How the qualification intelligence platform works — the flow, the AI layer, the data model, and the leader console. Click any element to see the detail behind it.
● Live — Betav0.5.0sdrhive.hyvara.aiDesign partner ready
Qualification flow
Platform architecture
AI and data layer
Leader intelligence
Release plan
Diagram 01
The qualification flow
For: Sales leaders · Design partners · Investors
Click any stage to see what happens there →
Entry
Anonymous visitor
qualify.hyvara.ai — no login
→
Authenticated SDR
sdrhive.hyvara.ai — full platform
↓
Input
Call notes pasted
Notes / Granola / Gong / CRM
+
Analysis tuning
Methodology · gates · intensity
↓
AI engine
Anthropic Claude — qualification engine
Scores every gate · generates coaching · flags gaps
↓
Gate scores
Confirmed
Gate fully evidenced
/
Partial
Needs follow-up
/
Missing
Gate not established
↓
Debrief
Debrief card + coaching thread
Pat on back · gap-fill · next questions
↓
Decision
Pursue
Add to follow-up queue
/
Needs more info
Re-qualify with coaching notes
/
Pass
Deal not worth pursuing
↓
Handoff
SDR submits handoff to AE
Score · gates · contact · recommendation
↓
AE review
Accept
AE takes the deal
/
Request more info
Gate selector + reason
/
Return
12-reason taxonomy
AE feedback loops back into SDR coaching flags and leader quality dashboard — the closed loop that makes qualification compound over time
Entry — Anonymous
Anonymous qualifier at qualify.hyvara.ai
Any visitor can paste call notes and receive one free analysis with no account required. The anonymous experience teases the full platform — methodology tuning, gate requirements, and SDR Offer Guardrails are all visible but locked, demonstrating depth before the conversion moment.
After 9 free analyses, a sign-up gate appears. The result remains visible but cannot be saved. This is the product demo: the SDR sees exactly what they're missing before they pay.
The anonymous flow is the top of the funnel — no friction, immediate value, converts to paid when the SDR is ready.
Entry — Authenticated
Authenticated SDR at sdrhive.hyvara.ai
After login, the SDR sees their full qualification history, growth metrics, follow-up queue, handoff pipeline, and leaderboard position. Their methodology configuration is inherited from their Sales Leader — they cannot change gate weights or coaching intensity if they're on a team plan.
Three experience layers:
Solo SDR — self-purchaser on Pro plan ($49/mo). Full tuning control. Self-managed.
Team SDR — company-purchased seat. Leader configures everything. SDR executes.
Leader — completely different shell. Coaching console, team pipeline, coaching flags, 1:1 prep, MBOs, QBR generation.
Input
Call notes — four input modes
The SDR pastes their call notes into one of four input tabs:
Call Notes — free-text paste from any source
Granola Export — direct paste from Granola AI meeting notes
Gong Transcript — paste from Gong call transcript
CRM Notes — paste from Salesforce or HubSpot activity notes
No CRM integration required. The SDR pastes what they have. The AI reads it regardless of format. The zero-friction input is intentional — the SDR should never have to clean or reformat their notes before getting an answer.
Input
Analysis tuning — methodology configuration
Before analysis runs, the tuning panel configures what the AI scores against:
Active Selling Blend — up to 5 active methodologies from 25+ library (Challenger, SPIN, Sandler, Value Selling, Never Split the Difference, and more). Weight sliders auto-constrain to 100%.
Gate Requirements — Required vs. Optional per gate. Leaders set this for team plans.
SDR Offer Guardrails — Approved and Restricted offer lists. Leader-managed.
Coaching Intensity — slider from Highly Encouraging to Brutally Honest
For team SDRs, all of this is locked by their leader. Solo SDRs control it themselves.
AI engine
Anthropic Claude — the qualification engine
The core AI layer runs on Anthropic Claude (claude-haiku-4-5 for speed, claude-sonnet-4-6 for depth). The model reads the call notes against the configured methodology and produces a structured JSON output covering:
Gate-by-gate scoring: CONFIRMED / PARTIAL / MISSING with evidence per gate
Coaching notes per gate: what to ask to confirm missing or partial gates
Overall qualification score (0-100)
Recommendation: Pursue / Needs More Info / Pass
Pat-on-back (positive reinforcement note)
Gap questions (specific follow-up questions for missing gates)
Follow-up tasks (action items auto-added to the follow-up queue)
Zero-retention API policy — call notes are processed and immediately discarded, never used for model training.
Gate result
Confirmed gate
A gate is confirmed when the call notes contain clear, specific evidence that the qualification criterion is met. For example, a MEDDPICC "Metrics" gate is confirmed when the notes include a specific business outcome the customer is trying to achieve — "reduce campaign cycle time from 12 days to 6 days" — not just "they want to be more efficient."
Confirmed gates show green in the analysis. The SDR's confirmation rate per gate is tracked over time and displayed in My Growth as a competency trend. High confirmation rates on a specific gate trigger peer coaching opportunities — the SDR may become a mentor to teammates who struggle with that gate.
Gate result
Partial gate
A gate is partial when there is some evidence in the call notes but it is incomplete, vague, or unconfirmed. For example, the customer mentioned budget exists "somewhere" but no specific figure or approval process was discussed.
Partial gates show amber with a specific coaching note — the exact question the SDR should ask in their next touchpoint to fully confirm the gate. These questions are the core of the coaching thread gap-fill experience: the SDR works through each partial gate in an iMessage-style dialogue, getting coached one gate at a time.
Gate result
Missing gate
A gate is missing when there is no evidence in the call notes that the qualification criterion was even discussed. For example, no mention of timeline, no discussion of decision process, no identification of an economic buyer.
Missing gates show red and are the primary driver of the coaching thread. For required gates (set by the leader), a missing gate can block the handoff — the SDR cannot submit a handoff until required gates are confirmed or explicitly waived by the leader.
Missing gates also feed the Leader Console coaching flags — the AI flags which SDRs are consistently missing the same gates and suggests targeted coaching interventions.
Debrief
Debrief card and coaching thread
After analysis, the SDR sees a debrief card before the full gate report — a warm, coaching-first experience designed to feel like feedback from a mentor, not a report card.
The debrief shows: pat on the back (positive reinforcement), one key coaching note, gate summary, and two buttons: Fill the Gaps and See Full Analysis.
Coaching thread (Pro+ only): an iMessage-style dialogue that walks through missing and partial gates one at a time. For each gate: "Did you discuss X? Yes / No / Sort of." The SDR answers and the AI provides the next follow-up question. Completing the thread populates the follow-up queue automatically.
The coaching tone adapts to the archetype signal — Socratic for high-ego SDRs, anchor-heavy for low-confidence SDRs, direct for calibrated SDRs.
Decision
Pursue — deal enters follow-up queue
When the SDR marks a deal as Pursue (or the AI recommends Pursue), the deal enters the Follow-up queue. Each item in the queue has:
SLA countdown — color-coded (amber approaching, red overdue)
Re-qualify button — opens the call sheet with prior context pre-loaded
Today's Calls queue — AI suggests up to 8 high-priority items per day based on score and SLA
Build My Day — one-click AI prioritization of the day's follow-up list
The follow-up queue is a qualification task manager, not a call log. It tells the SDR what to do next on each deal, not just that a deal exists.
Decision
Needs more info — re-qualify with coaching
When the deal is promising but gates are incomplete, the SDR marks it as Needs More Info. The deal enters the follow-up queue with specific gap questions auto-populated from the coaching thread.
When the SDR gets back on the phone with the prospect, they open the call sheet — a per-account view showing who to call, what was discussed in prior conversations, which gates still need confirmation, and a transcript upload field for the new call.
After re-qualifying, the SDR pastes the new call notes and runs a fresh analysis. The system compares gate scores across analyses to show whether qualification is improving or stalling.
Decision
Pass — deal removed from pipeline
When the analysis confirms the deal should not be pursued — missing critical gates, wrong buyer, no real need — the SDR marks it Pass. The deal is archived but not deleted.
Leaders can see pass rates per SDR in the Leader Console. Patterns in what gets passed reveal coaching opportunities: if an SDR consistently passes deals after two calls, they may need coaching on deal qualification earlier in the discovery process.
The Reject button in My Analyses is labeled DEL (not X) and requires inline confirmation — a small UX decision that prevents accidental deletions of valuable coaching data.
Handoff
SDR submits handoff to AE
When a deal is qualified above the leader's threshold, the SDR submits a formal handoff. The submission includes:
Qualification score and gate breakdown
Contact information and hierarchy
Key insight (one-line AI summary of why this deal matters)
Recommendation (Pursue / High Priority)
SDR notes and any transcript upload
The AE receives an email notification with a token-gated link — no login required to review. The AE Lite portal gives them a structured view of all pending submissions.
Handoff quality control: required gates must be confirmed before submission is allowed. The SDR cannot send an unqualified deal to the AE without the leader explicitly overriding the gate requirement.
AE review
AE accepts the handoff
When the AE accepts a handoff, the deal moves to active pipeline. The AE can rate the quality of the submission (1-5 stars) — this rating feeds back into the SDR's performance metrics and the leader's quality dashboard.
Accepted handoffs are tracked by SDR in the Leader Console's Handoff Quality tab — acceptance rate per SDR, time-to-accept, and pattern data on which SDRs send the best qualified deals over time.
The AE Lite portal is deliberately limited — it shows only what the AE needs to accept or return a deal. The banner on every AE Lite page says: "You're reviewing N SDR submissions this month. AE Hive gives you the full picture." — this is the upsell path to the $79/month AE Hive product.
AE review
AE requests more information
The AE can select specific gates that need more evidence, add a comment, and return the deal to the SDR's queue. The SDR sees exactly what the AE needs — no ambiguous rejections.
The back-and-forth lives in a threaded comment history per deal — visible to both SDR and AE, with email notifications on every action. The SDR responds, re-qualifies if needed, and resubmits.
Kickback reasons map to coaching categories in the Leader Console — if multiple AEs are kicking back deals from the same SDR for the same reason, the system flags it as a coaching priority for that SDR's 1:1 session.
AE review
AE returns the handoff
When the deal is genuinely unqualified, the AE returns it with a reason selected from the 12-reason rejection taxonomy. Each reason maps to a coaching category:
No real pain — coaching on Why Buy Anything gate
Wrong buyer — coaching on Economic Buyer identification
No timeline — coaching on Why Buy Now gate
Budget not established — coaching on Budget gate
Competitor already selected — coaching on competitive discovery
... and 7 more categories
The SDR can submit a rebuttal. The leader reviews and issues a ruling. The ruling is logged as part of the SDR's coaching record.
AE return rates flow directly into the SDR's Health Score in the sidebar — the leading indicator of whether their qualification is improving or degrading.
Diagram 02
Platform architecture
For: Engineering · CTOs · Design partners
Click any component to see its role →
Frontend
App framework
Lovable / React / TypeScript / Vite
Application UI and routing. Lovable Cloud deployment.
Routing
TanStack Router
Role-based routing: Solo SDR / Team SDR / Leader shells
Static hosting
Netlify
testhyve.hyvara.ai marketing site and team updates hub
SDR Hive is built on Lovable Cloud — a React + TypeScript + Vite stack that enables rapid product iteration through AI-assisted development. Every feature sprint is a Lovable session: Kevin architects the feature in plain language, Lovable generates the component, the team reviews and ships.
This architecture was chosen specifically because it lets a four-person founding team ship at a pace that would normally require 10-15 engineers. The Qualifier component alone is 2,400+ lines — it needs to be refactored into sub-components as a Days 31-60 priority to keep the codebase maintainable at scale.
Frontend
TanStack Router — role-based routing
TanStack Router handles the three-shell architecture: Solo SDR, Team SDR, and Leader each get a distinct authenticated experience after login. The role is determined at signup and stored in the Supabase profiles table.
Role-based routing post-login is a Days 31-60 priority — currently the shell is being finalized. The left sidebar (220px fixed, Lucide icons, section labels, amber active indicator, user avatar + plan badge at bottom) is the authenticated navigation anchor across all three roles.
Static hosting
Netlify — marketing and team updates
Netlify serves the testhyve.hyvara.ai marketing hub and the team updates site. These are static HTML files dropped directly into the Netlify folder — no build process, no framework, instant deploys.
The team updates site (this file's home) contains session readouts, architecture diagrams, product enhancement backlogs, and platform architecture documentation. It's the inside team's single source of truth for what was built and why — updated after every significant session.
DNS
Namecheap — domain management
All Hyvara domains are managed through Namecheap. Active domains and subdomains:
qualify.hyvara.ai — anonymous qualifier front door
testhyve.hyvara.ai — marketing landing page hub
Open item: Resend domain verification (hyvara.ai CNAME + TXT records) must be completed in Namecheap before handoff emails go live — this week priority.
Backend
Supabase — PostgreSQL + Edge Functions + Auth
Supabase is the entire backend — database, authentication, server-side functions, and storage. Key tables: analysis_history, handoffs, todos (follow-ups), coaching_threads, notifications, profiles.
Database indexes migration is a Days 31-60 priority — the analysis_history, handoffs, todos, and notifications tables need indexes for query performance at design partner scale.
Supabase Pro upgrade ($25/month) is recommended before the first design partner goes live — the free tier can pause, which would be catastrophic during a design partner demo. This is a this-week priority.
Security
Row-Level Security — data isolation
Row-Level Security (RLS) in Supabase ensures strict data isolation. Every table is gated by user_id — no user can ever query another user's data through the API, even if they have the correct credentials.
Team data visibility rules: Leaders see gate scores and analysis summaries for their team SDRs, but never the raw call note text. This is a core privacy commitment — the leader knows whether their SDR qualified a deal, not what the customer said word-for-word.
Terms acceptance is timestamped at first analysis and stored in the profiles table (terms_accepted_at). This is the legal timestamp for consent.
Payments
Stripe — subscription billing
Stripe handles all subscription billing. Currently in sandbox mode — moving to production is a Days 31-60 priority, timed to coincide with the first paying design partner.
Plan structure in Stripe:
SDR Pro — $49/month (1 SDR)
Team — $199/month (1 Leader + up to 5 SDRs)
Enterprise — custom pricing
AE Hive — $79/month (coming Q3 2026)
Active coupon codes: DESIGNPARTNER (first 3 months free), TEAMTRIAL (30-day team trial), SDRHIVE2026 (20% off annual).
Email
Resend — transactional email
Resend handles all transactional email from SDR Hive — handoff notification emails to AEs (including the token-gated AE Lite link), magic links for passwordless login, and approval notifications.
Resend requires domain verification (hyvara.ai CNAME + TXT records in Namecheap DNS) before emails go live. This is a this-week priority — without it, the handoff email flow is non-functional and AE Lite cannot be tested end-to-end.
AI engine
Anthropic Claude — qualification and coaching
Two Claude models in production:
claude-haiku-4-5 — used for speed-sensitive operations (quick qualification scores, coaching flag generation, debrief cards). Lower latency, lower cost.
claude-sonnet-4-6 — used for depth-sensitive operations (1:1 prep transcript analysis, QBR generation, full methodology analysis with coaching thread generation).
Zero-retention policy via Anthropic API — call notes are processed and immediately discarded. They are never used for model training. This is the core privacy commitment that makes enterprise adoption possible.
Anthropic monthly spend cap is being set in the Console this week — cost control before design partner scale.
Performance
Prompt caching — 60-80% API cost reduction
The analyze edge function (the main qualification API call) sends a large system prompt on every invocation — containing the full methodology library, gate definitions, and coaching instructions. This prompt is identical for every SDR on the same team configuration.
Prompt caching on the Anthropic API means the system prompt is cached after the first call and not re-sent on subsequent calls — only the new call notes are sent. This reduces API costs by 60-80% on the analyze function.
Implementation is a Days 31-60 priority — it's the single biggest performance and cost win available without any feature changes.
Analytics
Google Sheets / Apps Script — feedback and leads
Google Sheets via Apps Script serves as the lightweight analytics layer for early-stage feedback collection. When an SDR submits feedback from the in-app feedback tool (connected to Linear), the response is also logged to a Google Sheet for the team to review.
Lead capture webhooks from the marketing site (testhyve.hyvara.ai) log to Google Sheets — name, email, company, and interest area for every demo request or waitlist signup.
This is intentionally lightweight — no analytics platform overhead at the design partner stage. Proper analytics (Mixpanel, PostHog, or similar) is a Days 61-90 consideration when usage patterns are worth measuring.
Integrations
Salesforce / HubSpot — roadmap
Current state: Salesforce read-only integration is in the Days 31-60 roadmap — account name auto-fill from CRM when an SDR starts a new analysis. This eliminates manual data entry for the company name field.
Days 61-90: Bi-directional Salesforce sync — gate scores from SDR Hive written back to the CRM opportunity record. This closes the loop between qualification intelligence and pipeline reporting.
HubSpot follows the same pattern as Salesforce. Microsoft Dynamics is on the roadmap but lower priority.
Granola paste is already supported (copy-paste export, no API integration needed). Gong API integration is in the backlog for Days 31-60.
URL
qualify.hyvara.ai — anonymous front door
The anonymous qualifier is the top of the funnel. No login, no account, no friction. Any SDR who lands here can paste call notes and get a qualification analysis in 10 seconds.
The first analysis is always free. After 9 analyses, the sign-up gate appears — the result is visible but not saveable. This is the conversion moment: the SDR has experienced the value, seen the locked features (methodology tuning, gate requirements, guardrails), and understands exactly what they're missing on the free tier.
URL
sdrhive.hyvara.ai — authenticated app
The full SDR Hive application. After login, the SDR's role (Solo SDR / Team SDR / Leader) determines their shell. All three roles share the same domain but see completely different interfaces.
The left sidebar is the navigation anchor: 220px fixed, Lucide icons, section labels, amber active indicator, user avatar + plan badge at bottom. On mobile it collapses to a hamburger drawer.
URL
sdrhive.hyvara.ai — authenticated app
The full SDR Hive application. After login, the SDR's role determines their shell. All three roles share the same domain but see completely different interfaces behind the login wall.
URL
hyvara.ai/ae/[token] — AE Lite portal
The token-gated AE handoff review portal. No login required. The AE clicks the link from their email notification and lands on their personal dashboard showing all SDR submissions.
Each token is unique per AE and persistent — the same link always shows the same AE's pipeline. Tokens are generated in Supabase and stored against the AE's email address.
Building in Phase 2 (Days 31-60). The URL structure (hyvara.ai/ae/[token] vs a separate subdomain) is an open architectural decision for Kevin and Jon to confirm.
URL
testhyve.hyvara.ai — marketing hub
The Netlify-hosted marketing landing page that serves as the front door to all Hyvara products. Currently branches to SDR Hive and DCN. Will add AE Hive card when AE Hive reaches beta (Q3 2026).
Also serves the team updates hub — session readouts, architecture diagrams, and product documentation for the inside team. Static HTML, instant deploys, no build process.
Click any input or output to see what it contains →
Call intelligence inputs
Raw call notes (any format)
Granola AI export
Gong transcript
CRM activity notes
Configuration inputs
Qualification framework
Active selling blend
Gate requirements
SDR offer guardrails
Coaching intensity
Context inputs
Team case studies
Battle cards
Prior analysis history
1:1 meeting transcript
→
Anthropic Claude
claude-haiku-4-5 claude-sonnet-4-6
→
Zero-retention API — notes discarded after processing
Qualification score (0-100)
Composite across all gates and methodology weights
Gate results — per gate
CONFIRMED / PARTIAL / MISSING + evidence
Coaching notes
What to ask to confirm each missing gate
Debrief card content
Pat on back · coaching note · gap questions
Follow-up tasks
Auto-added to queue with SLA and context
Recommendation
Pursue / Needs More Info / Pass
SDR archetype signal
High Ego / Low Ego / Calibrated — drives coaching tone
1:1 prep output
Action items · talking points · deal review
Different from Gong and Chorus: Those tools record and transcribe calls retrospectively. SDR Hive intervenes at the moment of qualification — scoring gates, generating coaching, and producing follow-up actions before the deal ever reaches the AE. Retrospective ≠ real-time intelligence.
Input
Raw call notes
The SDR pastes whatever they have — bullet points, paragraph form, shorthand, fragmented sentences. The AI reads it regardless of format. No cleaning or reformatting required.
The zero-friction input is a deliberate product decision. The biggest failure mode in qualification tools is that SDRs have to do extra work to use the tool. If the SDR has to reformat their notes before pasting, adoption collapses. SDR Hive reads the notes exactly as they are.
Input
Granola AI export
Granola is an ambient meeting intelligence tool that auto-generates structured meeting notes. SDR Hive has a dedicated Granola tab — the SDR copies the Granola export and pastes it directly. The AI handles the Granola format natively.
Note: there was a Granola security issue fixed in v0.5.0 — the Granola export format was being mishandled in a way that could expose note content. This has been resolved.
Input
Gong transcript
Gong call transcripts can be pasted directly into the Gong tab. The AI reads the full conversation transcript and scores it against the configured methodology — pulling out key moments where qualification gates were addressed (or missed).
Gong API integration is in the Days 31-60 backlog — direct pull from Gong without copy-paste. Until then, paste is fully functional.
Input
CRM activity notes
SDRs can paste activity notes directly from Salesforce or HubSpot. The AI handles the typically terse, acronym-heavy format of CRM notes.
Salesforce read-only integration (auto-fill from CRM) is a Days 31-60 priority — this eliminates the copy-paste step entirely and lets the SDR pull notes directly from their open opportunities.
Configuration
Qualification framework
One primary framework is selected per team configuration:
BANT — Budget, Authority, Need, Timeline (foundational)
BMANTR — Budget, Metrics, Authority, Need, Timeline, Risk (extended BANT — David Jones' preferred framework)
The framework determines which gates are scored and how they're defined. Leaders set this for their team — individual SDRs on team plans cannot override it.
Configuration
Active selling blend
Up to 5 active selling methodologies can be blended simultaneously, each with a weight slider that auto-constrains to 100% total. The 25+ methodology library includes:
Never Split the Difference — Voss negotiation techniques
20+ additional methodologies (most locked on free tier)
The blend tells the AI which selling techniques to look for evidence of in the call notes — did the SDR use a reframe? Did they ask implication questions? Did they anchor on pain?
Configuration
Gate requirements
Each gate in the selected framework can be configured as Required or Optional by the Sales Leader. Required gates must be confirmed before the SDR can submit a handoff — the system blocks submission and shows exactly which gates are preventing the handoff.
This is the handoff quality control mechanism. It stops unqualified deals at the source rather than filtering them at the AE level. The AE accepts fewer garbage handoffs; the SDR learns exactly what qualified means for their team.
Configuration
SDR offer guardrails
The Sales Leader configures two lists: Approved offers (what SDRs are allowed to promise in a discovery conversation — trial access, executive sponsor meeting, specific product demo) and Restricted offers (what SDRs are never allowed to promise — custom pricing, features not in the roadmap, non-standard SLAs).
The AI scores whether the call notes contain any restricted offers and flags them in the debrief. This protects the AE from inheriting a deal where the SDR over-promised — one of the leading causes of deal failure at the transition point.
Configuration
Coaching intensity
A slider from Highly Encouraging to Brutally Honest that controls the tone of the AI's coaching output. At the Highly Encouraging end, the AI leads with positives, softens critical feedback, and frames gaps as opportunities. At the Brutally Honest end, the AI is direct about missing gates and clear that the deal is not qualified.
Combined with the archetype signal (High Ego / Low Ego / Calibrated detected from the SDR's language patterns), the coaching tone adapts per-SDR. A high-ego SDR on Brutally Honest gets Socratic questioning. A low-confidence SDR gets anchor-heavy positive reinforcement.
Context
Team case studies
Leaders upload PDF or Word case studies to the Knowledge Base. The AI extracts a structured story from each case study: talk track first, then a 2×2 Before / Trigger / Solution / After grid. Private case studies can be anonymized for competitive sensitivity.
Case studies feed the qualification analysis — when the AI detects a deal that closely matches a prior case study, it surfaces the relevant talk track and success metrics as part of the coaching output. "You have a deal that looks like the Northstar Holdings implementation — here's what worked there."
Context
Battle cards
Battle cards can be uploaded as documents or built via a 10-question voice narration wizard: elevator pitch, competitor comparisons, common objections and responses, discovery questions.
Once in the Knowledge Base, battle card content feeds the AI's private brain — when the analysis detects a competitive situation mentioned in the call notes, the relevant battle card content surfaces in the coaching output without the SDR having to search for it.
Context
Prior analysis history
Every prior analysis for the same account is accessible when the SDR re-qualifies a deal. The call sheet shows what was discussed in prior conversations, which gates were confirmed or missing, and the coaching notes from each session.
Over time, the system tracks gate confirmation rates per SDR across all analyses — this is the data that drives the My Growth competency map, the leaderboard, and the Leader Console coaching flags. The intelligence graph grows with every analysis run.
Context
1:1 meeting transcript
In the Leader Console's 1:1 Prep section, leaders can import a post-meeting transcript (Granola, Gong, or paste). The AI extracts action items from the transcript, separates them from discussion points, and populates the Action Items section.
Action items carry forward from session to session with age tracking. Items that haven't been completed after two weeks get a stall alert. This creates a year-long coaching record per SDR — 52 weeks of documented coaching moments, action item completion rates, and deal progression data at the end of the year.
AI engine
How the qualification engine works
The analysis edge function in Supabase sends a structured prompt to the Anthropic API containing:
The system prompt: full methodology configuration, gate definitions, coaching instructions, tone guidance, and output schema
The user message: the call notes pasted by the SDR
The model returns a JSON object with all output fields. The frontend parses this JSON and renders the analysis UI.
Prompt caching (Days 31-60) will cache the system prompt — only the call notes will be sent on repeat calls from the same team configuration, reducing API costs 60-80%.
The CleanMyMac-style loading animation (scanning rings, gate-by-gate reveal with ascending tones, score circle counting up with particle effects) is the product moment that makes the analysis feel alive, not like a form submission.
Output
Qualification score (0-100)
A composite score calculated from gate confirmation rates weighted against the configured methodology blend. A deal with all required MEDDPICC gates confirmed and strong Challenger evidence would score 85+. A deal with two missing required gates and no methodology evidence would score below 40.
The score circle counts up with particle effects on high scores — this is a deliberate product moment. The score is the truth in 10 seconds that the SDR has been waiting for.
Scores are tracked over time per SDR in My Growth and aggregate to the team view in the Leader Console. Trend lines show whether a team's average qualification score is improving week-over-week.
Output
Gate results — per gate
For every gate in the configured framework, the AI returns:
Evidence: the specific text from the call notes that supports the status
Coaching note: for PARTIAL and MISSING gates, the exact question to ask in the next conversation
Gates are displayed with an info icon that explains what each gate means (Why Buy Anything? MEDDPICC?) — educational for new SDRs, reference for experienced ones.
Output
Coaching notes
For every PARTIAL or MISSING gate, the AI generates a specific coaching note — not a generic reminder, but the exact question the SDR should ask in their next conversation with this prospect.
Example for a missing Why Buy Now gate: "You've confirmed there's a need, but you haven't established urgency. In your next call, ask: 'What happens to your team if this problem isn't solved by Q3?' — connect their stated pain to a specific deadline or consequence."
This is the core differentiation from retrospective tools like Gong. The coaching is prospective — it tells the SDR what to do next, not just what they missed.
Output
Debrief card content
The debrief card appears before the full analysis — a warm, coaching-first screen that the SDR sees immediately after the analysis completes.
The debrief contains:
Pat on back — one specific positive observation about what the SDR did well
One coaching note — the single most important thing to work on next
Gate summary — confirmed / partial / missing counts at a glance
Two buttons — Fill the Gaps (coaching thread) and See Full Analysis
The debrief is the difference between a report card and a coaching conversation. The SDR hears positive feedback first, then specific direction — not a list of everything they got wrong.
Output
Follow-up tasks
After analysis, specific follow-up tasks are auto-generated and added to the SDR's Follow-up queue. Each task includes:
What to do (call, email, research, re-qualify)
The specific gate to address
The account name and contact
An SLA countdown (color-coded amber/red as deadline approaches)
The Today's Calls queue (v0.6 feature) lets the AI suggest up to 8 high-priority follow-up items per day based on score and SLA — the SDR clicks "Build My Day" and gets a prioritized call list without having to decide what to work on.
Needs More Info — promising signals but gates incomplete, follow up before handoff
Pass — insufficient evidence of fit, urgency, or budget to justify AE time
The recommendation is advisory — the SDR can override it. But overrides are logged, and if an SDR consistently overrides Pass recommendations to submit handoffs that AEs then reject, that pattern shows up in the Leader Console as a coaching flag.
Output
SDR archetype signal
The AI infers the SDR's communication style from their call note language patterns and flags one of three archetypes:
High Ego — confident, direct language. Coaching uses Socratic questioning — the SDR is led to discover the gap themselves rather than being told.
Low Ego — tentative, qualifying language. Coaching uses anchor-heavy positive reinforcement — lots of confirmation before introducing the challenge.
Calibrated — balanced, analytical language. Coaching is direct — here's what's missing, here's how to get it.
The archetype adapts the coaching tone per-SDR automatically, without any manual configuration.
Output
1:1 prep output
When the leader imports a post-meeting transcript into the 1:1 Prep section, the AI generates:
Action items — extracted from the transcript, separated from discussion points, with age tracking and stall alerts
Talking points — agenda suggestions for the next session based on deal review and coaching flag data, with expandable Why context
Deal review — gate status per account for the SDR's active pipeline
The leader can export the full 1:1 prep as a printable HTML doc or copy to clipboard for use in any format.
Diagram 04
Leader intelligence engine
For: Sales leaders · VP Sales · Investors
Click any module to see what it surfaces →
Team pipeline
Deal pipeline — full team view
Every analysis from every SDR in one table. Click through to gate detail per account. Filter by SDR, score range, recommendation. The leader never has to ask an SDR "what's in your pipeline."
See detail →
SDR performance
Per-SDR performance cards
Score, analyses, gate completion %, handoff acceptance %, trend line per SDR. Always shows all team members — real data for active SDRs, placeholder for new hires. No SDR is invisible.
See detail →
Coaching flags
AI-generated coaching flags
The AI reads every recent analysis and flags patterns — SDR consistently missing the Why Buy Now gate, SDR submitting unqualified handoffs, SDR's score trending down. Send to SDR button turns green when sent.
See detail →
1:1 prep
AI-generated 1:1 preparation
Transcript import, action items with age tracking and stall alerts, talking points with Why context, deal review, career snapshot. A year-long coaching record per SDR, built automatically from every session.
See detail →
Handoff quality
Handoff quality analytics
Total handoffs, acceptance rate, needs-more-info rate, avg time to handoff. Per-SDR breakdown. Top rejection reason summary. The data that tells the leader which SDRs are sending qualified deals and which aren't.
See detail →
Peer enablement
Peer coaching network
System identifies who's strongest at each gate and flags peer coaching matches. When Jack's gate rate improves after watching Fran's tip, Fran earns an Impact badge — the rarest because it proves the content worked.
See detail →
MBOs
Objectives and SPIFF management
Per-SDR objective tracking with quarter selector. Objective templates library. Bonus Calculator with adjustable pool and instant payout calculation. SPIFF configuration. History tab with closed quarters.
See detail →
QBR
Quarterly business review generation
Quarter / year selector. Generate QBR button produces a full narrative from team data. Previous QBRs listing. Copy deck prompt → exports AI prompt for use in Claude / ChatGPT / Gemini to build slides.
See detail →
Knowledge base
Team knowledge management
Case studies, battle cards, and playbooks uploaded and managed by the leader. AI extracts structured content from documents. Private brain feeds the qualification analysis for every SDR on the team.
See detail →
Branding
Team branding and configuration
Organization name, logo upload, brand colors, website color extraction via Claude API. The team's SDR Hive instance reflects their company brand — not a generic tool interface.
See detail →
Where SDR Hive sits in the Hyvara revenue relay
SDR Hive← you are here
→
AE Hive Q3 2026
→
SE Hive Q4 2026
→
DCN Live
→
CSM Hive 2027
Leader console
Team pipeline — full deal view
The Team Pipeline tab shows every analysis run by every SDR on the team in one table. The leader can filter by SDR, score range, recommendation status, and date range.
Clicking any row opens a gate detail dialog — the full analysis for that account, including evidence per gate, coaching notes, and the SDR's decision (Pursue / Pass / Needs More Info).
This is the data the leader needs to answer the question their VP will ask in the weekly pipeline review: "What do we actually have in early-stage pipeline, and how qualified is it?" SDR Hive gives a specific, gate-level answer to that question — not a list of meetings that happened.
Leader console
SDR performance cards
A card per SDR showing:
Average qualification score (this week, trend vs last week)
Analyses run (activity indicator)
Gate completion % (are they improving on their weak gates?)
Handoff acceptance rate (the AE's verdict on their work)
Trend line — 4-week sparkline showing whether performance is improving or declining
Always shows all team members — real data for active SDRs, placeholder cards for new hires who haven't run their first analysis yet. No SDR is invisible in the dashboard.
The SDR Health Score in the sidebar (v0.6 feature) shows three micro-indicators at a glance: avg score this week, calls this week, handoff rate 30 days.
Leader console
AI-generated coaching flags
The system reads every analysis run in the past 7 days and generates coaching flags — specific, actionable interventions the leader should make before the next 1:1.
Examples:
"Marcus has missed the Why Buy Now gate on 4 of his last 5 analyses. He may not understand urgency discovery — consider reviewing his last 3 call recordings together."
"Sarah's handoff acceptance rate dropped from 78% to 42% this week. Two AEs cited 'No economic buyer identified' as the reason. This needs a coaching conversation today."
"James hasn't run an analysis in 6 days. Activity has stalled."
The leader can Send to SDR (button turns green when sent — the SDR sees the flag in their interface) or Dismiss (removes with a fade animation, logged as acknowledged).
Leader console
AI-generated 1:1 preparation
The 1:1 Prep section is pre-selected to the most urgent SDR on load (highest flag count or most overdue action items). Sections in order:
Post-Meeting Transcript Import — prominent, orange border. Paste the Granola or Gong export from the last 1:1. AI extracts action items.
Action Items — check off, carry forward, age tracking, stall alerts for items unresolved after 14 days.
My Notes — free text with dictation button.
Talking Points — AI-suggested agenda with expandable Why context (? button explains why each topic matters now).
Deal Review — gate status per account for this SDR's active pipeline.
Career & Goals — MBO %, badges, readiness trend.
Open as Doc exports a printable HTML version. Copy to Clipboard. This is the 52-week coaching record — every session documented, every action item tracked.
Leader console
Handoff quality analytics
The Handoff Quality tab answers the question the VP Sales asks every week: "Why are our SDR handoffs being rejected?"
Metrics shown:
Total handoffs submitted (team total, per SDR)
AE acceptance rate (% accepted vs kicked back vs returned)
Needs-more-info rate (partial quality issue)
Average time from analysis to handoff submission
Top rejection reason summary — which of the 12 rejection categories is most frequent
The top rejection reason is the most actionable number in the dashboard. If "No timeline established" is the top reason across 60% of rejections, the leader knows exactly what to coach in next week's 1:1s.
Leader console
Peer enablement engine
The system tracks gate confirmation rates per SDR across all analyses. When SDR A has >75% confirmation rate on a specific gate and SDR B has <40%, a peer coaching match is flagged.
The leader sees:
All suggested peer matches — can approve, modify, or override pairings
Content gap alerts — gates with no peer content yet ("Nobody has shared a Why Buy Now tip")
Contribution analytics — who is sharing, what is being watched, what is producing improvement
The badge system rewards contribution: First Tip (Bronze), Mentor (Silver — tip reaches 5 views), Content Leader (Gold — 100 total views), Impact (Legendary — a teammate's gate rate improves after consuming your content). Impact is the rarest badge because it requires measurable downstream improvement — not just content creation.
Leader console
MBOs and SPIFF management
Management by Objectives (MBOs) are quarterly performance targets set by the leader per SDR. SDR Hive tracks progress automatically from the qualification data — no manual updates.
The Bonus Calculator lets the leader input the bonus pool size and instantly calculate payout amounts based on current MBO completion percentages. This replaces the spreadsheet the leader currently builds manually at quarter-end.
SPIFFs (Sales Performance Incentive Funds) are short-term behavioral incentives — "Double your Why Buy Now confirmation rate this week and earn a $50 bonus." The active SPIFF banner appears on the SDR's Leaderboard view. The SPIFF history tab shows all closed quarters and payout history.
Leader console
Quarterly Business Review generation
The leader selects a quarter and year, clicks Generate QBR, and receives a full narrative QBR document generated from the team's actual performance data — qualification scores, handoff acceptance rates, gate improvement trends, coaching flag history, MBO completion.
No PPTX or PDF generation (avoiding the Gradial cautionary tale — don't overbuild). Instead, the QBR exports a Copy Deck Prompt — a structured AI prompt the leader can paste into Claude, ChatGPT, or Gemini to generate a full slide deck. The leader owns the tool they already use; SDR Hive gives them the data and the prompt.
Previous QBRs are stored and accessible in the History tab.
Leader console
Knowledge base management
The leader uploads and manages the team's private knowledge base — case studies, battle cards, and playbooks. The AI processes each document on upload and makes its content available to the qualification engine for every SDR on the team.
Case studies are extracted into a structured format: talk track, 2×2 Before/Trigger/Solution/After grid. Private case studies can be anonymized — the client name is hidden but the story structure is preserved.
The leader can request a team tip on a specific gate directly from the Knowledge Base view: "Nobody has shared a Why Buy Now tip yet — request a tip from your top performer on this gate." This connects the knowledge gap alert to a peer enablement action in one click.
Leader console
Team branding configuration
The Branding tab lets the leader configure how SDR Hive appears to their team:
Organization name — shown in the left sidebar and emails
Logo upload — replaces the Hyvara logo in the authenticated interface
Brand colors — primary and accent colors for the team shell
Website color extraction — paste the company website URL, Claude extracts the brand color palette automatically
This is a retention mechanic as much as a branding feature. When the interface shows the leader's company name and colors, it feels like their product — not a third-party tool they're renting. Switching means rebuilding their brand configuration, their methodology, their knowledge base, and their entire team history from zero.
Next in relay
AE Hive — Q3 2026
AE Hive is the deal progression layer — what happens after the SDR handoff is accepted. The AE owns the relationship, the timeline, and the close. AE Hive gives them:
Mutual Success Plan builder — documents what was promised and how success is measured
Stakeholder map — every contact, their role, their concerns, their position on the deal
Deal risk scoring — flags when a deal goes quiet, contacts go dark, or timeline slips
DCN handoff — when partner relationship capital is needed, AE initiates DCN from within AE Hive
The SDR Hive handoff context flows forward into AE Hive automatically — gate scores, methodology confirmation, contact hierarchy, key insights. The AE never has to re-ask discovery questions the SDR already answered.
Next in relay
SE Hive — Q4 2026
SE Hive is the technical discovery and proof layer — between AE Hive and the deal close. The SE runs deep discovery, designs POC scope, manages demo alignment, and produces a technical win.
Key capabilities planned: demo alignment to stakeholder-specific concerns (AI captures who said what in the intro meeting), POC scope guardrails against scope creep, RFP ingestion engine (70-90% auto-generated from knowledge base), and SE onboarding acceleration (field-ready in 10-14 days instead of 90+).
Live in relay
Deal Collaboration Network — Live MVP
The DCN is the partner co-sell layer — when the AE needs partner relationship capital to close a deal. The SDR Hive qualification output becomes the deal context that powers the DCN Stage 4 intel exchange.
When an SDR qualifies a deal and passes it to an AE, the handoff contains scored gate data, methodology confirmation, contact hierarchy, and key insights. When that AE enters the DCN and requests partner relationship capital, that SDR Hive context is the deal narrative the partner evaluates before engaging.
Better SDR qualification = richer DCN intel = faster partner engagement = higher deal velocity. The two products compound each other.
Roadmap
CSM Hive — Q3 2027
CSM Hive closes the loop — from implementation through renewal and back to the top of the funnel. The CSM inherits the full conversation history from SDR through SE through implementation: what was promised, what was built, what the customer struggled with, what delighted them.
The CSM Hive solves the founding insight: no Customer Success Manager should ever ask "why did you buy this product?" when the answer is documented in the Hyvara intelligence layer from the first SDR discovery call. The relay makes that promise real.
Diagram 05
Release plan — v0.6 · v0.7 · v0.8
For: Inside team · Jon · Design partners
Click any feature, bug, or decision to see the full detail →
Bug fixes — ship before any new features
P0 · v0.6.0
Leaderboard double (you) bug
is_me flag fires for multiple rows. Only authenticated user's row should be marked.
P0 · v0.6.0
Coaching Flags — Loading… forever
Leader console Coaching Flags tab stuck on loading state. Debug and fix data fetch.
P1 · v0.6.0
Account name truncation in My Analyses
Company name truncates at 4 characters. Expand column or use tooltip.
P1 · v0.6.0
Career Path missing from TOOLS nav
Was in v0.5.0 SDR nav spec but did not render. Add it back.
v0.6.0
The Coaching Companion
Design partner: Colin Schap — Ensemble
P0 · All tiers
Analysis system prompt update
New JSON fields: pat_on_back, coaching_notes, gap_questions, follow_up_tasks, archetype_signal
P0 · All tiers
Debrief card
First screen after analysis — pat on back, coaching note, gate summary, two action buttons
P0 · Pro+
Coaching thread / gap-fill modal
iMessage-style thread. One gate per exchange. Yes/No/Sort of + mic.
P1 · All tiers
Today's Calls queue
New tab in Follow-ups. Up to 8 items/day. Build My Day button.
P1 · All tiers
SDR Health Score in sidebar
Three micro-indicators: avg score, calls this week, handoff rate 30d
P1 · Pro+
Archetype-adaptive tone
High ego: Socratic. Low ego: anchor-heavy. Calibrated: direct.
v0.7.0
The Volley
Design partner: Peter Webster — Slalom
P0 · Team
The Volley — SDR to AE thread
SDR triggers handoff. Creates connection thread visible to both sides.
P0 · Team
AE Lite email link
Non-seated AEs receive time-limited link. No seat required to respond.
Manager cannot see analysis until both SDR and AE have signed off on handoff.
P1 · Team
Persona field in handoff payload
Brand or Partner persona travels with the handoff brief
P1 · Team
Coaching thread in handoff brief
AE sees score, gaps, and what the coaching thread revealed
v0.8.0
Gamification + Multi-Call
Design partner: PwC (TBD contact)
P0 · Pro+
Score progression across calls
Multi-call transcript blending. Gate movement shown visually across sessions.
P0 · Pro+
Prep My Call screen
AI writes opener, middle questions per missing gate, and close ask before the call.
P1 · Team
Weekly coaching digest via Resend
Monday morning email to SDR and Team Leader. Score trend, coaching moment, 1:1 focus.
P1 · Team
Leader Console restructure
Four sections: Team Dashboard, Managing Up, Team Ops, Platform Config
P1 · Team
1:1 Prep pre-loaded from coaching thread
Leader opens 1:1 Prep and sees week's coaching data already surfaced. No manual prep.
P2 · Pro+
Tone calibration learning
System learns per-SDR archetype over time from response patterns
Open architecture decisions — must resolve before build
Gates v0.7 · Owner: Kevin
Volley thread location — own route (/volley) or sidebar nav?
Determines whether the Connection Spine gets a dedicated UX home or lives inside existing nav patterns. Must decide before v0.7 prompts are written.
Gates v0.7 prompt 1 · Owner: Kevin + Colin
AE Lite token expiry — 7 days or 30 days?
Affects how urgently AEs need to respond and what happens to stale Volleys. Thread locks or stays read-only on expiry?
Gates v0.8 prompt 1 · Owner: Kevin
Multi-call grouping — how does Call 2 link to Call 1?
Account name match (automatic) vs SDR manually links them vs follow-up queue item triggers the link. Architecture implications for call_sessions table.
Gates v0.8 prompt 1 · Owner: Kevin + Jon
Leader Console section names and tab allocation
All four sections (Team Dashboard, Managing Up, Team Ops, Platform Config) need confirmed names and confirmed tab lists before the restructure prompt is written.
The is_me flag in the leaderboard query is returning true for multiple rows — not just the authenticated user's own row. The result is that multiple SDRs on the leaderboard appear highlighted as "you" simultaneously.
Fix: Correct the is_me flag logic so it only fires true for the single row where user_id matches the authenticated session. All other rows must return false. Run a smoke test across all three leaderboard period filters (Week / Month / Quarter) before marking resolved.
Bug · P0 · v0.6.0
Coaching Flags tab — Loading… forever
The Leader Console Coaching Flags tab is stuck in a perpetual loading state. The data fetch never resolves — the tab shows a spinner that never clears.
Fix: Debug the data fetch on the Coaching Flags component. Check the Supabase edge function for the flags query — likely a missing await, an unhandled null response, or a row-level security policy blocking the leader's read access to their team's flag data. Fix and confirm the tab loads with both real data and the seeded demo data.
Bug · P1 · v0.6.0
Account name truncation in My Analyses
Company names in the My Analyses table are truncating at approximately 4 characters — "Acme Corporation" renders as "Acme" with no indication that text has been cut.
Fix: Expand the account name column width, or implement a tooltip that shows the full company name on hover. The truncation is caused by a fixed column width that is too narrow for typical enterprise account names. Do not truncate at fewer than 20 characters without a visible indicator (ellipsis + tooltip).
Bug · P1 · v0.6.0
Career Path missing from TOOLS nav
Career Path was specified in the v0.5.0 SDR navigation spec under the TOOLS section but did not render in the final build. The nav item is missing entirely.
Fix: Add the Career Path nav item back to the TOOLS section of the left sidebar. It should link to the Career Path section within My Growth. Confirm the route resolves correctly and the section scrolls into view on click.
v0.6 · Prompt 1
Analysis system prompt update
The first prompt in the v0.6 sequence. No UI changes — only the system prompt and JSON schema are modified.
New fields added to the analysis JSON output:
pat_on_back — one specific positive observation about what the SDR did well in this call
coaching_notes — per-gate coaching notes for PARTIAL and MISSING gates
gap_questions — specific follow-up questions for missing gates, ready to use in the next conversation
follow_up_tasks — array of tasks to auto-add to the follow-up queue
archetype_signal — High Ego / Low Ego / Calibrated detected from SDR's language patterns
Companion tone instruction added to the system prompt: "Ask First, Tell Second. The SDR hears coaching, not judgment."
Run this prompt first. Confirm the JSON schema is correct before firing Prompt 2.
v0.6 · Prompt 2
Debrief card
The first screen the SDR sees after analysis completes — before the full gate report. Rendered as a new component between the loading animation and the existing analysis view.
Contents:
Pat on the back (warm, positive opener from the pat_on_back field)
One coaching note (the single most important gap from coaching_notes)
Gate summary (confirmed / partial / missing counts at a glance)
Two buttons: Fill the Gaps (opens coaching thread — Pro+ gated) and See Full Analysis
Free users see Fill the Gaps with a lock icon and a trial prompt. Steps 1-3 (debrief card) remain free. Steps 4-5 (gap-fill thread + queue population) are Pro+.
Success target: 80%+ of SDRs view the debrief card rather than skipping directly to full analysis.
v0.6 · Prompt 3
Coaching thread / gap-fill modal
iMessage-style dialogue that walks through missing and partial gates one at a time. Pro+ only.
Flow: for each missing or partial gate — "Did you discuss [gate]? Yes / No / Sort of." SDR responds (tap or mic). AI provides the next follow-up question. Completing the thread populates the follow-up queue automatically with specific tasks per gate.
Data: saves to new coaching_threads table (user_id, analysis_id, messages jsonb, archetype, completed). Requires SQL migration before prompt fires: CREATE TABLE coaching_threads (...) ALTER TABLE todos ADD COLUMN is_today boolean DEFAULT false ALTER TABLE todos ADD COLUMN sort_score numeric
Success target: 60%+ of analyses result in completed gap-fill within first 2 weeks with Colin.
v0.6 · Prompt 4
Today's Calls queue
A new tab in Follow-ups. The SDR pulls up to 8 high-priority items per day from their follow-up queue. The AI suggests the list based on qualification score and SLA urgency — the highest-scoring deals with the most overdue follow-up timers surface first.
Build My Day button: one click generates the SDR's prioritized call list for the day without requiring them to manually sort through their full queue.
Score sorting also added to the main Follow-ups view: sort by Score asc/desc, Date Added, SLA. Default sort: score descending. Color-coded score rings already built in v0.5 — no new UI work needed there.
v0.6 · Prompt 5
SDR Health Score in sidebar
Three micro-indicators visible in the left sidebar at all times — the SDR always knows where they stand without navigating to My Growth.
Three indicators:
Avg qualification score this week (vs last week — up/down arrow)
Calls this week (analyses run — activity indicator)
Handoff rate 30 days (% of Pursue decisions that became submitted handoffs)
Clicking the Health Score navigates to My Analytics for the full view.
This is the last prompt in the v0.6 sequence. Run after Prompts 1-4 are confirmed working.
v0.6 · P1 · Pro+
Archetype-adaptive tone
The archetype_signal field from the system prompt update drives tone adaptation in the coaching thread and debrief card:
High Ego — SDR uses confident, assertive language. Coaching uses Socratic questioning — the SDR is led to discover the gap themselves rather than being told directly.
Low Ego — SDR uses tentative, qualifying language. Coaching uses anchor-heavy positive reinforcement — lots of confirmation before introducing the challenge.
Calibrated — SDR uses balanced, analytical language. Coaching is direct — here's what's missing, here's the specific question to ask next.
The archetype is detected per-analysis from the SDR's call note language, not a permanent profile setting. It can shift call to call as the SDR's confidence develops.
v0.7 · P0 · Team
The Volley — SDR to AE handoff thread
The Volley is the first instance of the Connection Spine inside SDR Hive. When an SDR marks a deal as Pursue and is ready to hand it off, they manually trigger a Volley — creating a connection thread visible to both the SDR and the AE.
The thread contains:
The qualification score and gate breakdown
The coaching thread data (what the SDR knows, what gaps remain)
The contact information and persona field (Brand vs Partner motion)
Architecture decision required before build: does the Volley live at /volley route or inside the existing sidebar nav? This must be resolved before the prompt is written.
v0.7 · P0 · Team
AE Lite email link
AEs who don't have a Hyvara seat receive a time-limited token link via Resend email when a Volley is sent to them. No login required — the link takes them directly to their AE Lite view.
Requires new Supabase table: volley_tokens (token, volley_id, expires_at, opened_at).
Open decision: token expiry — 7 days or 30 days? What happens when the token expires? Thread locks (AE can no longer respond) or stays read-only (AE can view but not act)?
Edge case to resolve: when the AE is already an SDR Hive user (e.g. colin.ensemble), does the Volley go through the Lite email link or directly to their in-app inbox? Decision needed before Prompt 1.
v0.7 · P0 · Team
Five SAL outcomes
When an AE receives a Volley, they have five possible responses:
Accept — deal is qualified, AE takes it. SDR notified. Deal moves to active pipeline.
Kickback-for-gates — deal is promising but specific gates need more evidence. AE selects which gates. SDR notified with specific gap.
Reject-not-now — deal is real but timing is wrong. SDR can re-Volley in a future quarter.
Reject-not-a-fit — deal is not a fit for this AE's territory or product. SDR coaching flag generated.
Each outcome generates a notification to the SDR and logs to the volley outcomes record. Outcome data feeds the Leader Console Handoff Quality tab.
v0.7 · P0 · Team
Dirty laundry protection
The manager cannot see the full analysis until both the SDR and the AE have agreed on the handoff outcome. Upward visibility is gated on both parties signing off.
Why this matters: SDRs will not use the Volley if they believe their manager sees every low-scoring analysis before it reaches the AE. The dirty laundry protection gives SDRs control — they decide when they're ready to send, and the manager only sees the agreed outcome, not the raw qualification journey.
Implementation: the volleys table status field gates leader visibility. Status = 'agreed' (both sides signed off) unlocks the analysis view for the leader. Status = 'pending' or 'in_review' keeps it private to SDR and AE only.
Success signal: SDRs do not game scores to avoid manager visibility after launch.
v0.7 · P1 · Team
Persona field in handoff payload
The Volley brief includes a persona field that tells the AE which motion the SDR was running: Brand AE (direct software sale) or Partner motion (co-sell opportunity that may need DCN involvement).
The persona field connects to the Connection Spine — when an AE receives a Volley tagged as Partner motion and responds with Refer-to-partner, the system has the context to route toward DCN matching without requiring the AE to re-explain the deal context.
v0.7 · P1 · Team
Coaching thread data in handoff brief
The AE receives not just the gate scores but the full coaching thread context — what the SDR knows, what gaps remain, and what the gap-fill dialogue revealed about the deal.
This gives the AE a richer picture than a score alone. Instead of "MEDDPICC score: 72" they see "Timeline: PARTIAL — customer mentioned Q4 budget cycle but didn't confirm a specific decision date. Coaching thread: SDR will ask about board approval process in next call."
The AE can make a much better accept/kickback decision with this context than with a number alone.
v0.8 · P0 · Pro+
Score progression across calls
Multi-call transcript blending — the SDR pastes the second call transcript and the system re-scores the deal showing which gates moved (confirmed, improved, or regressed) compared to the first call.
Visually: a gate progression chart showing Call 1 → Call 2 → Call 3 gate status per gate. The SDR sees their qualification arc, not just a snapshot.
Requires new Supabase table: call_sessions (analysis_ids[], final_score, gates_progression jsonb).
Architecture decision required: how does the system know Call 2 is related to Call 1? Options: account name match (automatic), SDR manually links them, or follow-up queue item triggers the link. This must be decided before the prompt is written — the data model differs significantly between options.
v0.8 · P0 · Pro+
Prep My Call screen
A full pre-call coaching screen the SDR opens before their follow-up conversation. The AI writes a complete call script based on the prior analysis and coaching thread.
Three sections:
Opener — callback to the last conversation. Specific reference to something the prospect said, not a generic "following up on our last call."
Middle — target questions per missing gate. The exact questions from the coaching thread, sequenced for a natural conversation flow.
Close ask — the specific commitment the SDR should ask for at the end of this call, aligned to where the deal is in the qualification process.
The SDR sees this before picking up the phone — not a retrospective report but a prospective script.
v0.8 · P1 · Team
Weekly coaching digest via Resend
Monday morning email to both the SDR and their Team Leader. Sent via Resend using a template built in v0.8.
Contents for SDR: score trend (this week vs last week), top coaching moment from the week's analyses, 1:1 focus area (what to bring to this week's 1:1).
Contents for Team Leader: team score summary, top performer this week, SDR who needs the most attention (lowest score trend or highest flag count), recommended 1:1 prep focus areas per SDR.
Architecture decision required before build: exact digest format and template. Confirm with Colin (Ensemble design partner) before Resend is wired. Success target: 65%+ open rate on Monday morning digest.
v0.8 · P1 · Team
Leader Console restructure
The current Leader Console has too many tabs doing too many jobs. v0.8 restructures it into four clear sections:
Architecture decision required: confirm section names and which existing tabs go into each section before the restructure prompt is written. The current tab-to-section mapping needs to be defined first.
v0.8 · P1 · Team
1:1 Prep pre-loaded from coaching thread
The Team Leader opens 1:1 Prep for a specific SDR and finds it already populated with the week's most relevant coaching data — no manual prep required.
Pre-loaded content: the SDR's gate trends from the week's analyses, the coaching flags generated by the AI, any action items carried forward from the prior 1:1, and the recommended talking points based on what's most urgent for this SDR right now.
The leader can add their own notes and questions on top of the pre-loaded content. The 1:1 session becomes a 10-minute conversation grounded in actual data rather than a 30-minute prep session starting from scratch.
v0.8 · P2 · Pro+
Tone calibration learning
The system begins building a per-SDR archetype profile over time, rather than detecting the archetype fresh from each analysis. After 10+ analyses, the system has enough signal to identify a stable pattern — the SDR who consistently uses high-ego language doesn't need Socratic coaching re-detected on every call.
The learned archetype becomes the default but can still shift based on current-analysis signals. If a typically confident SDR sounds uncertain in a specific call, the system detects that and adjusts the coaching tone for that session.
This is P2 and explicitly out of scope for v0.7. Do not build early.
Open decision · Gates v0.7
Volley thread location
Two options:
Option A — Own route (/volley): The Volley gets a dedicated page in the app. The sidebar nav adds a Volleys item with an unread count badge. Clicking it opens the full Volley inbox. This is the cleaner UX and sets up the Connection Spine as a first-class platform feature — not buried in an existing tab.
Option B — Inside existing nav: The Volley lives inside the existing Handoffs tab as a sub-section. Lower build effort. But treats the Connection Spine as a feature of Handoffs rather than a platform-level concept.
Recommendation: Option A. The Connection Spine is the platform. Give it a home. But Kevin must confirm before v0.7 Prompt 1 is written — the route structure affects every subsequent prompt in the release.
Open decision · Gates v0.7 prompt 1
AE Lite token expiry
Two questions to answer:
1. Expiry duration: 7 days creates urgency — AEs respond quickly or the link dies. 30 days is more forgiving but stale Volleys accumulate. Recommendation: start with 30 days for design partners, tighten to 7 days once AE response behavior is understood.
2. Post-expiry behavior: Lock (AE gets an expired link error, must request a new Volley from SDR) vs Read-only (AE can view the brief but the action buttons are disabled). Read-only is safer for the SDR relationship — the AE isn't confused, they just can't act without a refresh.
Kevin and Colin must align on this before v0.7 Prompt 1. It affects the volley_tokens schema and the AE Lite page behavior.
Open decision · Gates v0.8 prompt 1
Multi-call grouping method
Three options for how the system links Call 2 to Call 1:
Option A — Account name match (automatic): When the SDR pastes new notes with the same company name as a prior analysis, the system automatically offers to blend with the prior call. Low friction. Risk: false positives if two different deals at the same company are treated as one thread.
Option B — SDR manually links: SDR selects "This is a follow-up to [prior analysis]" from a dropdown before running the new analysis. Explicit and accurate. Higher friction.
Option C — Follow-up queue item triggers link: When the SDR clicks Re-qualify on a follow-up item, the new analysis is automatically linked to the original. Most contextually correct — the follow-up queue is the natural home for multi-call tracking.
Recommendation: Option C. The follow-up queue is already the right mechanism. Kevin must confirm before v0.8 Prompt 1 — the call_sessions table schema differs significantly between options.
Open decision · Gates v0.8 prompt 1
Leader Console section names and tab allocation
The four proposed sections need confirmed names and confirmed tab lists before the restructure prompt is written. Current proposal:
Team Dashboard (looking down): SDR Performance cards, Team Pipeline, Coaching Flags, Handoff Quality, Leaderboard Managing Up (looking up): QBR generation, MBOs, SPIFF management, Badges & Rewards, QBR history Team Ops (operational): Team roster, SDR onboarding, Archive flows, PIPs Platform Config (configuration): Methodology, Guardrails, Knowledge Base, Branding, Billing, Integrations
Kevin and Jon must confirm or revise this mapping before the v0.8 restructure prompt fires. The current tab list must be fully mapped — nothing can be left unassigned.
Design partner · v0.6
Colin Schap — Ensemble
Active design partner. Credentials are live. Colin is the first real user of SDR Hive outside the founding team — his feedback on v0.6 directly shapes v0.7 scope before it's written.
What we need from Colin on v0.6:
Does the debrief card feel like coaching or a report card? (Qualitative — must be positive before v0.7 scoping)
Today's Calls adoption — is he using Build My Day daily?
Any regressions on existing features (analysis, handoff, PDF)
Target: design partner validation within 2 weeks of v0.6 ship.
Design partner · v0.7
Peter Webster — Slalom
Enthusiastic signal — Peter has expressed strong interest. He's the right design partner for v0.7 because Slalom's co-sell motion (Brand AE + Partner AE collaboration) is exactly the workflow the Volley and dirty laundry protection are designed for.
What we need from Peter on v0.7:
Volley workflow validation — does it match how Slalom actually decides to hand off deals to AEs?
Dirty laundry protection — does the manager visibility gate work as intended in practice?
AE Lite link experience — what's the AE response rate and how quickly do they act?
Handoff quality — acceptance rate on first Volley (target 50%+ accepted on first send)
Note: Peter independently asked about buying the DCN IP in an earlier conversation — flag as live open question alongside the design partner relationship.
Design partner · v0.8
PwC — TBD contact
PwC is next in the design partner sequence for v0.8. Contact TBD — the relationship needs to be established before v0.7 ships so the intro and onboarding can happen in parallel with v0.7 design partner validation with Slalom.
Why PwC for v0.8: the weekly digest, 1:1 prep pre-load, and Leader Console restructure are all manager-level features. PwC has large SDR teams with active Sales Leader oversight — exactly the use case these features are built for.
What we need from PwC on v0.8:
Manager digest adoption — 65%+ open rate on Monday morning email
1:1 prep pre-load usage — 70%+ of leaders using pre-loaded prep rather than building manually
Leader Console restructure feedback — NPS improvement vs old layout
