Write a PRD

PRD: AI-Generated Meeting Summaries with Decisions and Action Items

Audience: Engineering and Design leads Status: Ready for build planning Owner: Product

1. Summary

Huddle records and transcribes meetings with speaker labels, and the attendee list comes from the calendar invite. Today, people write their own notes after calls, and those notes are often late, incomplete, or missing. This feature generates a summary after each meeting that lists the decisions made and assigns action items to attendees. Action items are emailed to their owners, with Sales' requirement for immediate delivery balanced against the accuracy limits described in Section 4.

2. Goals and Non-Goals

Goals - Produce a summary within 5 minutes of meeting end for 95% of meetings. - Every action item names an owner who was present in the meeting. - Every decision and action item links to the transcript moment that supports it. - Give Sales automatic, immediate delivery for high-confidence action items.

Non-Goals (v1) - Calendar write-back or task creation in external tools (Jira, Asana). Design for it, but do not build it. - Summaries for meetings without a transcript. - Sentiment, coaching, or participation analytics.

3. Definitions

• Attendee: A person present in the meeting, determined by the participant session log (the platform's record of who joined), not only by the calendar invite. The calendar invite is a hint, not proof of attendance.
• Identified participant: A participant tied to a Huddle account or a verified email address.
• Unidentified participant: A guest who joined by name only, with no verified identity.
• Action item: A task with an owner, a description, and optionally a due date, derived from transcript content.
• Decision: A statement that the group agreed on a course of action, with supporting transcript evidence.

4. Known Limits and Design Responses

5. Functional Requirements

FR-1: Roster resolution Build the attendee roster from the participant session log at meeting end. Merge in calendar invitees for context, but do not treat invitees who never joined as attendees. Each roster entry is tagged as identified or unidentified.

FR-2: Summary generation Produce a summary containing: - A 3–5 sentence overview. - A decisions list. Each decision has a one-sentence statement, a timestamp, and a transcript link. Decisions are not assigned to individuals unless the transcript clearly states who made them. - An action items list (FR-3).

FR-3: Action item extraction Each action item includes: description, owner (a roster member), due date if stated, timestamp, transcript link, and attribution confidence (high, medium, low).

FR-4: Hard constraint: no non-attendee owners (must-pass) Before any summary is stored or sent, a deterministic validator checks that every action item owner matches a roster entry from FR-1. If an owner does not match, the item is removed and logged as a pipeline error. This check does not depend on the model or on confidence scores. The model is never the final authority on membership.

FR-5: Unidentified owners If the model attributes an action to an unidentified participant, the item is kept, but the owner is shown as "Unidentified participant (Name as joined)." These items are never emailed. They appear in the organizer's review view only.

FR-6: Confidence gating for delivery Each action item is sent automatically only if all of the following are true: - Owner is an identified roster member. - Attribution confidence is high. - The supporting transcript segment has no overlapping speech flag. - The item's source is not a multi-person room device.

Items that fail any check go to the organizer's review queue. The organizer can confirm, reassign to another roster member, or dismiss. Confirmed items are then sent.

FR-7: Multi-person sources When an action item comes from a multi-person source, the owner field is set to "Confirm owner" and the item is held for review. The review UI lets the organizer or the item's likely owner claim it from the roster.

FR-8: Delivery - Auto-sent items go to the owner's email within 5 minutes of meeting end. - Emails contain: the meeting name, the action item, the due date if any, a link to the summary, and a link to the transcript moment. - Each meeting produces one email per owner, grouping that owner's items. - Emails go only to identified roster members. Admins can disable delivery to external email domains.

FR-9: Review queue Organizers see a summary page with held items, can edit or confirm them, and can see which items were auto-sent. Edits made after sending generate a correction email to the owner.

FR-10: Controls and privacy - Admins can turn summaries on or off per workspace and per meeting type. - Any participant can opt out of their own speech being summarized. Their items are excluded from the summary and their name is removed from the roster. - Summaries and transcripts follow the same retention and access rules as the source recording.

6. Sales Requirement: Immediate Auto-Send

Sales wants action items emailed as soon as the meeting ends. This PRD supports that for high-confidence items (FR-6, FR-8). It does not send low-confidence items immediately, because a wrong owner gets a task they didn't agree to, and an email that is later corrected erodes trust. Sales can see the auto-send rate in the admin dashboard. If the held rate is too high for their workflows, the threshold can be tuned with data from the pilot. It should not be lowered without evaluation.

7. Acceptance Criteria

• AC-1: Across a 300-meeting eval set, zero action items name an owner absent from the participant log. This is a hard gate for launch.
• AC-2: Every decision and action item has a working transcript link that lands within 10 seconds of the cited statement.
• AC-3: For a labeled eval set, owner precision on auto-sent items is at least 97%.
• AC-4: Summaries are generated within 5 minutes of meeting end for 95% of meetings with transcripts under 90 minutes.
• AC-5: An uninvited identified joiner appears in the roster and can receive items. An unidentified joiner never receives email.
• AC-6: Opt-out removes the participant's items and name from the output within one processing cycle.

8. Metrics

• Adoption: Percentage of eligible meetings with a summary delivered.
• Quality: Action items edited or reassigned by the organizer, by confidence band.
• Trust: Post-send corrections per 100 auto-sent items. Target: under 2.
• Usage: Percentage of owners who open the linked transcript from an email.

9. Open Questions

1. Should the review queue notify the organizer by email, Slack, or in-app only?
2. When an owner is an external email address, what is the policy under the admin domain setting?
3. How will we handle meetings where the transcript contains no clear owner for a decision or action?
4. Do we need a per-workspace setting to require organizer approval for all items?
5. Should multi-person room devices be fixed at the hardware or SDK level, which would change FR-7?

10. Phasing

• Phase 1 (pilot, 20 workspaces): FR-1 through FR-5, FR-8 through FR-10, with all items held for organizer review. This validates the constraint and attribution quality.
• Phase 2: Enable confidence gating (FR-6) and auto-send for Sales and other pilot workspaces, based on Phase 1 precision data.
• Phase 3: Multi-person source handling at scale (FR-7), and general availability.

Product Requirements Document (PRD): AI-Generated Meeting Summaries & Action Items

1. Document Overview & Objective

1.1 Product Name

Huddle Intelligent Summaries (Internal Codename: Project Cue)

1.2 Target Audience

Engineering Leads, Frontend/Backend Engineers, and Product Designers.

1.3 Objective

Build an automated, post-meeting intelligence feature for Huddle that processes meeting transcripts to extract structured decisions, generate comprehensive summaries, and assign action items to attendees. The feature must account for real-world messy transcript data (speaker errors, uninvited participants) and adhere to strict safety boundaries regarding action item assignments and automated distribution.

1.4 Background & Scenario

Huddle is a video-meeting application currently utilized by 3,000 enterprise and mid-market companies. * Existing Infrastructure: Meetings are automatically recorded and transcribed with basic speaker labeling. Attendee lists are reliably synchronized from the originating calendar invite. * Current State & Pain Point: Users currently rely on manual post-meeting note-taking, leading to dropped tasks, misaligned follow-ups, and fragmented alignment. * Value Proposition: Automating meeting wrap-ups directly inside Huddle saves hours of manual administrative overhead per week⁴, driving higher platform stickiness and daily active use.

---

2. Core User Stories & Personas

2.1 Personas

• The Engineering Manager (User): Attends 5–6 back-to-back technical syncs daily. Needs clear, unequivocal records of architectural decisions and precise engineering follow-ups without digging through raw, messy transcripts.
• The Account Executive / Sales Rep (User): Operates on rapid customer cadences. Requires immediate, zero-friction distribution of post-meeting action items to stakeholders without manual typing.
• The Engineering & Design Leads (Builders): Require explicit edge-case handling, system constraints, deterministic UI layouts, and scalable pipeline specs to build and ship the feature efficiently.

2.2 Key User Stories

1. As a meeting participant, I want Huddle to automatically generate a structured summary and action item list the moment a meeting concludes, so I don't have to spend time writing manual recap notes.
2. As an attendee with assigned tasks, I want clear ownership labels on my action items so there is never any ambiguity about who is driving a deliverable.
3. As a sales representative, I want my action items emailed to me automatically as soon as the meeting ends, so I can immediately execute client follow-ups.
4. As a meeting host, I want to review, edit, or override incorrect speaker attributions or bad action assignments before summaries are locked or shared externally.

---

3. Detailed Functional Requirements

3.1 Data Inputs & Pre-Processing Pipeline

The AI summarization engine triggers asynchronously via webhook immediately upon meeting termination (`recording.ended`).

• Input A: Transcript with Speaker Labels: Full text payload containing timestamped utterances, text strings, and speaker IDs. Known limitation: Speaker labels suffer from an 8% error rate, primarily caused by cross-talk, overlapping dialogue, or multi-person huddle rooms.
• Input B: Calendar Invite Attendee List: Ground-truth list of officially invited participants (Email, Name, Role).¹ Known limitation: 15% of meetings include an uninvited participant (e.g., ad-hoc joiners via shared link).
• Input C: Meeting Metadata: Duration, title, organization ID, and recording URL.

3.2 Summarization & Extraction Logic (LLM Pipeline)

The backend processing pipeline must execute three logical passes: 1. Meeting Overview Generation: Condense the transcript into a 3–5 sentence executive summary alongside 3 key thematic highlights. 2. Decision Extraction: Isolate explicit agreements, architectural choices, or policy resolutions made during the call. Format as an bulleted list of declarative statements (e.g., "Decided to migrate the primary auth service to OAuth2 by Q3"). 3. Action Item Extraction & Assignment Engine: * Extract explicit tasks, owners, and deadlines. * Crucial Constraint (Safety Guardrail): The engine must never assign an action item to someone who was not present in the meeting. * Resolution Strategy for Uninvited Attendees: If an uninvited participant (who joined without a calendar invite) is identified as the owner of a task, the system must cross-reference their presence against the audio/transcript participant logs. If verified present, they are added to a dynamic "Extended Participants" list for assignment validation. If their identity cannot be verified or they were absent, the task defaults to unassigned or falls back to the meeting host with a UI warning flag².

3.3 Sales Request: Automated Instant Emailing

• Requirement: Sales teams have requested that action items be emailed automatically to their owners the second a meeting ends.
• Product Specification & Guardrail: Due to the 8% speaker label error rate and potential hallucination risks, fully autonomous external emailing on meeting close creates high business risk.
• Implementation Design:
• Implement an Opt-In Workspace Policy configured by organization admins.
• For workspaces with "Auto-Send Sales Recaps" enabled, summaries and assigned action items are dispatched via email to internal organization members immediately upon processing completion.³
• For external participants or non-opt-in workspaces, a 10-minute "Grace Period / Review Window" notification banner appears in the Huddle desktop client, allowing hosts to review and edit assignments before dispatch.

---

4. User Experience & Interface (UX/UI) Specifications

4.1 Post-Meeting View (Huddle Web & Desktop App)

Upon meeting conclusion, a new tab labeled "AI Summary" appears alongside the existing "Transcript" and "Recording" tabs within the meeting details drawer.

``` +-------------------------------------------------------------------+ | Huddle Meeting: Q3 Architecture Sync [Share] [Export]| +-------------------------------------------------------------------+ | Tabs: [Overview] [Transcript] [Recording] | |-------------------------------------------------------------------| | EXECUTIVE SUMMARY | | The team reviewed the Q3 scaling bottlenecks. Consensus was | | reached on moving auth infrastructure to OAuth2. Database latency | | issues will be mitigated via Redis caching layers. | | | | KEY DECISIONS | | • Migrate core auth service to OAuth2 by August 15. [Edit] | | • Adopt Redis cluster for session state caching. [Edit] | | | | ACTION ITEMS | | [x] Investigate OAuth2 token expiry edge cases | | Owner: Sarah Jenkins (Calendar Invite) | Due: Next Tuesday | | [ ] Set up staging Redis cluster | | Owner: Unassigned ⚠️ (Verify Attendee) | Due: Friday | | | | [Send Sales Recap Email Now] [Approve & Lock] | +-------------------------------------------------------------------+ ```

4.2 Interactive States & Edge-Case UI Handling

• Unassigned / Unverified Owner State: If the model detects an action item owner who was an uninvited participant or has an ambiguous speaker label, render a yellow warning badge: `⚠️ Unverified Owner — Click to Reassign`.
• Inline Editing: Every decision and action item card features an inline pencil icon allowing hosts or assignees to edit text, change owners via a searchable dropdown of confirmed attendees, or delete hallucinated items.

---

5. Technical Architecture & Engineering Guidelines

5.1 Pipeline Flow & Data Schema

1. Webhook Listener: Receives `meeting.ended` event containing metadata and storage URIs.
2. Context Aggregator: Merges calendar invite list (`invitedusers[]`) with actual audio track speaker diarization outputs (`detectedspeakers[]`).
3. LLM Structured Extraction: Prompt engineering optimized for JSON output schema enforcement (using strict JSON mode / function calling):

```json { "meetingid": "string", "summary": "string", "decisions": ["string"], "actionitems": [ { "task": "string", "assigneeemail": "string | null", "assigneename": "string", "isverifiedattendee": boolean, "due_date": "ISO-8601 | null" } ] } ```

5.2 Error Handling & Fallbacks

• Speaker Label Inaccuracy (8% Error Rate): Mitigation via UI feedback loops. If an assignee name has low confidence score (< 0.85) from the diarization model, flag the action item with a review indicator in the UI.
• Uninvited Participant Handling (15% Occurrences): The validation layer must query the roster of active audio streams. If an email is missing from the calendar invite but present in audio metadata, tag them as `guest_participant` and permit assignment only if their voiceprint or explicit self-identification ("Hi, this is Dave from Finance joining late") is logged in the transcript text. Otherwise, fail-safe to unassigned.

---

6. Non-Functional Requirements & Constraints

• Latency: Summary generation and database write must complete within < 120 seconds of meeting termination for a standard 45-minute meeting.
• Security & Privacy: Meeting transcripts and summaries must be encrypted at rest (AES-256) and in transit (TLS 1.3). Customer data must never be used to train foundational models across tenant boundaries.
• Reliability: 99.9% pipeline uptime. In the event of an LLM timeout or parsing failure, the system must gracefully fall back to displaying raw transcripts with an error banner stating "AI summary generation delayed. Click to retry."

---

7. Metrics & Success Criteria (KPIs)

1. Adoption Rate: % of completed meetings where users view, edit, or export the AI summary (Target: > 40% within 30 days post-launch).
2. Action Item Completion Rate: Increase in marked-complete action items compared to legacy manual workflows.
3. Accuracy & Correction Rate: Frequency of manual edits made to action item owners and speaker assignments (Target: < 5% manual override rate, validating low false-positive rates for uninvited attendee checks).
4. Sales Workflow Efficiency: Reduction in post-call administrative turnaround time reported by enterprise sales teams using the automated email dispatch feature.

Product Requirements Document (PRD)

Feature: Automated Meeting Summaries, Decisions, and Action Item Attribution Product: Huddle Video Conferencing Target Audience: Engineering Leads, Design Leads, Product Ops Status: Ready for Review Target Release: Q3

---

1. Executive Summary & Objective

Today, Huddle’s 3,000 corporate customers rely on manual note-taking. This results in lost context, untracked deliverables, and administrative drag.

This project introduces an AI-powered pipeline that automatically generates: 1. A concise meeting summary. 2. A structured list of key decisions. 3. Assigned action items mapped strictly to verified attendees.

Key Objectives

• Reduce post-meeting admin time by 80%.
• Achieve $\ge 90\%$ user acceptance rate on action item assignments (measured by lack of manual re-assignment or deletion).
• Hard Rule Compliance: 0% assignment of action items to non-attendees.

---

2. Participant Verification & System Constraints

Our models must handle two structural data flaws: 1. Diarization/Labeling Error Rate: ~8% of transcript lines have inaccurate speaker tags (common during crosstalk or multi-person hardware rooms). 2. Ad-Hoc Joiners: ~15% of meetings feature participants who were not on the original calendar invite.

2.1 The "Eligible Assignee" Rule Engine (Non-Negotiable)

``` [Calendar Invite] [Client Telemetry Logs] \ / \ / v v +----------------------------------------+ | Participant Reconciliation Engine (PRE)| +----------------------------------------+ | v [Eligible Assignee Roster (EAR)] - Authenticated Users (User ID + Email) - Verified Guests (Session ID + Display Name) | v +----------------------------------------+ | LLM Extraction & Assignment Pipeline | +----------------------------------------+ | +--> Action assigned to User in EAR? --> ACCEPT | +--> Action assigned to External Party? | v Route to In-Room Proxy (or flag as Unassigned Dependency) ```

Constraint Definition: An action item can only be assigned to an entity present in the Eligible Assignee Roster (EAR) for that specific meeting session.

#### Step 1: Ingest Telemetry at Call Termination (`call.ended` event) Do not rely on the calendar invite as the source of truth for presence. Telemetry establishes physical attendance: * Calendar Attendees ($P{cal}$): Used only as an identity-matching fallback. * Connected Clients ($P{conn}$): Every participant with connection duration $> 60\text{ seconds}$ in the meeting telemetry logs. * Eligible Assignee Roster ($EAR$): Defined strictly as $EAR = P_{conn}$.

#### Step 2: Resolving the 15% Uninvited Attendees * Authenticated Joiners: If an uninvited joiner is logged into a Huddle account, resolve their `userid`, display name, and email directly into the $EAR$. * Guest Joiners (Unauthenticated): If a user joins via guest link without an account: * Capture their client display name and the session ID. * Tag them in the $EAR$ as `Guest: [Display Name]` with `userid = null`. * They are eligible for assignment in the UI, but will require an email address input if notifications are to be sent.

#### Step 3: Mitigating the 8% Diarization Noise Diarization errors mean the speaker label alone cannot be trusted for commitments (e.g., if Bob says "I'll take that," but the audio engine tags Alice). * The LLM prompt must cross-reference explicit verbal attribution ("Jane, can you take that?" $\rightarrow$ "Sure") against semantic context rather than relying exclusively on transcript speaker IDs. * When semantic confidence falls below threshold ($\tau < 0.75$), flag the action item as Unassigned rather than guessing.

#### Step 4: External Dependencies (Out-of-Room Third Parties) If a meeting participant says: "We need legal to review this, I'll have Sarah from Legal look at it," Sarah was not present and cannot be assigned the task. * Enforcement: The system assigns the task to the speaker (the in-room proxy) as: Assignee: Current Speaker (`[Speaker Name]`) Task: "Coordinate with Sarah (Legal) to review [X]." * If the speaker cannot be deduced with high confidence, set `assignee = UNASSIGNED` with a note: `External dependency: Sarah (Legal)`.

---

3. Product & Feature Specifications

3.1 LLM Processing Pipeline

Within 60 seconds of call termination, the processing service executes the following workflow:

``` [Call Ended] | v [Fetch Clean Transcript + Telemetry EAR] | v [LLM Context Construction] ├── System Prompt & Guardrails ├── Strict EAR List (Names + Emails + IDs) └── Diarized Transcript | v [Structured Output JSON Generation] ├── Executive Summary (Paragraph, <= 150 words) ├── Key Decisions (Bullet points) └── Action Items (Array: Task, Assignee ID, Due Date, Evidence Quote) | v [Deterministic Post-Processing Validator] ├── Cross-reference every Assignee ID against EAR └── Overwrite invalid assignments to "UNASSIGNED" | v [Summary Artifact Created & Persisted] ```

3.2 Action Item Post-Processing Validation Logic

Before writing to the database, run a deterministic verification function:

```python def validateactionitems(extractedactions, eligibleassigneeroster): validatedactions = [] for action in extractedactions: assigneeid = action.get("assigneeid") # Validation: Assignee must explicitly exist in the EAR if assigneeid and assigneeid in eligibleassigneeroster: action["status"] = "assigned" action["assignee"] = eligibleassigneeroster[assigneeid] else: # Fallback for hallucinated or external assignments action["status"] = "unassigned" action["assignee"] = None action["flag"] = "ASSIGNEENOTINROOM" validatedactions.append(action) return validated_actions ```

---

4. Sales Request Resolution: Auto-Email Notifications

The Conflict

Sales requested that action items be emailed automatically to assignees immediately when the meeting ends. However, because speaker labels fail ~8% of the time, immediately sending unreviewed emails will blast misattributed commitments to clients and internal executives, eroding trust.

The Solution: "Grace Period Auto-Dispatch"

We satisfy Sales' need for immediate execution while protecting data integrity through a time-delayed trigger and host bypass mechanism.

``` [Meeting Ends] │ ├─► AI Generates Artifacts (Target: <60 seconds) │ ├─► Start 5-Minute Grace Period │ │ │ ├─► Host edits/approves? ──► Send instantly with updates │ │ │ └─► Timer expires? ────────► Auto-dispatch without host intervention │ └─► (Workspace Setting Override: "Instant Send" with 0-min delay available) ```

1. Host Review Window (Default): * A 5-minute countdown starts the moment the summary finishes processing. * The meeting host gets a browser/desktop push notification: "Your Huddle summary is ready. Auto-sending in 5:00 minutes. Click to review." * If the host makes edits or clicks "Send Now", the countdown ends and emails dispatch immediately. * If the host takes no action, the emails dispatch automatically at the 5-minute mark. 2. Sales/Workspace Configuration Flag: * Enterprise Admins can toggle the grace window per team (e.g., Sales team can set Grace Period to `0 minutes` for true instant dispatch; default for other workspaces is `5 minutes`). 3. Safety Fallback: Action items marked as `UNASSIGNED` are omitted from individual assignment emails and instead bundled exclusively into the primary meeting recap sent to the host and organizer.

---

5. User Experience & Design Specs

5.1 Post-Meeting In-App Summary Tab

Located inside the Huddle desktop and web apps under the meeting historical detail view.

``` +-----------------------------------------------------------------------+ | Meeting Summary: Product Architecture Sync [Share] [Edit]| | Today at 10:00 AM • 42 mins • 6 Attendees | +-----------------------------------------------------------------------+ | EXECUTIVE SUMMARY | | The team finalized the data migration strategy for Q3. Key risks | | around API rate limiting were addressed by introducing Redis caching.| +-----------------------------------------------------------------------+ | DECISIONS MADE | | • Chose Redis over Memcached for distributed cache. | | • Deferred mobile client updates to Sprint 44. | +-----------------------------------------------------------------------+ | ACTION ITEMS Auto-sending in [ 04:12 ] [x]| | [ + Add Action Item]| | [x] Task: Implement Redis cache layer | | Assignee: [ (Avatar) David Chen v ] Due: [ Friday v ] | | Source: "David: I can set up the cluster by end of week." | | | | [ ] Task: Verify Legal compliance on data retention | | Assignee: [ Unassigned v ] Due: [ Set Date v ] | | ! Flag: External dependency (Sarah / Legal was not present) | +-----------------------------------------------------------------------+ ```

5.2 Key UI Components

1. Assignee Dropdown Menu: * Filtered exclusively to individuals in the $EAR$. * Section 1: Authenticated Attendees (1-click select). * Section 2: Ad-Hoc / Guest Attendees (Displays `[Guest Name] - prompt for email`). * Option to select "Unassigned". * Explicitly disallows searching the global company directory within the "Assignee" field to prevent violating the non-attendee constraint. 2. Context Tooltip ("Source"): * Hovering over an action item shows the transcript snippet that triggered the assignment, allowing the host to quickly verify ambiguous attributions caused by diarization errors.

---

6. Email Dispatch Specs (Sales Notification System)

6.1 Individual Assignee Notification

• To: Assignee email
• Reply-To: Host email
• Subject: `[Action Required] Action item from: {Meeting Title}`

``` Hi {Assignee First Name},

The following action item was assigned to you during {Meeting Title}:

Task: {Task Description} Due Date: {Due Date or "Not specified"} Context: "{Relevant transcript excerpt}"

View full meeting notes and recording: {Link to Huddle Meeting Record}

Did the AI get this wrong? Click here to reassign or remove: {One-Click Edit Link} ```

---

7. Telemetry, Edge Cases & Error Handling

---

8. Success Metrics

1. Attribution Precision: $\ge 98\%$ of assigned action items correctly reflect real commitments made by that person (measured via sampling and low manual-edit rates).
2. Zero-Hallucination Rate: $100\%$ compliance with the hard constraint: 0 automated emails dispatched to individuals outside the $EAR$.
3. Dispatch Velocity: 95th percentile of summary and action item deliveries completed within $< 90\text{ seconds}$ post-meeting (or exactly at the expiration of the chosen grace period).
4. Host Intervention Rate: $< 12\%$ of summaries require manual host corrections during the grace period window, indicating high trust in baseline accuracy.