PRD Template for Chat

Example: "Our users currently switch to Slack or email to discuss work happening in our product, creating context fragmentation. In interviews with 30 power users, 87% said they switch between our product and a messaging tool more than 10 times per day. Context from these conversations (decisions, clarifications, feedback) is lost because it lives outside the product. 62% of users reported making decisions based on stale information because the relevant discussion happened in a Slack channel they were not in."

Example: "Primary: 60% of product-related discussions happen within the product (currently 0%) within 6 months of launch. Secondary: Reduce average context-switches per session from 10+ to under 3. Traceability: Every decision discussed in chat is linked to the artifact it affects. Latency: Messages delivered in under 200ms."

Example: "As a team member, I want to start a threaded discussion on a specific paragraph of a document so that the conversation stays attached to the content it references. Acceptance criteria: threads can be started by selecting any text in a document; the thread appears in a side panel alongside the document; all participants are notified of new replies; threads are searchable and visible in the activity feed."

Example: "FR-1: Messages must be delivered in real-time via WebSocket with a REST fallback for reconnection. FR-2: Messages must support rich text formatting (bold, italic, links, code blocks) and @mentions. FR-3: Users must be able to share files up to 25MB within a conversation. FR-4: Presence indicators must show online/offline/away status updated within 30 seconds. FR-5: Message history must be persistent and paginated (load 50 messages at a time, scroll up for more). FR-6: Users must be able to edit messages within 15 minutes and delete their own messages at any time."

Example: "NFR-1: Message delivery latency must be under 200ms at p95 for users in the same region. NFR-2: The chat system must maintain 99.9% uptime. NFR-3: Zero message loss — every message sent must be persisted and eventually delivered, even if the recipient is offline. NFR-4: Chat must support 1,000 concurrent WebSocket connections per server instance. NFR-5: Message search must return results within 500ms for conversations with up to 100,000 messages."

Example: "Metric 1: Percentage of product-related discussions in-product. Baseline: 0%. Target: 60%. Metric 2: Daily active chat users / daily active product users. Target: 50%+. Metric 3: Message delivery success rate. Target: 99.99%. Metric 4: Average response time in threads. Target: under 30 minutes during business hours. Metric 5: Context switches per session. Baseline: 10+. Target: under 3."

Example: "Real-time messaging will use Socket.io with Redis Adapter for horizontal scaling across multiple server instances. Messages are persisted in PostgreSQL with a compound index on (conversation_id, created_at) for efficient pagination. Presence is tracked via Redis with a 30-second heartbeat. Offline messages are delivered on reconnection by querying messages newer than the client's last-seen timestamp. File attachments are uploaded to S3 and referenced by URL in the message payload."

Example: "Risk: Users do not adopt in-product chat because Slack is deeply embedded in their workflow. Likelihood: High. Impact: High. Mitigation: Start with contextual comments (attached to documents and tickets), not general-purpose chat. This provides unique value that Slack cannot replicate. Risk: Message loss during WebSocket reconnection. Likelihood: Medium. Impact: High. Mitigation: Implement client-side message queue with server-side delivery confirmation. Unconfirmed messages are retried on reconnection."