ethos
← Back to Projects
🗣️

Standup

Async standups without the noise

Thesis

Async standups are noisy or ignored. AI can synthesize updates and surface only what matters.

The Problem

Daily standups are broken: Sync standups interrupt deep work. Async standups become walls of text nobody reads. Blockers get buried. Half the team doesn't update consistently. You want to know: what's happening, who's stuck, what needs attention? You don't want to read 15 updates to find one blocker.

Implementation Approaches

Approach 1

Smart Digest (Recommended)

Recommended

AI synthesizes async updates into actionable digest

⏱️1-2 weeks
📊Medium complexity

Implementation

  • →Team submits updates (Slack, form, or bot prompt)
  • →AI synthesizes: key progress, blockers, attention needed
  • →Personalized digest: show me what I care about
  • →Highlight: related work, dependencies, conflicts
  • →Skip the noise, surface the signal

Pros

  • +Solves the 'wall of text' problem
  • +Personalized, everyone sees what matters to them
  • +Catches blockers and dependencies automatically
  • +Flywheel: learns what's important per team

Cons

  • −Still requires people to submit updates
  • −Synthesis quality depends on update quality
  • −May miss nuance in brief updates
Approach 2

Activity-Based Status

Generate status from tool activity, no manual input

⏱️2 weeks
📊High complexity

Implementation

  • →Connect: GitHub, Jira, Slack, Calendar
  • →AI infers: what did each person work on?
  • →Synthesize into team digest
  • →Highlight: stalled work, missing activity

Pros

  • +Zero input required from team
  • +Based on actual work, not self-reports
  • +Catches silent blockers (no activity = problem)
  • +More accurate than manual updates

Cons

  • −Lots of integrations to build
  • −May feel like surveillance
  • −Misses work not in connected tools
Approach 3

Intelligent Prompts

AI asks better standup questions based on context

⏱️1 week
📊Low complexity

Implementation

  • →Instead of generic 'what did you do?'
  • →AI asks: 'How's the payment integration going?'
  • →Based on: tickets assigned, recent commits, calendar
  • →Gets more useful updates with less effort

Pros

  • +Better input = better output
  • +Shows team their work is visible
  • +Reduces cognitive load of 'what should I report?'
  • +Works with existing standup tools

Cons

  • −Still requires manual response
  • −Prompt quality depends on data access
  • −Incremental improvement, not transformation

Validation Plan

Hypothesis to Test

Teams will pay $8/user/mo for AI-synthesized async standups that surface blockers

Validation Phases

1

Manual Synthesis

1 week
  • •Get 5 days of async standup updates from 2 teams
  • •Manually synthesize into daily digests using Claude
  • •Show to team leads: 'Is this useful?'
  • •Compare: time to read digest vs all updates
✓Digest preferred over reading all updates, catches blockers
2

Slack Bot MVP

2 weeks
  • •Build Slack bot for standup collection
  • •Auto-synthesize into digest
  • •Test with 3 teams for 2 weeks
  • •Measure: engagement, blocker detection, usefulness
✓80%+ update rate, leads find digests actionable
3

Full Product Launch

2 weeks
  • •Add personalization (what I care about)
  • •Add activity-based signals
  • •Launch publicly with $8/user/mo pricing
  • •Track: signup, retention, feature usage
✓50+ users, 70%+ weekly retention

Kill Criteria

Stop and move on if any of these become true:

  • ✕Teams don't submit updates consistently
  • ✕Synthesis misses important blockers
  • ✕Digest becomes another thing to ignore
  • ✕Competitive tools already do this well