PORTFOLIO
PORTOverview
In fast-moving consultancy environments, case studies are a persistent bottleneck. Valuable project knowledge lives in Slack threads, Drive folders, and people's heads — and rarely makes it into a reusable format before the team moves on.
Casey is the internal AI agent I designed and built to close that gap. One Slack command triggers a fully automated pipeline: Casey reads the channel history and Drive documents, extracts structured content via Claude, drafts a Notion case study page, and creates a ready-to-use slide — all without leaving Slack.
Role & Responsibilities
End-to-end pipeline
The hardest part wasn't the AI. It was designing a workflow people would actually finish.
Consultancy teams are fast-moving. Any process that requires opening a new tool, filling in a form, or switching context is one that quietly gets skipped. The brief was deceptively simple: make capturing a case study as close to zero effort as possible. That meant the interaction had to live entirely in Slack — the one place everyone already is — and the output had to land in the exact formats the team already uses.
Three design decisions that made it work
01 — Control
Casey drafts, the team approves. A single "Update slide" button lets anyone regenerate the slide after editing the Notion page, keeping control with the people who know the project best.
02 — Timing
A daily HubSpot trigger detects deals closing in the next 7 days and nudges the right Slack channel automatically. The case study habit starts before the project wraps, not months later.
03 — Audience
The Notion page captures the full story. The slide is strictly client-facing impact. Financial data never leaks onto a pitch deck.
Impact
0
Extra tools
Entire workflow runs inside Slack, Notion, and Google Slides. No new login, no new tab. Teams adopted it without a training session.
<2m
Draft to publish
A full case study — Notion page plus client slide — in under 2 minutes. Previously a manual effort that often never happened at all.
100%
Company rollout
Deployed across all internal project channels, public and private. Shipped and live in production within weeks of the first commit.
