I redesigned our project intake process after realizing that the original structured intake form wasn’t getting consistent adoption.

Rather than keep pushing a process people weren’t using, I built an AI-powered workflow that let the team submit requests the way they naturally communicated, through Slack messages or voice notes, while still giving the Project Manager a structured ClickUp task to review.

<aside> <img src="/icons/new-alert_green.svg" alt="/icons/new-alert_green.svg" width="40px" />

My role: Intake process redesign, automation design and build, SOPs, training, rollout, and follow-up

</aside>

The Problem

I had already built a standardized project intake form and process in ClickUp, but adoption was inconsistent. New work often came out of client meetings or quick leadership conversations, and people were more likely to send a Slack message or voice note than stop and complete a detailed intake form. Some requests never made it to the PM, and client projects could slip through the cracks before anyone formally kicked them off.

Note: We didn’t have a reliable count of missed requests because some work never entered the system, but we did see clients following up on deliverables that hadn’t been formally initiated.

Report on projects that were not delivered to the client on time. One from this date range never made it to ClickUp because of the current broken process.

How I Approached It

The original process asked for too much behavior change. I wanted to make submitting work easier without leaving the PM to sort through incomplete messages and guess what needed to happen next.

So I designed the process around how the team was already communicating. Someone could describe the request in Slack; AI would organize the information into a usable intake task, and the PM would review it before creating the full project. The goal was to lower the friction on the front end without removing human judgment on the back end.

Original structured Intake Form

What I Built

I created two intake paths:

1. Typed request: Slack → AI processing → structured ClickUp task
2. Voice request: Slack voice memo → transcription → AI processing → structured ClickUp task

AI was prompted to turn the request into a project summary, action items, suggested owners and timing where possible, and notes or assumptions for anything that still needed human review.

The PM reviewed the request, confirmed ownership and dates, fixed anything that needed context, and then created the full project in ClickUp.

Zapier Automation

Implementation

I wrote the SOPs, created the training plan, trained the team, rolled out the new intake process, and followed up after launch to make sure people understood how to use it and where the process was still creating friction.