Home›Companies›Autostep
Autostep

The P&L for knowledge work

Autostep maps how work actually happens across a company, finds repetitive tasks and bottlenecks, and ranks them by cost and impact. Then it fixes them by building agents inside existing tools, changing the process, or finding the right software.
Active Founders
Aidan Pratt 🥑
Aidan Pratt 🥑
Founder
Aidan is the founder and CEO of Autostep. He studied Machine Learning and Computer Science at Georgia Tech and has been building and shipping products since he was a kid - hardware hacks, developer tools, AI agents - with dozens of projects launched along the way. Aidan has worked at seed to Series A startups with strong PMF. aidanpratt.com
Company Launches
Autostep: Measure the ROI of Knowledge Work
See original launch post

I'm Aidan, founder of Autostep. We are raising to enable the P&L of knowledge work (link below).


Most companies can tell you what they spend on AI. Almost none can tell you what came back.

Autostep is a desktop app that installs across a company and maps real work. It finds repeated tasks, bottlenecks, waiting, and admin loops, then ranks them by cost and impact.


From there, Autostep helps decide what should happen next: build an agent, change the process, standardize the workflow, or use software the company already pays for.


Product video: https://youtu.be/xxzvZKfVdKM

The problem we are solving:

  • The blocker for AI adoption is proof and discovery. Before rollout, teams need the workflow to target and the capacity returned after the fix.
  • They still struggle to answer the board-level question: Did this return results? Autostep answers that from the work itself.


How it works:

  1. Install once across the org. No integrations
  2. Discover the cost of human & agent tasks.
  3. Measure Human ROI: where people lose time and how much capacity can come back.
  4. Measure Agent ROI: which AI workflows earn their cost and which burn money.
  5. Act: remove the step, standardize it, route it to the right tool, or build the agent.

The long-term bet is that Autostep becomes the system of record for work: the layer a company uses to see capacity, measure AI payoff, and eventually pay people for working smarter. This is different than automating a bad process faster.


We are already working with companies that have the same pain: leaders want AI to work, and operators need the scorecard before they add one more workflow to manage.


Our ask:

  • If you know a COO, CFO, ops leader, or founder trying to prove whether AI is paying back, we would love an intro.
  • If you build agents or workflow automation products, let's talk. Autostep finds the work and can route teams to the right fix.
  • If you are an angel who believes every company will need an AI ROI layer, Aidan would love to compare notes.

Book a demo: https://calendly.com/aidan_/autostep

Investors: https://calendly.com/aidan_/p26

Example human tasks:

uploaded image

Recommendations:

uploaded image

Previous Launches
Welcome to Autostep - Uncover Your Repetitive Tasks Ready For AI and Automate Them
YC Photos
Autostep
Founded:2025
Batch:Spring 2026
Team Size:2
Status:
Active
Location:San Francisco
Primary Partner:David Lieb