For support and onboarding teams
Keep your support team on the cutting edge of AI.
We test new AI on your own tickets and onboarding work, set up what wins in your help desk, and train each person on it. When better AI ships, we do it again.
See it in 2 minutes
How Orcapo works, start to finish.
Why teams fall behind
AI gets better every month. Your team’s tools stay the same.
01
Switching is a project.
Every new tool means rebuilding automations and changing how people work.
02
Everyone has to relearn.
Each switch means training people again, so most teams never make it.
03
Hype is hard to spot.
New AI ships every week. Nobody has time to test what is real on your own work.
Where we start
The work we test first on teams like yours.
We confirm yours in the first round.
01
First replies to tickets
Today
Written from memory or copied from old replies.
After a round with Orcapo
Drafts from your help center, checked and sent by the agent.
02
Sorting and tagging tickets
Today
Read and routed one by one.
After a round with Orcapo
Sorted and tagged first, with the agent checking the edge cases.
03
Onboarding docs and training
Today
Rewritten for each new customer.
After a round with Orcapo
Drafted from your own docs and past onboardings.
How it works
We work in rounds. Each one keeps your team on the best AI.
- 01Map
- 02Test
- 03Roadmap
- 04Set up
- 05Train
- 06Next round
01 · Map
We learn how your team works today.
It starts with one email. You connect your tools as read-only, and each person fills out a 5-minute form. We map every role and find where the time goes.
Company
An example support team, 10 people
Tools you pay for
Your rules
Approved tools only
No customer data in public AI
Roles
Time on repeated tasks
People
AI experience, person by person
Support agent
New to AI · 60 tickets a week
Support agent
Uses ChatGPT daily · handles billing
From a 5-minute form each person fills out.
02 · Test
We test new AI on your own past work.
Every week we go through new AI tools, models and research. We run the best options on your past work, in a safe copy of your tools. An option passes only if it beats how you work today, and your team picks the better result without knowing which tool made it.
Test · First replies to tickets
50 past examples- Baseline
Today’s way
How your team does it now
- Not better
Option A
A new AI agent
- Beat today
Option B
A newer model, set up for you
- Too many errors
Option C
A startup’s new tool
Blind vote
Your team picked Option B 9 times out of 12, without knowing which tool made each result.
03 · Roadmap
You approve every change first.
The winners go into a roadmap. Each change shows the test it won. Nothing goes live until you say yes.
Roadmap · An example support team, 10 people
Waiting for your approvalThe main idea: each change won its test on your own past work. Nothing starts until you approve it.
- Weeks 1 and 2 Support agents: First replies ClaudeYour help desk
- Weeks 3 and 4 Onboarding: Onboarding docs ChatGPTNotion AI
- Weeks 5 and 6 Support lead: Sorting the queue ClaudeZapier
- Weeks 7 and 8Automations you approve
Needs your OK
Which tools may see customer data.
Which AI seats to add, and which nobody uses.
04 · Set up
We set it up in the tools you already use.
The automations and agents run in your own accounts. A person checks the work before it reaches a customer.
Set up from what won
Sorts and tags new tickets
Runs in your own help desk and Zapier
Drafts first replies from your help center
Drafts only, the agent sends
Drafts onboarding docs for new customers
Runs in your own docs
Approvals
Before anything runs
The owner approves anything that touches customer data.
05 · Train
Each person learns on their own work.
Each person gets a short guide made from their own work. Then they practice on a safe copy of their tools, with an AI coach that answers questions. It is not a generic course.
Practice lab
Sample data- Sample ticket: password reset
- Sample ticket: refund and a bug report
- Sample ticket: export not working
Draft the first reply from the help center.
Paste the ticket into your reply setup.
Done. The reply covers the refund.
Read the ticket again. What else does the customer report?
Two issues in one ticket: answer both, and log the bug.
Live coach · next to Your help desk
OnStep 2 of 3: draft the reply
Can I promise a fix date?
No. Link the status page and say you will update them.
Show me onceor do it yourself
06 · Next round
When better AI ships, the next round starts.
Every month we flag new AI worth testing, and you pick up to three. If one wins its test on your work, it replaces the old tool. If nothing wins, nothing changes.
Monthly check-in
New AI worth testing for your team
Using this round’s change
Pick up to 3 to test
- A newer model for first replies
- A new tool for onboarding docs
- An agent for sorting the queue
If one wins its test, it replaces the old tool. Nothing changes until you say yes.
Before and after
What changes for your team.
Before
Tools picked years ago, and never revisited
AI seats that few people use
Guessing which new tools are real
Training that is not about their work
After
The tool that won on your own work, for each task
Set up in the tools you already use
Each person trained on their own work
A new round whenever better AI ships
Who we are
Two Berkeley founders who build with AI every day.
Google and MIT
Our CTO is first author on AI research with both.
Databricks
Our CEO was an AI and machine learning consultant there.
Berkeley SkyDeck
Supported by Berkeley SkyDeck, Founders Inc. and UC LAUNCH.
Abhiraam Eranti
Co-founder and CTO
First author on AI research with Google and MIT. Technical lead on a U.S. Treasury app for 100,000+ blind and low-vision Americans.
Waleed Chaudhry
Co-founder and CEO
AI and machine learning consultant at Databricks. UC Berkeley Haas School of Business.
We also built LessonOrca, an AI tutoring company that writes a plan for each student and rewrites it after every session. Orcapo uses the same system for teams at work.
Questions
Questions
Is this a course?
No. We find, test and set up the tools, then train each person on their own work, in the tools they already use. There is no course library.
Why not just buy ChatGPT or Claude seats?
Seats give your team a tool. Nobody picks the tasks, tests the results or checks that people use it. An AI vendor will also recommend its own tools. We recommend whatever wins on your own work, even when that is a rival’s tool.
How often do our tools change?
Only when something wins its test on your work and you approve it. Most months, little changes. We keep the screens your team knows wherever we can.
What happens to our data?
You connect your tools as read-only, and we test new AI in a safe copy, so nothing real can break. Automations run in your own accounts, and you keep everything we set up.
Who approves changes?
You do. Every change shows the test it won, and nothing goes live until you say yes.
How much time does it take each person?
A 5-minute form at the start, a short vote on test results, and about 15 minutes of training for each new tool.
What happens on the audit call?
A short video call. We ask how your team works today and show you where better AI would save the most time. There is nothing to sign.
Our promise
We only recommend a tool if it beats your current one on your own work.
Nothing changes until you say yes.
Book a free AI audit.
Pick a time