Days 1–30
Shadow
A ranked map of automation opportunities, each with an estimate of hours saved.
The mission · The 90-Day Placement
A senior AI operator, embedded in your leadership team for 90 days.
The 90-Day Placement puts one senior AI operator in your leadership meetings. Thirty days to find where the hours go, thirty to prove the best automations on your real data, thirty to put them in production. Day 91 is your call.
What is an embedded AI operator?
An embedded AI operator is a senior person who knows AI and works from inside your leadership team, in the meetings where you decide where time and money go.
Most companies split the problem in two. The people who know where the hours go (your leadership) don't have time to learn AI. The people who know AI (vendors, consultants) don't know where the hours go. So AI pilots stall and very little reaches production.
The fix is one person who understands both, with a deadline: automations in production by day 90.
How does the 90-Day Placement work?
It runs in three 30-day phases: shadow, prove, ship.
Days 31–60
Prove
Working prototypes on your real data. You pick what ships.
Days 61–90
Ship
Automations in production, documented, with trained owners.
Days 1–30: shadow. The operator sits in your leadership meetings, follows the workflows, and tracks where hours and dollars go: the report rebuilt by hand every week, the data copied between two systems. Ranking by hours saved puts the place where AI pays back first at the top of the list.
Days 31–60: prove. The top opportunities become prototypes on your real data, not a demo account. Demo accounts have clean records and no exceptions. Yours don't. Ideas that die here die before anyone builds them into production.
Days 61–90: ship. What you picked goes to production. The operator documents it and trains the people who will own it, because an automation nobody understands gets switched off the first time it breaks. You keep everything we build.
Why the operator works inside your leadership team
An operator in the room sees what decides what gets automated: the complaint in a leadership meeting, the workaround nobody mentions. From outside, you get the version people describe in a kickoff interview. Inside, you see the version that runs.
Incentives matter too. Consultancies bill by the hour, so slow is profitable. Vendors sell seats, so the answer is always their product. We're paid for a fixed 90-day engagement, so the only way to look good is to ship. More in why not a consulting firm.
Previously, in 1882… Edison switched on Pearl Street Station, his first commercial power plant in the US. American factories didn't get the productivity jump until the 1920s, after they rebuilt the floor plan around electric motors.
Redrawing a floor plan is hard to do from outside the building.
What does the operator need from your team?
Mostly access: a seat in leadership meetings, your workflows and tools, real data, and someone who can say yes. The operator can only automate what they can see.
- A seat in leadership meetings. That's where priorities and the real cost of slow work come up.
- Access to workflows and tools. Watching the work beats reading the process doc.
- Real data. Prototypes run on it in days 31–60, so security and IT should clear access early.
- Someone who can say yes. At day 60, you decide what goes to production. That decision needs an owner.
- Owners for each automation. They get trained in days 61–90 and keep it running afterward.
Not ready for a full placement? Start with shorter AI consulting, like an AI roadmap ranked by return.
What happens on day 91?
You pick one of three paths: extend the placement, hire the operator full-time, or take it from here. Your call.
Option 1
Extend
If the map still has opportunities worth shipping, keep the operator working down the list.
Option 2
Hire the operator
They already know your meetings, your data, and what they built. Bring them on full-time.
Option 3
Take it from here
Your team has the documentation, the training, and everything we built.
Casting call
Who is the 90-Day Placement for?
Companies that want automations running, not a deck about them.
It fits US companies whose leadership will let an outsider into the room and act on what they find. Good fit if:
- You've bought AI licenses or run pilots, and very little reached production.
- At least one process eats your team's week.
- You'd like the option of hiring the person who built your automations.
Not a fit if:
- You want a strategy deck and nothing else. A big consulting firm will gladly sell you one.
- You need one answer, like which vendor to sign with. That's model and vendor selection.
- You need a permanent AI engineer now. Go straight to AI hiring.
- Nobody can grant access to real data. A prototype on a demo account proves nothing.
Mission briefing
Frequently asked questions
Is an embedded AI operator the same as a fractional AI lead?
Similar seat, different job. A fractional lead advises; our operator also builds the automations and is accountable for having them in production by day 90.
What kind of work gets automated?
Whatever ranks highest on the map from days 1–30. We don't arrive with a product to install, so the list comes from where your hours actually go.
Which AI models and tools does the operator use?
Whatever works best this month for each job. We're paid for the engagement, not for seats, so there's no product to push.
What if a prototype doesn't work?
Then it doesn't go to production. That's what days 31–60 are for: you decide what ships after seeing it run on your real data.
Where these facts come from
- Pearl Street Station, Thomas Edison's first commercial power plant in the United States, began generating electricity on September 4, 1882. Wikipedia: Pearl Street Station
- Paul A. David, "The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox," American Economic Review, May 1990. Paper (PDF)
- US productivity growth leapt in the 1920s, four decades after electricity was commercialized, once factories were redesigned around electric motors. Tim Harford, Financial Times (via History News Network)
Keep reading
AI Consulting
AI consulting, without the 90 days
AI consulting for US companies: an AI roadmap ranked by return, model and vendor selection, and build sprints where a senior team ships one automation.
Read itWhy AI pilots stall
Why AI pilots stall
Alexander Torrenegra on why most AI pilots never reach production, what 1920s factory electrification teaches leaders, and how to fix it.
Read itWhy not a consulting firm
Why not a consulting firm? Or the vendor?
AI consulting firm, AI vendor, or a senior operator inside your team? Compare who does the work, how each gets paid, and when each one is the right call.
Read itAnswer the call
Put an operator in your leadership meetings.
Bring one process that eats your team's week. We'll pick which part to start with, in a 90-Day Placement or a shorter consulting project.
Schedule a call