Service / Operations AI
Operations AI Engagement
For 50–500-person organizations where individual AI use is dramatic but organizational capability isn't compounding. A bounded engagement that ships at least one workflow actually running on AI — without committing to an enterprise transformation budget.
When AI stops compounding
Individual AI use isn't becoming organizational capability
The pattern is consistent. Individuals on your team are using AI tools, and the most disciplined of them are dramatically more productive than they were a year ago. None of that is shared. None of it is governed. The organization, measured at the level that matters to the business, is producing about the same output as it was a year ago.
At the same time, the AI offers in your inbox keep getting bigger. Enterprise transformation programs that don't fit your shape. Vendor platforms that promise a horizontal solution to a problem that's actually specific to how your team works. Slide decks where you'd rather have working artifacts.
This engagement is the in-between. Bounded, opinionated, executed by one person, sized for a single-quarter operating budget.
Buyer states we'll probably recognize together
- Individuals on your team are using AI tools and getting dramatic personal productivity gains. The organization is producing about the same output as a year ago.
- Three different people have built three different prompt libraries and three different workflows. None of them are shared. None of them survive when that person takes vacation.
- Leadership wants "AI strategy" but every concrete proposal turns into a vendor evaluation or a transformation slide deck. Nothing actually ships.
- You can name the top five workflows that are eating your team's time, but encoding them feels like a project nobody has the bandwidth to lead.
- You're between 50 and 500 people — too small for a McKinsey engagement, too large for ad-hoc. The shape of help you need doesn't exist on the market.
Good fit when
- You're 50–500 people. Big enough to have real internal process surface area; small enough that one focused engagement can actually move the needle.
- Leadership is committed to using AI inside the organization, not just buying licenses for it.
- There's at least one team that wants this and has time to engage. Engagements built around an unwilling team don't ship.
- You're prepared to assign an internal owner for the artifacts that get produced. Without an owner, encoded processes decay within a quarter.
What changes by the end
From scattered personal tooling to at least one shared, governed workflow
Before
Three different prompt libraries, no shared context, productivity locked inside individual contributors, and a leadership team that can't tell whether the AI investment is producing anything at the organizational level.
After
One or two workflows actually running on AI with multiple people on the team using them, a ranked adoption roadmap, named owners, and a 90-day continuation plan your team can execute without me.
What you receive
Three phases. One quarter of work, compressed.
Discovery and prioritization. A real working pilot. Governance and a 90-day continuation plan. By week six your team is shipping on its own.
-
01
Discovery & prioritization
Weeks 1–2
I work with you to map the workflows on the agreed surface area — usually one or two teams, sometimes a single function. Each workflow gets scored on AI leverage (how much time and inconsistency it costs today) and encode-ability (how cleanly it can be captured as a skill, given current context and tooling). The output is a ranked adoption roadmap with the next quarter's work clearly identified.
-
02
Pilot
Weeks 3–5
We pick one or two of the highest-priority workflows from the roadmap and actually ship them. I work with the person who runs each workflow best, capture how they actually do it, encode it as a skill, and put it into use with their team. The point is proof and pattern, not coverage. By the end of the pilot, more than one person on the team is producing better, more consistent output on a workflow that mattered.
-
03
Governance & handoff
Week 6
The pilot artifacts get owners, review cadences, and a place to live. The roadmap becomes a 90-day continuation plan your team can run without me. I close with a working session that walks the leadership team through what was built, what was learned, and what the next two quarters of work look like.
Methodology
The framework I use is published in full
The methodology behind this engagement is The Pragmatic AI Migration Playbook — eight chapters covering the maturity model, the four migration tracks (process encoding, knowledge architecture, governance, progressive automation), the compounding effect, and a 90-day execution plan. Read it before the discovery call. If your team can execute it on its own, you don't need this engagement.
Read the PlaybookNot a fit
When to skip this engagement
- You're under 25 people. The engagement is overkill at that scale; an AI Decision Review or some informal advice is usually a better fit.
- You're over 1,000 people with a real PMO and change-management apparatus. The shape of help you need is different from what this engagement provides.
- You want me to build a long-running custom AI platform. That's an implementation engagement, not an adoption one. I can refer you to partners.
- Leadership wants the engagement to produce a strategy deck without committing to actually changing how the work happens. That kind of work isn't useful, and I won't do it.
Common questions
What teams ask before booking
-
Why the explicit 50–500 framing?
Because the shape of help that's actually useful at that scale is different from enterprise transformation work, and almost nobody offers it. Enterprises have PMOs, change-management functions, and budgets to absorb 12-month engagements. Sub-50 organizations don't have enough internal process surface to justify a structured engagement. The 50–500 band is consistently underserved, and it's the band where the methodology fits naturally. -
How is this different from a Strategy Sprint?
A Strategy Sprint is product strategy — should we bet on this AI feature, and if so, how? An Operations AI Engagement is internal — how do we get our own organization compounding on AI? Different audience, different deliverable. Some companies need both, sequenced; most need one or the other. -
How is this different from "AI transformation" engagements I've seen pitched?
Three things. First, it's bounded — 4–6 weeks, fixed-price, no scope creep into a multi-quarter program. Second, it actually ships working artifacts, not slide decks. Third, it's done by me end-to-end, not sold by a partner and delivered by a junior team. The methodology is also different — it's the Pragmatic AI Migration Playbook, applied to your specific context, not a generic transformation framework. -
What does it cost?
Fixed-price, scoped on the discovery call. The price depends on the agreed surface area (one team vs. two), the number of workflows in scope for the pilot, and access logistics. Sized to fit a one-quarter operating budget for an SMB, not an annual transformation budget. -
What kinds of processes work for the pilot, and what kinds don't?
Process encoding works best where the work is repeatable, the criteria for "good" output are knowable, and someone on the team does it visibly better than others. Spec writing, code review, customer-issue triage, sales-call summarization, weekly reporting, and competitive analysis all tend to work well. Things that depend heavily on real-time judgment under unique constraints — incident response, executive negotiation, novel R&D — are usually not the right pilot candidates. -
What access do you need?
Time with two to four people from the team that owns the chosen workflows, including whichever person currently does the work best. Access to the systems and source materials those workflows touch. A shared place to put the artifacts (a repository, a Notion workspace, whatever you already use). I don't need administrative access to your stack. -
Do you write code as part of the pilot, or just advise?
I write code, configure tools, and ship working artifacts. The whole point of including a pilot phase is that you get more than a roadmap — you get one or two real automations actually running by the end. The handoff is to your team, who own and extend them from there. -
Does the engagement lock us into a specific AI vendor or platform?
No. The methodology is platform-agnostic. We use whatever your team is already on (Claude, Cursor, ChatGPT, Copilot, Gemini, etc.) and the artifacts produced are portable across them. If your stack changes in a year, the encoded processes and compiled context move with you. -
Can the pilot extend if we want to ship more?
Yes. The default engagement is bounded so the team gets a clean handoff and can continue independently. If after the pilot you want me to come back for a focused extension — another two to three weeks on a specific additional workflow, for example — that's a clean follow-on engagement, scoped separately. -
Do you sign NDAs?
Yes — happy to sign yours, or use a simple mutual NDA if you don't have a standard.
Start the conversation
Tell me about your team
A few sentences on your org size, the team you'd want the engagement focused on, and the one or two workflows you'd most want to encode. I'll respond personally within a day or two.