The method
Five things that separate field-ready AI practitioners from everyone else.
This is not a maturity model or a readiness checklist. It is a set of hard-learned distinctions about what actually moves enterprise AI from conversation to working workflow.
Most AI engagements start with the wrong question.
"Where can we apply AI?" sounds reasonable. It isn't. That question produces a long list of possibilities and no decision. The right question is: where is work already leaking measurable value that someone is currently tolerating? Those two questions take you to completely different conversations — and completely different outcomes.
The accessible context assumption fails first. Not the AI.
Every workflow we have looked at has a visible pain point. The blocker is almost never model capability. It is that the context the AI needs to help — the documents, emails, prior files, guidelines, exception history — is distributed across systems that were never designed to talk to each other. Mapping context availability before building anything is not optional. It is the first deliverable.
Governance is a feature, not a constraint.
Regulated industries move slower on AI not because they are risk-averse, but because the cost of a wrong decision is real and traceable. The practitioners who move fastest in those environments are the ones who arrive with a governance frame already articulated — what AI does, what it does not do, who owns the decision, what gets retained for audit. That framing does not slow adoption down. It is what gets a pilot approved.
"First viable" is a deliberate scope. Not a limitation.
The temptation is to show the full AI transformation vision: end-to-end automation, real-time decisioning, enterprise-wide rollout. That vision loses deals and stalls pilots. The question that unlocks forward movement is: what is the one workflow where AI can demonstrably help, the human stays in control, and we can measure whether it worked within 90 days? Everything else is a later conversation.
The 80/20 rule is the only way to scale the practice.
An experienced practitioner entering a commercial insurance client should not start from scratch. Eighty percent of what matters — the value leakage patterns, the governance constraints, the evidence-assembly problem, the broker communication friction — is predictable from industry research. The other twenty percent is what the field conversation discovers: the specific system limitations, the political owner, the exception that breaks the pattern. Building the 80 into a reusable prior is what makes it possible to run ten engagements instead of one.
The six criteria that define a first viable workflow
A workflow earns the designation when it clears all six. Score is directional — the field conversation either confirms or breaks the hypothesis.
Visible pain
Accessible context
Bounded risk
Measurable outcome
Adoption owner
Reusable pattern potential
These six criteria emerged from field pattern synthesis — looking across engagements at what separated workflows that reached pilot from ones that stalled or failed. They are not a theoretical framework. Score is directional: it reflects hypothesis strength from source research and should be tested against what the field conversation actually reveals, not treated as a final answer.
How a field engagement runs
Before the first meeting
- Load the industry value chain and known leakage patterns.
- Identify the top two or three workflow candidates by score.
- Prepare field questions that test whether the hypothesis holds at this client.
In the first meeting
- Confirm visible pain with recent examples, timestamps, and rework loops.
- Test context accessibility — where do people leave the system to assemble evidence?
- Find the adoption owner — the specific person who can sponsor a narrow pilot.
Leaving the meeting
- One workflow named, scoped, and governed.
- A 30/60/90 path the client can approve without a transformation budget.
- A governance boundary that compliance will not push back on.
What this tool is for
- Encoding the 80 percent prior so every practitioner starts oriented.
- Generating a brief that is field-ready, not presentation-ready.
- Building a library that compounds with every engagement, not resets.