From a plan we build with you to ways of working your team owns
We work in ChatGPT and Claude, with the people who do the work. First we find where AI genuinely helps and put it in a plan. Then we deliver it together, train your team on it, and hand it over working.
How an engagement runs
- Discovery call.
We learn your context and the work you want to improve, and tell you whether the assessment is the right next step.
- Two-week AI Opportunity Assessment.
We sit with the people doing the work, evaluate the candidate workflows, and plan what implementation needs.
- AI Action Plan and readout.
You receive the plan we built with you and a live session walking through it.
- Four weeks of delivery.
We build the agreed workflows, train your people, and set up measurement.
- Results review.
We review the results with you after handover.
- Ongoing advisory support.
Continued help when you want it.
Find the work that matters: the AI Opportunity Assessment
Where AI can create meaningful operational value for one team, which workflows to address first, what implementation requires, and how success will be measured.
Who takes part
An executive sponsor who owns the outcome, the leader of the function, a handful of the people who do the work, an internal champion, and IT or security when access or policy makes them relevant. We use the least access necessary and can work through walkthroughs and examples where direct access isn't appropriate.
What happens over the two weeks
We align on the function you want to improve, the people involved, and the business consequence you're after, and we look at how AI is used today.
We sit with the people who do the work: interviews, walkthroughs, screen sharing, and real examples. We map the recurring work, the handoffs, the repeated briefing, the delays, and the rework.
We evaluate the candidate workflows for value, frequency, feasibility, data readiness, risk, adoption burden, and measurability, and set aside the ones that don't hold up.
We select two or three priority workflows and define, for each, the intended outcome, the baseline or how we'll capture one, and who owns it.
We plan the implementation: ChatGPT or Claude, the knowledge and instructions it needs, access and permissions, training, ways to check the output, ownership, effort, and risks.
We present it back in a live readout, resolve any corrections, and agree the recommended sequence.
What you receive
The AI Action Plan we built with you: the findings, the prioritized workflows, what each one requires, how it will be measured, the risks, and the delivery sequence. Plus a live executive readout. Detailed production design is part of delivery. And we can execute the plan together.
The assessment fee is credited toward implementation when you continue with us.
If we can't find a responsible, measurable, commercially viable path within the two weeks, we tell you why and refund the assessment.
See how this ran at Thunderbird Press.Build it into the way you work: delivering your AI Action Plan
Over four weeks we deliver the plan we built with you, typically on up to two priority workflows in one function. You're getting a working capability, and it comes in six parts.
Working workflows
The agreed workflows running in ChatGPT or Claude, tested on representative work, with the right checks and hand-offs to a person where they belong.
Company context and working instructions
Your approved knowledge organized for the workflows; reusable instructions, projects, assistants, or agents where they help; clear ownership of sources and versions.
Evaluation and measurement
A baseline or a way to capture one, representative test cases, acceptance criteria, and a method for measuring workflow and adoption outcomes.
Live training and adoption
Role-appropriate live sessions where people complete real work. Confusion, failures, and workarounds get corrected. Your internal champion is ready to keep supporting the team.
Governance and ownership
A named owner for each workflow, clear permissions and sign-offs, and simple procedures for questions, problems, and changes that fit the risk.
Operational handover
Final instructions and operating materials, maintenance responsibilities, known limitations, and the measurement and follow-up plan.
What done looks like
The agreed workflows work on representative tasks, the included people are trained, the internal owner can support normal use, and the checks, testing, and measurement are in place. Business outcomes are measured over the agreed period, and early results are captured where the period allows.
Results review
After handover we review the results with you: adoption, the agreed measures, early outcomes, and what to do next, timed to how long the workflow needs to show its effect.
Keep it working: ongoing advisory support
Ways of working change as the business, the team, and the tools change. When you want continued help, ongoing advisory support keeps a senior person from ASMR close to the work.
Leadership advisory calls
Regular conversations about strategy, performance, priorities, and what to do next.
Team working sessions
Focused time on existing workflows, responsible use, adoption barriers, and agreed improvements.
Platform monitoring and updates
We watch what changes in ChatGPT and Claude, explain what matters, and update your workflows.
Opportunity scoping
We identify and prioritize new workflows as the organization develops.
Adoption tracking
What's being used, what isn't, why, and where a workflow or a person needs attention.
Quarterly reviews
Measured outcomes, adoption, open risks, and priorities for the next quarter.
Arranged after delivery, shaped around the capacity you want.
Questions people ask
Do we need to know which workflow to start with?
No. You need a function or team with recurring work you want to improve and a consequence you can feel. Finding and prioritizing the workflows is what the assessment does.
Do you need access to our systems?
Not necessarily. We use the least access needed. Interviews, guided walkthroughs, screen sharing, and representative examples are often enough. Where direct access matters, we agree it with your IT or security owner.
What if you can't find anything worth doing?
Then we tell you why, and we refund the assessment. We'd rather stop than hand you a plan we don't believe in.
How many people are involved?
Enough to represent how the work is actually done and to make the decisions: a sponsor, the function leader, a few of the people doing the work, a champion, and IT or security when relevant. One person can wear more than one hat.
Which platform will you recommend?
ChatGPT or Claude, based on what the work needs, what you already use, and what your people will actually adopt. We work in both.
How long before we see results?
Working workflows, trained people, and running measurement are in place at the end of the four weeks. Business results are measured over the agreed period, because they depend on adoption as well as the capability, and we review them with you after handover.
What happens on a discovery call
We get to know your context: what the company does, your role, the team you have in mind. We understand the work and the problem: what happens today and where it hurts. We tell you honestly whether we can help, and what the right starting point is. Then we agree the next step.