AI Coaching & Workflows
Bring a task you repeat, avoid, or cannot yet use AI for reliably. I help you define the outcome, map the current method, choose where AI earns a place, and build the prompts, checks, and documentation needed to run it with confidence.
Request a fit callInteractive process
Preview any stage with the numbered controls. Click, tap, or press Enter to keep that explanation selected.
We identify the decision, document, analysis, or workflow you need to produce, then define what good looks like and when it is needed. Only then do we decide whether AI has a useful role.
Together, we map the current steps, skill level, approved tools and licences, data sensitivity, quality risks, and how often the task repeats. This reveals what should stay human and what can be supported safely.
We choose only useful AI touchpoints and connect them with prompts, human checks, and documentation. A repeatable design beats a collection of disconnected chats.
Outputs are tested against sources, assumptions, privacy requirements, and relevant rules. Reliable use depends on knowing when the model should not be trusted.
You should be able to explain, run, adapt, and stop the workflow yourself. The tool may assist, but judgment, accountability, and final decisions remain human.
Recommended starting point
A 90-minute online session plus a short written next-step plan. We examine your current work, identify one realistic workflow, clarify risks, and define what to do next.
You receive the complete format and provide structured private feedback after the session.
What we work on
Every project begins with an assessment. The right format then depends on whether one person needs a working method, a team needs shared rules, or a proven workflow should become reusable material.
For one person and one concrete task. We assess the current method, privacy and quality requirements, then build one repeatable workflow with explicit human checks.
For 2-6 people who need one common way of working. We assess roles, skill levels, approved tools, licences, hand-offs, and shared risks before designing the team method.
After the diagnostic has proven the method, it can be translated into prompt templates, checklists, workflow notes, or carefully bounded support materials. This is scoped separately.
Capability ladder
A broad introduction of about two hours can cover how language models work, prompting, hallucinations, verification, privacy, and limits.
Move from isolated chats to repeatable, documented ways of working with ChatGPT, Claude, Microsoft Copilot, or another suitable environment.
Combine task steps, human checks, reusable prompts, and realistic ownership so the workflow can survive beyond one session.
Feasibility reviews and guided prototypes only after scoping data, permissions, oversight, dependencies, maintenance, and failure modes.
A grounded next step
The short intake identifies the aim, constraints, and right format before any paid work begins.
