M.A.I. Consulting
Scope estimate

Build the engagement you need and see its scope in delivery days.

Choose the scope and the page shows the delivery days, where each one comes from, and what is included. Priced per engagement: you receive a written quote after a free scoping call, where the scope is fixed in writing.

1 · What kind of organisation are you?

Your answer tells us who we are speaking with. Individuals and organisation representatives take the course for fewer than ten participants.

2 · How many staff would this cover?

Size drives how many departments we work with and how many sittings a session needs, so it moves the day count.

3 · Readiness assessment

The assessment produces the baseline score, the inventory of tools in use and the ranked gap list the rest of the work is built on. Two depths, with different day counts.

Four essential dimensionsThe four competencies of AI fluency, scored across the organisation with a short interview round. It answers what staff can and cannot do, and what to train first. It does not cover governance, data protection, workforce impact or psychological safety.
Ten dimensionsThe full instrument: sixty questions, staff interviews and a documented tool inventory, with the optional standards mapping in the vocabulary of ISO/IEC 42001, the NIST AI RMF and EU AI Act Article 4. Choose this when a board, a funder or an auditor will read the report, or when the question is governance rather than skills. It is a management report, not certification preparation.

See what the report contains →

4 · Training

Training is scoped by where the group starts and where it needs to reach. The seven levels, the days between them and what each one changes are set out on the training page.

What each level means
Level 1No working use of an assistant yet. Nothing to unlearn, and the objections are specific enough to answer with the team's own documents.
Level 2Occasional use with no shared method: a summary here, a search there. Output varies by person and by day because the input does.
Level 3Structured prompting and verification. The person briefs the work as they would a colleague, gives the source document, and checks figures, facts, language and defensibility before anything is sent. This is the floor we train to.
Level 4Persistent set-ups. Projects hold the reference documents, custom instructions hold the standing rules, skills hold the house method. Nothing is briefed twice.
Level 5Delegated work. An agent carries a job from brief to draft while the person supervises, and several workstreams run at once.
Level 6Building for others. Designs agents and skills for colleagues, connects the sources, and reviews the team's AI-assisted output. This is what makes the capability survive turnover.
Level 7Terminal-based agentic work with no coding background. Claude Code operates on folders of files rather than one conversation: batch document work, the same review across a year of reports, a repeatable pipeline.

5 · AI use policy

General means one umbrella policy: permitted uses, red lines, data classification, disclosure, escalation, approved tools. Specific annexes add rules for a particular department or use (casework data, fundraising, HR records) where the umbrella is not enough.

See the policy structure and sample clauses →

6 · Custom agents, skills and tools

Three different things, counted separately: about 3 delivery days per agent, 1.5 per skill and 2 per tool, scaled by the size of the organisation and of the department that will use them, because testing and handover grow with both. Each is configured on the platform you already use, tested on your documents and handed over with an operating guide. We configure existing platforms; writing software and training models are outside what we do. See worked examples of each →

7 · Fluency Support

A monthly retainer after the engagement: office hours for whoever is stuck, playbook revisions as the tools change, coaching for your internal champions, and a re-score at six months. One or two delivery days a month. The six-month re-score adds 2 delivery days once. See what a support month contains →

8 · Which tasks are you trying to improve?

Each task you pick is built into the training material and the playbooks: the patterns, the worked examples and the checks are prepared for that task on your documents. The first three carry most of the preparation, and further ones take less time.

9 · Anything we should know

Languages, deadlines, sites, data constraints, a funder requirement, or a task not listed above. This is sent with your estimate.

How to read the estimate

How to read the estimate.

The estimate is a count of delivery days and a list of what you hold at the end. It is not a quote. Priced per engagement: once we have spoken for thirty minutes and know your actual departments, data and constraints, you receive a fixed scope and a written quote after that free scoping call.

What moves the day count: the level you are training to, the size of the organisation and of the departments involved, the depth of the assessment, the number of policy annexes, how many agents, skills and tools you want configured, and how many months of support. When training and policy, or configuration and training, are scoped together, some preparation and sessions are shared, and the breakdown shows the days that overlap.

The smallest engagement is three delivery days. For any engagement, and for individuals as well as groups, we can host the training in person and build it around what the participants need. The arrangements are agreed on the call and confirmed in the written quote.

Would you like a written quote?

A free thirty-minute scoping call. We check the scope, tell you if any part of it is unnecessary, and send the fixed scope and a written quote the same week.

Book a call See the services