Training & Capability Building
AI training for East African organisations where the output has to hold up — in front of a regulator, a funder, a board, or a customer. Delivered personally by Naomi Mbugua Mutinda.
18+ years in pharmaceutical marketing across Cipla Sub-Saharan Africa and GSK Kenya. AI trainer at Strathmore Business School.
How to Use AI in Business (AI for Founders)
Founders and business owners, executive education
Leveraging AI for Branding
Women entrepreneurs, professionals and leaders
Also speaks regularly on brand and marketing strategy — Big Baraza SME Conference, Sndbx | The Village Experts, Centonomy Entrepreneur Programme.
Asked for something it doesn't have, AI does not leave a blank. It writes a fluent sentence containing a number nobody in your organisation can trace. One unsupportable figure in front of a funder, a regulator or a client is not a typo — it is a due-diligence problem.
Four companies in the same sector, all using the same tool with the same prompts, all now sounding identical. AI does not dilute your positioning deliberately. It defaults to the average, and the average is your competitors.
The team is enthusiastic for two weeks, then reverts to the old way. Not because the tool failed, but because nobody built it into a workflow, and there was never anything to point back to.
None of these are tool problems. They are method problems, and method is teachable.
Every output produced in these sessions is labelled: Fact. Inference. Estimate. Assumption. Four words, applied to every sentence that carries a number or a claim.
It sounds like housekeeping. It is the difference between AI-assisted work you can put in front of a board and AI-assisted work you have to quietly check afterwards — which costs more time than writing it yourself.
This is not a technique borrowed from an AI course. It is how I have worked for eighteen years in pharmaceutical marketing, where an unsupported claim is a regulatory matter, not an editing note. It transfers directly, and it is the first thing I teach.
Before anyone writes a prompt, we get what the organisation actually knows out of people's heads and into a form AI can work from responsibly.
Most bad AI output is a decision that was never made, dressed up in good prose. We fix the decision.
Only now does anyone open a chat window.
The final exercise is always done without me in the room. If it only works while I'm standing there, it hasn't worked.
This is close to the reverse of how most teams reach for AI. It is also the reason the outputs hold up.
C.R.A.F.T. — the structure taught in every session, built for people who do not want to learn a new discipline, only to get better answers out of one tool. Participants leave able to construct a working prompt without a cheat sheet, which is the only test that matters two weeks later.
C.R.A.F.T. is how you get a good answer. The claims discipline is how you know whether to trust it. Most training teaches only the first.
Every session is a build session. Participants work on live material from their own organisation, in pairs, and something real exists at the end of it.
If the only thing that leaves the room is enthusiasm, the training has failed.
The trigger: Your team is producing more content than ever and it has stopped sounding like you.
Fits: AI for Marketing Workshop — half or full day, on your own brand material.
See formats and investmentThe trigger: You are the brand, you have no time, and generic AI output undermines the one thing you're selling — your voice.
Fits: 1:1 coaching, or the branding session format.
See formats and investmentThe trigger: You are running a programme, association or business school cohort and need someone who can hold a room of senior people who have already tried ChatGPT and been underwhelmed.
Fits: Speaking engagement or programme module.
See formats and investmentThe trigger: Your output is read by funders, regulators or boards. You cannot afford invented numbers.
Fits: Multi-session programme with a claims register and verification protocol built in.
See formats and investment| Service | What You Get | Investment |
|---|---|---|
| Discovery call | Anyone · 30 minutes to scope the actual problem | Free |
| AI for Marketing Workshop | Marketing and brand teams · Half or full day on your own brand material. Prompt library, voice guide and finished work leave with you. | KES 75,000–150,000 |
| Capability programme | Teams needing durable change · Multi-session build programme with claims register, verification protocol and unassisted production. Scoped to your material. | Scoped after discovery call |
| Speaking / programme module | Conferences, associations, business schools · Keynote or teaching session, tailored to the cohort | Custom quote |
| 1:1 coaching | Founders and individual leaders · Prompting, workflow and the tools relevant to your role | KES 15,000/session |
Workshops and programmes are scoped after a discovery call. Pricing reflects preparation on your material, not delivery hours alone — a session built on your own brand and evidence base requires work before the room.
Delivered as keynotes (30–60 minutes), half-day or full-day workshops and masterclasses, conference panels, and corporate team training — in person or virtual.
Need systems built rather than capability taught? See AI automation — or meet our consultant network.
A short structured interview with each participant about the work they know best. This becomes the raw material for the morning. About 45 minutes each, done in their own time.
What the tool is actually doing, and the three ways it fails. No hype, no doom.
Building the evidence base. Working in pairs on live material. This is where the claims register gets built.
Prompting on real briefs. The frameworks, then immediate application to work already sitting in the team's queue.
Unassisted production. Pairs produce finished work with no facilitator input. This is the assessment.
Workflow assignment. Each participant leaves with one named task, one method, and a date.
Most AI trainers know AI. Far fewer have been personally accountable for a revenue number. Every framework taught here comes from someone who has carried a P&L in a regulated industry, where an unsupported claim is a compliance matter rather than an editing note.
Full profile: Naomi Mbugua Mutinda. Pharma discipline in practice: the Cipla Kenya case study.
1:1 coaching is KES 15,000 per session. Workshops are KES 75,000–150,000 depending on length and preparation required. Multi-session programmes are scoped after a discovery call. Pricing reflects the preparation done on your own material before the session, not delivery hours alone.
Because the sessions are built on your material, not on a standard deck, and because the method is different. Most AI training teaches generation. This teaches verification — how to produce AI-assisted work whose claims can be traced and defended. If your output is only read internally, cheaper training may genuinely be enough.
That is the most common reason people call. Training fails when nothing leaves the room and no workflow is assigned. Every session here ends with built artefacts and a named task with an owner and a date. The final exercise is always done without me in the room.
Both. Workshops work for aligning a team on shared brand material. 1:1 coaching is better when a single leader wants deep, personalised capability.
The workshop is designed for marketing and brand teams. Coaching, programmes and speaking sessions work for founders, operations leaders and cross-functional teams. If your work involves content, communication, research or funding applications, it applies.
I do, personally. Milk & Honey works through a network of senior consultants on advisory engagements, but training is delivered by me — because the method is mine and the room needs someone who has used it commercially, not someone briefed on it.
The claims discipline taught in these sessions came out of regulated pharmaceutical marketing — this is what it produced commercially.
Read morePublished rates across advisory, AI training and coaching — and what sits behind each price.
Read moreA 30-minute call, no charge. Bring an example of AI-assisted work your team has produced recently and we'll look at it together. If training isn't the answer, I'll say so.