Question 1

Who are these programmes actually built for?

Accepted Answer

Most of our work is with specific departments that want AI applied to their real day-to-day work. Marketing teams use it for campaign briefs, content production and audience research. Sales teams for prospect research, call prep, proposal drafting and CRM hygiene. Customer support for reply drafting, ticket triage, knowledge-base upkeep and QA. Operations, HR, Finance and Legal each have their own version. We also run cross-department cohorts when leadership wants shared language, joint use cases and ideas that only surface when Marketing, Sales and Support are in the same room.

Question 2

Can you tailor the use cases to one department, or do we have to run a generic AI course?

Accepted Answer

Tailoring is the default, not an upsell. Before day one we interview two or three people from the target function, review a sample of their actual work (with anything sensitive redacted), and build the exercises around the workflows that eat their week. A marketing cohort will not sit through sales examples, and vice versa. When a client wants a company-wide programme we deliberately mix functions, then split into department pods for the build days so each group still leaves with something usable on Monday.

Question 3

Why run a cross-department cohort instead of training each team separately?

Accepted Answer

Two reasons. First, most high-value AI use cases sit between functions, not inside them: the handover from Marketing to Sales, from Sales to Support, from Support back to Product. You only see those when the people involved build together. Second, a shared cohort creates a common vocabulary and a shared quality bar for prompts, reviews and governance, which is what stops every team from reinventing the same wheel six months later.

Question 4

How do you measure the business impact?

Accepted Answer

For each participating team we agree two or three target workflows up front and capture a baseline: hours spent, cycle time, error rate or volume handled. During the build days every prototype is scored against that baseline including token and licence cost per run. Within seven days you receive a written report with prototypes built, validated time savings, projected annual value, total cost of ownership and the recommended next step per idea. That is what your CFO, COO or transformation lead can act on.

Question 5

How do you handle data confidentiality and regulation?

Accepted Answer

We work inside your approved environment: enterprise ChatGPT, Copilot or Gemini, a private deployment, or open-source models in your own cloud. In the workshop we use synthetic or anonymised data so nothing production-sensitive leaves your perimeter. We cover the operational side too: data residency, GDPR and sector rules, prompt and output logging, human-in-the-loop checkpoints, and clear no-go rules for public models.

Question 6

Our team has used ChatGPT for months with little to show for it. Why will this be different?

Accepted Answer

Stalled adoption usually comes from the same four gaps: no shared backlog of use cases, no prompt standards, no link to the systems people actually work in, and no measurement. We tackle each directly. We start from your highest-volume workflows, build reusable prompt libraries and light automations against them, document what works, and leave the team with a prioritised roadmap. The output is a small portfolio already in use, not another round of experiments.

Question 7

Who should attend from each department?

Accepted Answer

The strongest pods mix three profiles: people who own the target workflow day to day (a campaign manager, an SDR lead, a support team lead), someone with authority to change process and tooling in that function, and one or two curious builders who will carry prototypes forward. We keep each pod to 8 to 16 people so facilitation stays deep. Executive sponsors are welcome for the opening framing and the final demos, and most join both.

Question 8

What is the realistic total cost once ideas go into production?

Accepted Answer

Beyond the programme fee, ongoing cost usually includes model usage (token cost scales with volume and model choice), enterprise licences such as ChatGPT Enterprise, Copilot or Claude for Work, integration and hosting, monitoring, and the internal time for governance and review. We model these for every prioritised use case during the programme, so weak unit economics get filtered out before they become a line item.

Question 9

What support is available after the programme ends?

Accepted Answer

Each engagement ends with a written report, the prototype assets and a prioritised roadmap. From there clients usually pick one of three paths: continue internally using the materials, retain us for structured implementation support on the highest-value ideas (architecture, governance, rollout), or run a follow-on cohort for an adjacent department. The default scope ends at the report. Anything beyond that is opt-in.

Question 10

What does success look like 90 days after delivery?

Accepted Answer

For a departmental cohort, a typical 90-day picture is two to four workflows in regular use inside that team, measurable hours returned per week, a governance baseline the team actually follows, and a leadership-ready business case for one or two larger moves. For a cross-department cohort you also get one or two joint workflows across functions and a shared review standard, which is usually where the compounding value comes from.

Your people will not remember AI training. They will remember using it.

Why traditional AI training rarely sticks

Passive learning fades fast

Generic content, no real work

No confidence to keep going

How capability is actually built

Learn

Apply

Build

Adopt

People retain what they use, not what they hear

Retention through application

Confidence from doing

Judgement, not just skill

Adoption that outlasts the programme

Confident people, capable organisations

For your employees

For your organisation

Pick the right starting point

AI Foundations Sprint

AI Impact Lab

AI Implementation Support

Capability building, not another workshop

Capability is the product

Real work, from day one

Confidence, not dependence

Judgement, not just tools

A shared language across teams

Designed to compound

Practitioners who have built AI inside real businesses

Khurrum Ghori

Rahaf Al-Najjar

Anna Rehermann

Taking strong ideas further, if you want to

Practical answers before we start

Let's talk about building AI capability in your team