The mid-size advantage
Large groups have bigger budgets. Mid-size companies have something more useful: shorter decision chains, leaders who know the operations personally, and the ability to change a process across the company in weeks rather than years.
The risk is the opposite of the large-company risk. Instead of a heavy central programme, mid-size companies tend to end up with a dozen small experiments, no owner and no measure of value. The roadmap below is designed to fix that.
Before day one: agree the baseline and the ambition
Spend one leadership meeting agreeing two things:
- Where you stand. A short readiness assessment, completed by two or three leaders, gives a shared baseline and surfaces disagreements early.
- Where you need to be in 18 months, dimension by dimension. Not every company needs to lead on AI. Every company needs to know which gaps block its strategy.
The output is three priority gaps. Everything in the next 90 days serves those three.
Days 1–30: mobilise
- Name a sponsor and an owner. A member of the executive committee sponsors the programme. A business owner, not only IT, is accountable for each priority domain.
- Publish an AI usage policy. One page is enough to start: approved tools, what data may not be shared, when a human must review.
- Make one approved tool available. If staff do not have a safe tool, they will use an unsafe one.
- List current use cases and pilots. Score each on value and feasibility. Stop the ones that score low on both.
- Choose two or three use cases in the priority domains, with a measurable target for each.
Days 31–60: build
- Deliver the first use cases with small cross-functional teams: a business owner, a process expert and the technical skills, internal or external.
- Fix the data each use case needs, and only that. Data programmes without a use case rarely finish.
- Redesign the process, not just the tool. Decide which tasks change, which roles change, and what the human now checks.
- Run a risk review before anything reaches customers or makes decisions on its own.
- Start a training programme, beginning with the leadership team and the teams involved in the first use cases.
Days 61–90: embed
- Measure the value of the first use cases against their targets, and decide to scale, adjust or stop.
- Set up the portfolio rhythm: a monthly review of use cases, value, spending and risks, owned by the sponsor.
- Report to the board with a short dashboard: adoption, value, incidents and regulatory readiness.
- Plan the next wave of use cases and the foundations they will need.
- Re-run the readiness assessment at six months to show progress against target.
Adapting the plan to your starting point
The same 90 days look different depending on where you start. The AI Readiness Check places organisations in one of four archetypes:
| Archetype | Starting point | Emphasis for the first 90 days |
|---|---|---|
| Explorer | Foundations and leadership choices both early | Ambition, ownership, policy and one or two quick, visible wins |
| Visionary | Strong ambition, weak foundations | Data, platforms and governance for the priority domains, before scaling pilots |
| Engineer | Strong foundations, weak business pull | Business ownership, process redesign and measured value |
| Leader | Both strong | Scaling, portfolio discipline and new business models |
Measuring progress
Keep the dashboard short. Five measures are enough for the first year:
- Share of employees using approved AI tools every week
- Number of use cases in production, and pilots stopped
- Value delivered against target for each use case in production
- AI incidents and policy breaches
- Readiness index against the 18-month target
Get your own roadmap
Start the assessment See a sample roadmap