AI Strategy
The board is asking about AI. Here's what to tell them.
Your board wants an AI strategy. Here's how to give them a clear, honest briefing that sets realistic expectations and gets the green light for action.
It usually starts with a question at the end of a board meeting. "What are we doing about AI?" Or maybe someone forwards an article about a competitor using AI and adds a note: "Should we be doing this?"
If you are the person who has to answer that question, this post is for you. Not a technology deep-dive. Not a list of tools. A practical guide to giving your board a clear, honest briefing about AI that sets realistic expectations and gets the green light for action.
Why the board is asking now
Board members read the same headlines everyone else does. They see that AI is being adopted across industries and that companies are reporting significant productivity gains. They worry about being left behind.
That concern is not unreasonable. The Bank of England's analysis of AI and the UK economy suggests that AI adoption will create measurable competitive advantages for early movers, particularly in professional services, manufacturing, and distribution. The question is not whether to act, but when and how.
What the board usually wants to know is:
- Are our competitors using AI?
- Should we be using it?
- What would it cost?
- What are the risks?
- What do you recommend?
Let me help you answer each one.
"Are our competitors using AI?"
Probably, yes. According to the ONS Business Insights Survey, around 20% of UK businesses are now using at least one form of AI, up from under 10% two years ago. Among mid-market businesses (50 to 250 staff), the figure is closer to 35%.
But "using AI" covers a wide range. Some businesses are using AI-powered tools without even realising it (smart email filters, chatbots on their website, predictive text in their CRM). Others have made deliberate investments in AI for specific processes.
The honest answer for most boards is: "Some competitors are using AI in specific areas. We do not know the full picture, but adoption is accelerating and we should have a plan."
"Should we be using it?"
Yes, but selectively. AI is not a strategy in itself. It is a tool that helps you do specific things better, faster, or cheaper. The question is not "should we use AI?" but "which of our processes would benefit from AI?"
The areas where AI delivers the clearest results for businesses like ours are:
- Reducing admin time on repetitive tasks (data entry, document processing, report generation)
- Improving customer response times through automated triage and communication
- Better decision-making through faster access to accurate data
- Reducing errors in manual processes
If your business has any of these pain points, and most do, then yes, AI can help. But it is not about adopting AI for the sake of it. It is about solving real problems that cost you time and money.
"What would it cost?"
This is the question that matters most to a board, and the one that gets the most misleading answers in the market. So let me be straightforward.
A pilot on one specific process typically costs £500 to £5,000 depending on complexity and takes 2 to 4 weeks. A focused implementation covering two or three key processes runs £10,000 to £30,000 over 2 to 3 months. Ongoing costs for AI tools, maintenance and support sit between £500 and £2,000 per month depending on scope.
For comparison, the cost of not acting is usually higher than people think. If your team spends 20 hours a week on manual processes that AI could handle, that is roughly £25,000 a year in salary cost alone, before you account for errors, delays, and opportunity cost.
The Confederation of British Industry's productivity report has consistently shown that businesses which invest in process efficiency outperform those that do not, even accounting for the investment cost.
"What are the risks?"
The board needs to hear about risks. If you pretend there are none, you lose credibility. Here is an honest assessment.
The risks that are real
Implementation risk
Like any technology project, AI can be delivered badly. The mitigation is to start small, prove value on a pilot before committing to a larger project, and work with a partner who has relevant experience.
Data quality risk
AI is only as good as the data you feed it. If your data is inconsistent, incomplete, or scattered across multiple systems, you may need to clean it up before AI can help. This is not a reason not to proceed. It is a reason to include data preparation in the plan.
Change management risk
Your team may be resistant to new tools, especially if they have been burned by technology projects before. The mitigation is to involve the team early, start with tools that make their lives easier (not harder), and give people time to adapt.
The risks that are overstated
Job losses
In our experience, AI rarely leads to redundancies in SMEs. What it does is free people from repetitive work so they can focus on higher-value tasks. Most businesses use the time savings to grow without increasing headcount, or to improve service quality.
AI going rogue
For the kind of AI applications we are talking about (processing invoices, drafting emails, sorting customer queries), the risk of AI making dangerous autonomous decisions is essentially zero. Humans stay in the loop. The AI handles the boring bits.
Regulatory risk
The UK government's approach to AI regulation is deliberately light-touch for business applications. There is no imminent regulation that would prevent you from using AI for operational efficiency. Stay informed, but do not let regulatory uncertainty stop you from acting.
"What do you recommend?"
Here is the recommendation I would give to any board.
Start with a pilot
Pick one process that is manual, repetitive, and measurable. Run a small AI pilot lasting 2 to 4 weeks. Measure the results. This costs very little, carries minimal risk, and gives you real data to base future decisions on.
Set a modest budget
Allocate £5,000 to £10,000 for an initial exploration phase. That is enough for a pilot project and, if the results are positive, a plan for the next phase. You are not committing to a major technology programme. You are investing in finding out what is possible.
Assign an owner
Someone in the business needs to own this. Not as a full-time role, but as a clear responsibility. This person coordinates the pilot, reports back to the board, and makes sure the project does not drift.
Set a timeline
"We will run a pilot in Q3, report back to the board in Q4, and make a decision about next steps based on the results." That is a clear, reasonable timeline that gives the board confidence you are treating this seriously without rushing.
Report honestly
When you come back to the board, report what worked, what did not, and what you learned. If the pilot delivered strong results, recommend expanding. If it did not, explain why and what you would do differently. Boards value honesty more than optimism.
The one-page board briefing
If you need to put something in writing for the board, here is a template:
Current situation: We have [X] manual processes that cost approximately [£Y] per year in staff time and errors.
Opportunity: AI tools can automate significant portions of these processes, based on evidence from similar businesses in our sector.
Recommendation: Run a focused pilot on [specific process] over [timeframe] at a cost of [£Z]. Measure the results and report back.
Expected outcome: If successful, the pilot should demonstrate [specific metrics: time saved, errors reduced, etc.].
Risks: [Honest list]. Mitigated by: [starting small, involving the team, working with experienced partners].
Next steps: [Clear actions with owners and dates].
According to Deloitte's boardroom AI readiness report, boards that receive clear, honest AI briefings are significantly more likely to approve investment than those that receive either hype-driven or overly cautious presentations.
Getting the data for your briefing
The hardest part of a board briefing is often getting the specific numbers for your business. How much are manual processes actually costing? Where are the biggest opportunities? What would a realistic pilot look like?
That is exactly what our free AI opportunity report provides. It gives you a clear, data-backed assessment of where AI can help your business, with specific recommendations and realistic cost estimates. It is designed to be the evidence base for a board conversation.
Mark Blair
Founder, gofasterwith.ai
Frequently asked questions
What numbers should I take into a board meeting on AI?
A pilot on one specific process typically costs £500 to £5,000 and runs for 2 to 4 weeks. A focused implementation across two or three processes runs £10,000 to £30,000 over 2 to 3 months. Ongoing tools, maintenance and support sit at £500 to £2,000 a month. Set those against the cost of doing nothing: 20 hours a week of manual work is roughly £25,000 a year in salary alone, before errors, delays, and missed opportunity get counted in.
How should I answer the board's question about competitor AI use?
Honestly, with the ONS Business Insights figures: about 20% of UK businesses now use some form of AI, and roughly 35% in the 50 to 250 staff bracket. But that ranges from accidental use of smart filters and chatbots through to deliberate process investment, so the truthful answer is that some competitors are using AI in specific areas, the picture is incomplete, and adoption is accelerating. That is more credible than claiming you know exactly what every rival is doing.
Which AI risks should I raise with the board, and which are overstated?
Raise three real ones: implementation risk (mitigated by piloting before committing), data quality risk (mitigated by building data prep into the plan), and change management risk (mitigated by involving the team early). Push back on three overstated ones: SME automation rarely causes redundancies, the AI applications most boards actually approve cannot make autonomous decisions because humans stay in the loop, and the UK's regulatory approach is light-touch for operational use. Pretending there are no risks loses credibility fast.
What kind of pilot should I propose if the board says yes?
Pick one manual, repetitive, measurable process. Set a 2 to 4 week timeline and a £5,000 to £10,000 budget. Assign a single named owner inside the business, not as a full-time role but as a clear responsibility. Agree what success looks like in advance: hours saved, errors reduced, response time. Then come back to the board with what worked, what did not, and what you learned. Boards approve next phases on honest reporting, not optimistic forecasts.