Your board has asked for an AI strategy. Three vendors have pitched in the last six months. The internal team cannot tell you which one to pick. And the budget is sitting in committee while the compliance clock keeps running.
This is the most common conversation we have with mid-market Consumer Products and Logistics tech leaders in the UK. The question is rarely should we do AI. It is who do we hire to do it.
There are three real answers — an AI company, a fractional Chief AI Officer, or a Big 4 consultancy. Each solves a different problem. Choosing the wrong one is how most UK CPG companies end up in Pilot Purgatory: 84% say they need to move faster on AI but only 3% have reached full AI deployment.
This guide is for VP Technology, CDO, CTO, COO, and Head of Operations buyers in UK CPG, FMCG, and Logistics companies between £100M and £2B revenue. By the end, you will know exactly which option fits your situation, what to ask before you sign anything, and the five warning signs that will tell you to walk away.
What is the difference between an AI company and fractional AI leadership?
An AI company builds and sells AI products. They have engineers, sales teams, and usually a platform. They are project-led. You hand them a brief, they hand back software. The contract ends when the software is delivered.
Fractional AI leadership is one or two senior AI executives embedded inside your business one to three days a week. They write the strategy, build the data foundation, ship the first working product, and train your team to run it after they leave. It is leadership, not delivery.
The two are not mutually exclusive. In most successful mid-market AI transformations, fractional leadership comes first — to decide what to build and why — and an AI company (or your internal team) builds it.
Quick comparison: four ways to get AI capability in your business
| Option | What you get | UK cost range | Time to first output | Best for |
|---|---|---|---|---|
| AI company | Built software, narrow scope, project-end | £25K–£250K per project | 8–24 weeks | You already have strategy, data, and an internal sponsor |
| Fractional AI leadership | Embedded CAIO + data engineering, strategy through adoption | £7,500–£18,000/month | Working prototype in 30 days | You need direction, data foundation, and change management — not just code |
| Big 4 consulting | Strategy report, named-brand credibility | £200,000–£500,000 minimum | 12–20 weeks | FTSE 250+ enterprises with internal delivery capability |
| Full-time CAIO hire | Permanent senior leader | £250,000–£400,000 salary + benefits | 4–6 months to appoint | Companies committed to a multi-year, multi-product AI roadmap |
A practical note for UK mid-market buyers: a £15K–£25K AI FlightPath™ Sprint usually sits below most procurement committee thresholds. A £200K Big 4 engagement does not. Speed of approval is itself a strategic advantage when the board is asking when, not whether.
When you need an AI company
You need an AI company when five conditions are already true:
- You have a clear AI strategy aligned to a specific P&L outcome — margin per SKU, cost per delivery, case fill rate, dwell time, demand forecast accuracy
- Your data is clean, centralised, and AI-ready (most companies discover at this point that it is not)
- You have a senior internal sponsor who owns the outcome, not just the budget
- You know exactly which use case to build and what success looks like in numbers
- You have change management capability in-house, or a credible plan for how the new tool will be adopted
If all five are true, hire the AI company. They will move faster than fractional leadership because the upstream thinking is done.
If any one of those is missing, you do not have an AI company problem. You have a leadership problem. And no AI company will fix that for you — they will deliver software that lands on a team that was not ready, and within six months that software will be quietly decommissioned. We have seen this pattern eleven times across UK CPG and logistics businesses in the last five years.
When you need fractional AI leadership
Fractional AI leadership is the right answer when the situation looks like this:
- The board has demanded an AI strategy and nobody internally can credibly write one
- Three AI vendors have pitched conflicting recommendations and you cannot evaluate them objectively
- Your last AI pilot delivered positive results in a sandbox and was never scaled
- Your data sits across siloed ERP, WMS, or CRM systems — or in Excel — and you suspect that is the real blocker
- You have hired a data scientist who spends 80% of their time on data preparation instead of analysis
- A new CIO or CTO is in their first 180 days and needs early wins
- You are PE-backed and the operating partner has a 100-day AI plan that you have not yet seen written down anywhere
The honest test: if you cannot explain, in one sentence, exactly which AI use case you are buying and what business outcome it will produce, you are not ready for an AI company. You need someone to do the navigation work first.
That is what fractional means. You get a Chief AI Officer with 20+ years of sector experience for one to three days a week, at a fraction of the £250K–£400K cost of hiring one full-time — and you get them in days, not the four-to-six months it takes to run a CAIO recruitment cycle.
How to choose an AI company for your FMCG or logistics business — a UK buyer's checklist
Whether you are evaluating an AI company directly or commissioning fractional AI leadership to evaluate them on your behalf, these are the ten criteria that separate the AI companies who deliver from the ones who disappear after the kickoff workshop.
Score each prospective AI company from 0 (no) to 2 (strong yes). Below 14 out of 20 is a walk-away.
The 10-point UK FMCG and logistics AI company checklist
| # | Criterion | What "strong yes" looks like |
|---|---|---|
| 1 | Sector references in your industry | They can name three CPG or logistics clients at your revenue band, and you can speak to one of them |
| 2 | Production deployments, not just pilots | They will show you a system running in production, used by real operators, with measurable outcomes — not a slide of "case studies" |
| 3 | Specific outcome metrics | Their case studies cite case fill rate, margin per SKU, cost per delivery, demand forecast accuracy — not "increased efficiency" |
| 4 | Data engineering capability | They have data engineers, not just data scientists. Most AI projects fail at integration, not modelling |
| 5 | Honest about your data readiness | In the first call they ask about your ERP, WMS, and master data — and tell you what they need before they can commit to outcomes |
| 6 | A named senior person on delivery | The person in the pitch is the person doing the work. Not a sales engineer who hands off to juniors |
| 7 | Change management capability or partner | They have a plan for how your team will use the tool after they leave. If they shrug at this question, you have your answer |
| 8 | Fixed-scope, fixed-price option | They will commit to a small, defined diagnostic or first-phase deliverable. If everything is "time and materials," they are protecting themselves at your expense |
| 9 | 30-day working prototype, not a 12-week roadmap | The right AI partner delivers something working inside 30 days. Anything longer to first output is a planning company, not a build company |
| 10 | Exit clause and IP ownership | You own the code, the data, and the models. You can leave with 30 days' notice. Lock-in is a red flag |
Questions to ask any AI company before you sign a contract
By the second meeting you should be asking these directly. The quality of the answers, not the polish of the deck, is what tells you whether to proceed.
1. "Show me a working AI product you have shipped in the last 12 months — running in production, used by real operators."
If they cannot, they are a strategy company, not a build company. The phrase "we are currently in pilot with…" is the most common warning sign of an AI company that has never reached production.
2. "What does our data need to look like for this to work? Where will you check before quoting?"
A strong AI partner will ask to see — or audit — your data schema before they price the work. An AI company that quotes without auditing your data is either guessing or planning to bill the data-fixing as a costly change order in week 4.
3. "Who, by name, will do the work? Will the person in this room be the person on delivery?"
Big 4 firms and many mid-tier AI companies pitch with partners and deliver with first-year analysts. Get the names. Get them in the statement of work. We have seen at least seven UK CPG transformations stall because the senior name in the pitch was unreachable by month two.
4. "What is your view on our data foundation? Be honest."
The right answer is uncomfortable. A vendor who tells you everything looks fine when it does not is selling you the project they want to sell, not the project you need.
5. "What happens at month 6 if adoption is at 8% instead of 80%?"
8% is the realistic adoption number for AI tools deployed without a change plan. If the AI company has no answer for this question, the entire engagement is structured to deliver software, not outcomes. You will own the failure.
6. "Who owns the code, the data pipelines, and the trained models when we part ways?"
You should. If the answer is "we retain ownership of the platform and license it to you," you are buying lock-in, not capability.
7. "Can we start with a fixed-price, fixed-scope, 30-day diagnostic?"
Any AI partner worth working with at scale will say yes. A two-to-four week diagnostic lets both sides find out whether the relationship works before committing to a six-figure programme. We structure our own engagements this way for exactly this reason the AI FlightCheck is designed to make the first decision low-risk on both sides.
8. "What is your CPG or logistics sector experience, and who specifically has it inside your team?"
Generic AI experience does not transfer cleanly to your S&OP process, your retailer compliance demands, or your last-mile cost structure. If their "sector experience" is one mid-sized client three years ago, they will be learning your industry on your invoice.
9. "What is the named-person handover plan at the end of this engagement?"
The best AI companies leave behind an internal team that can run the system without them. The weakest ones engineer ongoing dependency. Ask to see the handover plan in writing before you sign.
10. "What would make you walk away from this project?"
Any partner who says "nothing — we make every project work" is either lying or about to fail. The AI companies and fractional teams worth hiring have clear deal-breakers: bad data, no executive sponsor, no change capacity, unrealistic timelines. If they have none, they have not seen enough projects fail to know what kills them.
5 red flags when evaluating an AI company for mid-market CPG
These are not theoretical. Each one maps to a real pattern we have seen kill UK FMCG AI programmes — usually within the first 12 weeks.
Red flag 1: The pitch is more polished than the proof
If you have seen a beautiful 60-slide deck but cannot get a clear answer to "show me one product in production, used by real operators, with named clients we can call" — you are looking at a sales operation, not a delivery one.
The best AI companies and fractional teams will show you working software before they show you a deck. Live products, real users, measurable outcomes. Polish without proof is the most expensive purchase in the AI category.
Red flag 2: They have never worked in CPG, FMCG, or logistics — but they are "confident the methodology transfers"
It does not. CPG and logistics are not generic verticals. Retailer compliance, trade spend, S&OP rhythm, case fill rate, last-mile cost structure, agency labour models — these are not learnable on your project. The first three months of a generalist's engagement is them learning your sector. You are paying for it. And the institutional knowledge leaves with them.
This is why we focused AI Navi exclusively on Consumer Products and Logistics. Haja was Global Director of Data and Analytics at pladis ($3B+ CPG — McVitie's, Godiva, Ülker). Abhishek led data and AI strategy at the same business, after five years at Deloitte running supply chain traceability work. We do not learn your sector. We have sat in your chair.
Red flag 3: There is no data engineer on the team — only data scientists and "AI consultants"
Most AI projects fail at integration, not modelling. Without data engineering, your data scientist spends 80% of their time cleaning rather than analysing. The model never reaches production because the pipeline to feed it does not exist.
If the AI company you are evaluating has six "AI consultants" and no data engineer, they will be subcontracting that work or, more commonly, leaving it for your team to handle after the contract ends.
Red flag 4: They cannot tell you what your data foundation needs to look like
A serious AI partner will ask, in the first or second meeting, what ERP you run, what state your master data is in, how your transactional data is structured, and where the gaps are. They want to see a sample schema before quoting.
A weak AI company will skip all of this and quote on a fixed scope of work that will quietly fall apart when their team encounters your actual data. The renegotiation happens at week four, when it is too late for you to walk away cheaply.
Red flag 5: The price is either far below or far above the UK market
The UK 2026 benchmarks: fixed-price diagnostics £4,500–£10,000; 8–12 week sprint engagements £15,000–£50,000; ongoing fractional retainers £7,500–£18,000/month; Big 4 strategy engagements £200,000–£500,000 minimum.
An AI company quoting £3,000 for a full strategy and build is either inexperienced or undercutting to win the logo. Either way, the project will not finish at that price. Conversely, anyone quoting £150,000 for a single use case at a mid-market business is pricing for a market they should not be in.
What about Big 4 AI consulting?
If you are a FTSE 250 enterprise with £500K of unallocated AI budget, a mature internal data engineering team, and a board that needs a named brand on the proposal, the Big 4 is a reasonable choice. They are not what kills mid-market AI programmes.
What kills mid-market AI programmes is the £200,000 strategy report that sits on a shelf because the company has no internal delivery capability to act on it. The Big 4 produces strategy decks beautifully. The implementation gap is yours to close — and most mid-market CPG and logistics businesses cannot close it without an embedded leader.
Fractional AI leadership exists for exactly this gap. It is the senior AI executive you cannot afford to hire, available on the days you need them, with skin in the delivery game.
How much does it cost? UK 2026 pricing benchmarks
We publish our pricing because most of this market does not, and the opacity is part of what makes the buying decision so hard.
| Engagement type | UK 2026 price range | Typical duration | Outcome |
|---|---|---|---|
| Fixed-price AI diagnostic | £4,500–£10,000 | 2 weeks | Board-ready audit, prioritised roadmap, go/no-go on the bigger programme |
| AI sprint / first product delivery | £15,000–£50,000 | 8–12 weeks | Working AI product in production, internal team trained, ROI measurement framework |
| Fractional CAIO retainer | £7,500–£18,000/month | 3-month rolling | Embedded senior leadership 1–3 days/week, ongoing delivery and adoption |
| AI company — project basis | £25,000–£250,000 | 12–24 weeks | Built software, scope-limited |
| Big 4 AI strategy | £200,000–£500,000 minimum | 12–20 weeks | Strategy deck, implementation handed back to client |
| Full-time CAIO | £250,000–£400,000/year + benefits | Permanent (4–6 month hire cycle) | In-house leadership |
The key number to anchor on: a Big 4 engagement at £200K minimum, or a full-time CAIO at £250K–£400K per year, are not the only options for mid-market UK CPG and logistics. They are simply the two most visible.
If you are spending less than £100K on your first AI engagement, you want a partner who structures their entry point at sub-procurement-committee thresholds so you can move fast — and who is honest about whether your data and team are ready before they take your money.
How AI Navi works — and when we will tell you we are not the right fit
We are a fractional Chief AI Officer team for UK mid-market CPG, FMCG, and logistics businesses. Haja leads CPG and FMCG engagements (former Global Director of Data and Analytics at pladis Global, former Group Digital Director at Saint-Gobain UK). Abhishek leads logistics and supply chain engagements (former Global Head of Data and AI at pladis, five years at Deloitte Consulting, three at TNT Express / FedEx running global operations analytics).
We will tell you we are not the right fit if:
- Your business is below £50M revenue or above £2B — there are better-priced specialists at both ends
- You already have a strong internal CAIO and a clean data foundation — you need an AI company, not us
- You need a sector outside CPG, FMCG, manufacturing, logistics, or supply chain — our credibility lives in those rooms, and we will refer you to the right person elsewhere
- You are looking for an AI training programme or a workshop series — we build, we do not run workshops
Where we do fit, the engagement starts with a fixed-price, two-week AI FlightCheck™. You leave with a board-ready audit, a prioritised 90-day plan, and clarity on whether the bigger programme is right for you. Most clients do continue. Some do not. Both outcomes are fine.
Frequently asked questions
What is the difference between an AI company and a fractional Chief AI Officer in the UK?
An AI company builds and delivers AI software on a project basis, typically £25,000–£250,000 per project. A fractional Chief AI Officer is an embedded senior leader who provides strategy, data engineering oversight, and change management on a retainer of £7,500–£18,000/month. AI companies build what you tell them to build. Fractional CAIOs decide what to build, build the first version, and train your team to own it.
How much does a fractional Chief AI Officer cost in the UK in 2026?
UK fractional Chief AI Officer engagements range from £7,500 to £18,000 per month depending on the number of days per week. Fixed-price diagnostic engagements (typically two weeks) cost £4,500–£10,000. Sprint engagements (8–12 weeks delivering a working AI product) cost £15,000–£50,000. These prices are below most procurement committee thresholds, which is a key reason mid-market UK CPG and logistics buyers choose fractional over Big 4.
When should I hire an AI company instead of a fractional CAIO?
Hire an AI company when you already have: a clear strategy aligned to P&L, clean and AI-ready data, a senior internal sponsor, a defined use case, and in-house change management capability. If any of those five are missing, you need fractional AI leadership first — to fix the upstream problem — before an AI company can deliver value.
What are the biggest red flags when choosing an AI company in CPG or logistics?
Five red flags: no production AI products to show; no CPG or logistics sector experience; no data engineer on the team; cannot tell you what your data foundation needs to look like; pricing that is far below or far above the UK 2026 market range (£4,500 diagnostics to £50,000 sprints to £18,000/month retainers).
Is a Big 4 consultancy the right choice for mid-market AI?
Rarely. Big 4 AI engagements start at £200,000–£500,000 minimum, deliver strategy decks rather than working software, and assume the client has internal delivery capability — which most mid-market UK CPG and logistics businesses do not. Big 4 fits FTSE 250+ enterprises with mature internal teams. For £100M–£2B revenue businesses, fractional AI leadership and specialist AI companies are almost always faster and cheaper.
What questions should I ask an AI company before signing a contract?
The ten essential questions: show me a working AI product in production; what does our data need to look like; who specifically will do the work; what happens if adoption is 8% at month 6; who owns the code and models; can we start with a fixed-price diagnostic; what is your specific CPG or logistics experience; what is the handover plan; what would make you walk away. The quality of the answers, not the polish of the deck, is what tells you whether to proceed.
Where to start
If you are not yet sure which option fits your business, the lowest-risk first step is the AI Readiness Scorecard a 15-question diagnostic that gives you a Navigate / Execute / Land readiness score against the UK mid-market benchmark. It takes three minutes and tells you whether the next move is an AI company, a fractional CAIO, a Big 4 engagement, or none of the above yet.
Or if you would prefer a 30-minute conversation first, we run a free AI Navigation Call, no pitch, no slides, just an honest read on whether AI Navi is the right partner for your situation.