A founder has £40,000 of unallocated technology budget for the year. A consultant has proposed a £42,000 programme described as “comprehensive AI transformation”. She has read the proposal twice and does not yet have a clear picture of what she will hold in her hand at the end. The frame she needs is not what does this cost. It is at this tier, what should I expect to receive.
This is the single most useful re-frame an SME owner can make when comparing AI advisory proposals. Three real budget tiers exist at SME scale. Each tier buys a different shape of deliverable. Owners regularly pay one tier and expect the output of another, then experience the gap between expectation and delivery as a failed engagement. The fix is to know what each tier actually gets you before signing.
Why budget-to-deliverable mapping is unclear at SME scale
The market does not advertise the tiers cleanly. Proposals at £10,000, £25,000, and £50,000 are pitched in similar language. They reference roadmaps, implementation, governance, and outcomes. They promise broadly similar things in broadly similar tone. The tier structure is implicit, not stated, and the consultant has no commercial incentive to make it explicit, because making it explicit would surface that the £42,000 proposal is overpriced for the deliverable list and underpriced for the deliverable list it would need to actually shift the business.
The tiers are real. Below £10,000, you are buying a workshop or a coaching conversation, not a roadmap. Between £10,000 and £15,000, you are buying a roadmap. Between £20,000 and £30,000, you are buying a roadmap plus a single pilot. Between £40,000 and £60,000, you are buying a roadmap, pilot, and governance retainer for several months. The shape changes at each tier, and reading the proposal against the tier it sits in tells you whether you are getting fair value.
What 10k actually buys
At the £10,000 tier, you are buying a roadmap engagement. Six to eight weeks of work culminating in five outputs: business process interviews with named stakeholders, a data readiness assessment, a prioritised use case list with quantified business value, a recommended sequence with budget tiers attached, and a risk register. The deliverable is a strategic document, not running systems.
What you are not buying at this tier is implementation. No code is being written. No models are being trained. No tools are being deployed. If the £10,000 proposal in front of you promises implementation alongside the roadmap, the consultant is either over-promising and under-scoping, or doing both at a level neither will land properly. A serious consultant working at this tier will be honest that the deliverable is a document the business uses to make subsequent buying decisions, including which delivery partner to engage next.
The £10,000 tier is the right tier for a business that has not yet validated which use cases are worth implementing, or for a business that has tried AI, stalled, and now wants to step back and re-strategise before committing more capital.
What 25k actually buys
At the £25,000 tier, you are buying a roadmap plus one pilot implementation. Typical structure is £8,000 to £12,000 of roadmap work over six to eight weeks, followed immediately by £12,000 to £15,000 of pilot implementation over four to six weeks. The pilot delivers one use case, end to end, with a measurable outcome attached. Total elapsed time is typically ten to fourteen weeks. Total spend lands around £25,000.
The pilot at this tier should not be ambitious in scope. One process, one team, one measurable outcome. Reduce client report turnaround from twelve days to six. Cut admin time on a specific recurring task by half a day a week. Get one client report to a quality threshold without senior review. The constraint is deliberate. A small pilot with a measurable outcome produces evidence the business can use to decide whether to commit more capital. A large pilot without a measurable outcome produces a working system the business is unsure whether to expand.
The £25,000 tier is the right tier for a business that has done some AI experimentation, has a candidate use case it is reasonably confident in, and wants validated traction before committing to a multi-quarter programme.
What 50k actually buys
At the £50,000 tier, you are buying a roadmap, a pilot implementation, and a fractional governance retainer that sits over the programme for three to six months. Structurally, this looks like £8,000 to £12,000 of roadmap, £12,000 to £15,000 of pilot, and £15,000 to £24,000 of fractional CAIO time at £3,000 to £4,000 a month for three to six months. The fractional CAIO oversees sequencing, selects subsequent delivery partners for additional use cases beyond the pilot, and provides governance discipline across the whole programme.
What this tier really buys is programme-level structure rather than one bigger project. A business that committed all £50,000 to delivery without governance would end up with multiple disconnected implementations and no coordination across them. The fractional retainer is what holds the work together, ensures one pilot informs the next, and prevents the common pattern where each project is technically successful and the overall programme produces no compounding value.
The £50,000 tier is the right tier for a business with multiple AI initiatives in motion or planned, lacking senior internal coordination, and needing governance discipline alongside delivery muscle.
The pattern that works: buy small first
The smarter pattern at any tier is to buy the smallest first, then let the work earn the next tier. A £10,000 roadmap that produces clear, defensible recommendations is the strongest possible signal that the consultant is worth committing more to. A £10,000 roadmap that produces vague recommendations is the cheapest possible exit, and worth every penny because it saved you the £40,000 you did not commit.
Committing £50,000 to a consultant you have known for two weeks is the opposite of risk management. The amount of evidence you have on consultant quality two weeks in is not sufficient to justify a £50,000 commitment. The right move is the smaller commitment that produces evidence, then the larger commitment that the evidence has earned.
The pattern that fails is the post-failure overcorrection: paying for a comprehensive programme to “do it properly this time” after a previous engagement disappointed. This pattern produces a deck, an unspent licence, and a second disappointment. The fix is the same fix as the first time. Buy small, let the work earn the next stage.
If you are sitting with a proposal and want help reading what the budget should actually deliver, book a conversation.
