How do AI consultancies price projects in 2026?

AI consultancies in 2026 price projects using five models in active use: fixed-fee by deliverable, tiered ladder, value-based, hybrid (fixed PoC then time-and-materials), and hourly. Which model a consultancy uses tells you more about their business than their slide deck does — fixed-fee consultancies sell outcomes, hourly consultancies sell time, value-based consultancies sell a story.

Quick answer: Five pricing models are in use. Boutiques default to fixed-fee. Enterprise consultancies default to hybrid (fixed PoC then T&M). SME firms run tiered ladders. Hourly billing is dying at the boutique tier because AI compresses task time.

The five pricing models in 2026

1. Fixed-fee by deliverable

The buyer pays a fixed price for a defined outcome — an audit, a prototype, a built system. Scope is agreed up front; anything outside scope is a change order or a new engagement.

Works for: discrete, well-defined deliverables. AI audits, scoped builds, MVP prototypes, pSEO programmes, marketing system builds.

Pros: Budget certainty for the buyer. The consultancy is incentivised to ship efficiently. AI productivity gains accrue to the consultancy’s margin — which is the right outcome, because otherwise nobody invests in the tooling.

Cons: Requires real scoping work before signature, which means paid discovery or a slower sales cycle. Discourages mid-stream pivots — you cannot easily change direction without a new engagement.

Typical users: Formulaic, Tectome, most independent boutiques.

2. Tiered ladder

The consultancy publishes three or four packaged tiers (good / better / best), each with a fixed scope and price. The buyer picks the tier. Often combined with productised delivery — same template applied to each client.

Works for: SME-focused work where the buyer values speed and clarity over bespoke scope. Common in marketing AI, sales AI, and “starter pack” automation work.

Pros: Fast to buy. Easy to compare to alternatives. Lower friction sales cycle.

Cons: Almost certainly oversells some buyers and undersells others. Tiers force the buyer’s reality into the consultancy’s template. Inflexibility on edge cases.

Typical users: Elevate AI, Ampliflow, the broad SME tier.

3. Value-based pricing

The consultancy charges a percentage of attributable savings or revenue generated. Variants include performance bonuses on top of a base fee, or pure outcome-based deals.

Works for: narrow, measurable outcomes. Lead-gen systems where revenue is attributable. Cost-reduction automations where baseline is well-documented and savings can be verified.

Pros: Maximum alignment between buyer and consultancy. The consultancy only wins when the buyer wins.

Cons: Measuring attribution is hard. Most “value-based” deals end up being fixed-fee deals with a value-justification narrative — because the alternative is litigation over what was actually saved. Operationally, you need clean baseline data and an agreed measurement framework before the engagement starts. That is rare.

Typical users: Performance marketing agencies, some growth consultancies, occasional specialist boutiques.

4. Hybrid (fixed discovery + T&M build)

The consultancy charges a fixed price for discovery or proof-of-concept (typically £25,000–£150,000), then bills the build phase as time-and-materials or as a dedicated-team engagement (typically £30,000–£80,000 per month per team).

Works for: large, multi-year transformation programmes where the build scope cannot be specified up front. Enterprise programmes touching multiple business units, multiple systems, multiple stakeholders.

Pros: Realistic for genuinely large scope. Lets the buyer evaluate the consultancy on the PoC before committing to a long build. Allows mid-stream pivots.

Cons: Budget uncertainty for the buyer once the PoC is signed off. Incentive misalignment in the T&M phase — the consultancy is paid more for taking longer. Requires sophisticated buyer-side project management.

Typical users: Faculty AI, Neurons Lab, Big Four AI practices, BCG / Bain / McKinsey AI groups.

5. Hourly or day rate

The consultancy bills hours or days at an agreed rate. Buyer pays for time consumed.

Works for: advisory engagements, fractional CTO / Head of AI roles, defined low-scope research projects, and circumstances where the buyer genuinely wants control over which hours are spent on what.

Pros: Maximum buyer flexibility. The buyer can stop or redirect at any point. No incentive for the consultancy to under-scope a fixed engagement.

Cons: Penalises consultancies for getting faster — which in 2026 means the consultancy has to actively avoid efficiency gains to maintain revenue. Reverse-aligned incentives in any project longer than a few weeks. Buyer carries full project management overhead.

Typical users: Independent specialists, advisory-only consultancies, some legacy enterprise consultancies that have not migrated to fixed-fee.

Which model fits which engagement

Engagement type	Best-fit pricing model
AI audit / readiness assessment	Fixed-fee
Proof-of-concept on real data	Fixed-fee
Scoped build of a defined system	Fixed-fee, milestone-paid
Multi-system implementation programme	Hybrid (fixed discovery + dedicated team)
Enterprise AI transformation	Hybrid or T&M with milestones
Marketing AI starter package	Tiered ladder
Ongoing AI partner / embedded delivery	Fixed monthly retainer
Lead-gen system with revenue attribution	Fixed-fee + performance bonus
Fractional advisory	Day rate or fixed monthly retainer

2026 benchmark ranges by tier

These are indicative ranges from published prices and what UK and US buyers report paying in 2026. Treat as a sense-check, not a quote.

Tier	AI audit	PoC / Sprint	Scoped build	Embedded retainer (per month)
SME	£500–£5,000	£5,000–£15,000	£10,000–£40,000	£2,000–£6,000
Boutique	£3,000–£15,000	£15,000–£25,000	£25,000–£150,000	£8,000–£20,000
Premium boutique	£15,000–£50,000	£40,000–£100,000	£100,000–£500,000	£25,000–£60,000
Enterprise	£50,000–£200,000+	£100,000+	£500,000–£5m+	£50,000–£200,000+

Formulaic sits at the boutique tier: audit £3,500, sprint £15,000, scoped build £25k–£150k, partnership from £8,000/month. See pricing.

Red flags in AI consultancy proposals

Six signals worth taking seriously when evaluating a proposal.

1. Refusal to publish indicative pricing. “Pricing depends on scope” is true at the extremes — a £500 single-workflow audit and a £5m enterprise programme genuinely cannot share a number. But a consultancy that cannot tell you their audit price, their typical build entry point, or their day rate is asking you to negotiate from a position of no information. That is by design.

2. Free strategic work as a sales pitch. Free audits, free roadmaps, free architecture reviews. These are pre-sales investments that get amortised into the eventual engagement. You are not getting a free audit — you are getting a sales pitch and paying for it in the build fee.

3. Hourly billing with senior consultants on multi-week projects. Misaligns incentives. The consultancy is paid more for taking longer. The senior consultant whose AI tooling makes them 3x faster is now billing 3x fewer hours — they will not adopt the tooling.

4. Open-ended T&M contracts without milestone gates. Hybrid pricing is fine, but the T&M phase needs phase-gates with go/no-go decisions. Without them, the engagement runs until the budget runs out — which is often the consultancy’s revenue target, not your project’s natural endpoint.

5. Bundled infrastructure with no breakdown. A proposal that bundles LLM API costs, cloud hosting, and third-party tooling into the fee is hiding markup. Ask for the breakdown. Compare the answer to what the contract actually allows.

6. Lock-in language in the contract. Watch for: ownership of derivative IP staying with the consultancy, non-transferable source code, mandatory hosting on the consultancy’s infrastructure, exclusivity clauses on AI vendors. Each of these turns a build engagement into a long-term dependency.

Methodology vs price tier

Pricing model and price tier are independent variables. A consultancy can be SME-tier with a fixed-fee model, or enterprise-tier with hybrid. The matrix below shows the typical combinations in 2026.

	SME tier	Boutique	Premium boutique	Enterprise
Fixed-fee	Common	Default	Common	Rare
Ladder	Default	Occasional	Rare	Rare
Value-based	Rare	Occasional	Occasional	Rare
Hybrid	Rare	Occasional	Common	Default
Hourly	Occasional	Dying	Rare	Occasional (advisory)

The pattern is clear. Fixed-fee dominates at the boutique tier; hybrid dominates at enterprise; SME firms favour ladders. Hourly is in retreat everywhere except pure advisory.

What to do with this

If you are evaluating proposals from multiple consultancies, normalise on scope per pound, not headline price. Two £25,000 builds at different consultancies can mean very different deliverables. Ask each one:

• What is the deliverable?
• What is the timeline?
• What is included vs change-order?
• What is the infrastructure cost treatment?
• What happens at the end — what do I own?

If the answers are clear and comparable, you can pick on price. If they are not, you are not comparing the same product.

If you want to see how Formulaic’s pricing works in detail, see the pricing page — four tiers, all published, all fixed-fee. If you want to start with the cheapest first step, take the free AI Opportunity Scorecard. Three minutes. Or book a 30-minute call and we will tell you which tier fits your firm honestly, including when the answer is “not us”.