A mid-market UK law firm evaluating AI typically has two types of supplier pitching at them: the existing IT consultancy, and a newer breed of AI-specific consultants. Both claim competence on the problem. Neither quite solves it alone. The firms that buy well understand what each actually does, where the overlap ends, and in what order to engage them.
This guide sets out the practical difference between AI consulting and IT consulting for law firms — what each genuinely delivers, where pricing and accountability diverge, and how to sequence both without paying twice for the same work.
IT consultants keep the firm running. AI consultants change what the firm can do with the time IT frees up. Mistaking the first for the second is how law firms end up with a Microsoft Copilot rollout described as a digital transformation.
What IT Consultants Actually Do
An IT consultancy serving a UK law firm is effectively an outsourced infrastructure function. The remit is hardware, software licensing, network reliability, backup, security patching, and help-desk support. At mid-market firms without a dedicated IT director, the consultancy often also owns procurement decisions on case-management platforms, document management, and third-party integrations.
The methodology is operational: ticketed support, quarterly roadmap reviews, project work when a refresh or migration is needed. The economics sit on a monthly managed-service fee plus project-specific charges. The deliverable is an environment that does not go down.
What this skill set does well: keep the lights on, renew licences on time, patch vulnerabilities before regulators ask, and manage migrations competently when the firm changes platforms. What it does poorly: diagnose the gap between the firm's current operating model and what an AI-native firm looks like. That diagnosis is not an IT problem.
What AI Consultants Actually Do
An AI consultancy starts from a different premise. The deliverable is not uptime; it is capability. The engagement assesses which fee-earner workflows are candidates for AI augmentation, what the firm's data maturity will actually support, where the commercial return is, and how adoption should be sequenced. The output is a roadmap, a set of working systems, or both.
The methodology is strategic and implementation-blended. A short diagnostic produces the readiness picture, a roadmap sequences the initiatives, and — increasingly, for agentic AI consultancies — the same team builds the systems against the firm's actual workflows rather than handing over to a separate implementation partner.
The economics reflect the shape. Engagement-based pricing in the five-to-six-figure range, with retainers for ongoing oversight rather than monthly managed-service fees. The deliverable is measurable change in how the firm operates — billable-time recovery, intake conversion, first-pass review speed, matter-opening throughput. Not an invoice for server upgrades.
Why Firms Confuse the Two
Three forces blur the line. First, IT consultancies have added "AI" to their pitch because clients are asking. Second, AI consultancies often need to touch IT infrastructure to deploy, which looks to the firm like IT work. Third, law-firm partners without a technical background cannot easily distinguish "we'll help you roll out Microsoft Copilot" from "we'll redesign your intake workflow with an AI agent."
The test is simple. If the consultancy's proposal describes tools, licences, and deployment, it is IT work. If the proposal describes workflows, outcomes, and capability, it is AI work. Both are legitimate. They are not substitutes.
The Pricing Gap
A managed-service IT contract for a 20-80 fee-earner UK law firm typically sits between £3,000 and £12,000 per month. It covers help desk, infrastructure, and software management. Project work on top — a case-management migration, a Microsoft 365 tenant consolidation — adds one-off costs in the £10,000 to £40,000 range.
AI consulting for a firm of the same size looks different. A readiness audit runs in the £1,500 to £5,000 range. A 12-month roadmap engagement sits between £5,000 and £20,000. Build engagements that deliver working agents — intake, billing assistance, contract triage — typically fall in the £15,000 to £40,000 range for the initial system, with monthly oversight retainers between £750 and £2,500 depending on scope.
Neither is cheap. The point is that the numbers solve different problems. A firm paying its IT consultancy £8,000 per month is not paying for strategic AI advice. A firm paying an AI consultancy £20,000 for a roadmap is not getting help-desk cover. Buying one and expecting the other is the commonest source of "we spent money on AI and nothing happened" complaints.
The Deliverable Gap
An IT deliverable is verifiable by uptime. If email works, the backup ran, the patches are current, and the printer is online, the engagement is functioning. The verification is binary and the cadence is continuous.
An AI deliverable is verifiable by outcome. If the intake agent classifies enquiries with 85%+ accuracy and surfaces qualified matters to partners faster than the previous process, the engagement is working. If fee-earners recover 15-25% more billable time through AI-assisted time capture, the engagement is working. If contract review throughput improves measurably without a drop in quality, the engagement is working. The verification is outcome-based and requires a baseline measurement — which is almost always the first thing missing from firms that have previously tried AI and given up.
The Accountability Gap
IT accountability is to an SLA. Response times, resolution times, availability percentages. The firm knows what "failure" looks like because the SLA defines it.
AI accountability is to outcomes the firm usually has not measured before. If nobody can say what the firm's intake-to-engagement conversion rate was before the AI agent, claims of improvement are vibes. Good AI consultancies build the baseline into the engagement before they build the system. Firms that skip this step buy opaque work that is impossible to evaluate and therefore impossible to defend at partnership level.
How to Sequence Both
Most mid-market UK law firms need both IT and AI support. The right order is:
First, make sure IT is not broken. If case management is unstable, email is unreliable, or document storage is fragmented, AI adoption will fail regardless of how good the consultancy is. AI amplifies the substrate it runs on. Broken substrate produces broken agents.
Second, run an AI readiness assessment separately from the IT provider. The readiness view needs to be honest about the IT landscape, which means it should not be produced by the people responsible for that landscape.
Third, scope the AI engagement. Define the specific workflows, the measurement baseline, the build approach, and the ongoing oversight model. The IT consultancy may execute parts of it under the AI consultancy's direction — that is normal — but the scope is owned by the AI side.
Fourth, keep the contracts separate. An "AI services add-on" inside an existing managed-service contract creates accountability confusion. Separate contracts make success (and failure) legible.
Red Flags: IT Dressed as AI
Several signals indicate that an "AI consulting" pitch is in fact IT work with a rebrand:
The proposal centres on rolling out a named Microsoft, Google, or vendor platform without a prior diagnostic of whether that platform fits the firm's workflows.
The deliverables are licences and configuration rather than measurable workflow change.
The team proposed is the same team that manages the firm's existing IT, just billed at a higher rate.
The success criteria are "platform deployed" rather than "fee-earner capacity recovered" or "intake conversion improved."
The pricing is structured as an uplift to the existing managed-service fee rather than as a defined engagement against defined outcomes.
None of these signals disqualify the supplier from doing the IT rollout. They disqualify the supplier from calling the rollout an AI strategy engagement. Firms that miss this distinction pay AI consultancy prices for IT consultancy work.
What to Buy First
For a mid-market UK law firm that has never done structured work on AI before, the lowest-risk first purchase is an AI readiness audit from a consultancy that does not also hold the firm's IT contract. Small commitment, independent view, clear decision framework on what to do next. Most firms discover that their IT baseline needs tidying up before significant AI work can start — which is useful information regardless of where the firm takes the AI engagement afterwards.
The worst first purchase is a vendor-led pilot of a named AI platform. It looks like progress, it feels like progress, and six months later the firm is back where it started plus a licence renewal it cannot justify.
The Short Version
IT consultants keep the firm running. AI consultants change what the firm can do with the time IT frees up. Mid-market UK law firms need both, but paid separately, scoped separately, and measured against the different things each is actually accountable for. When the two get conflated in one contract, firms end up paying AI prices for IT work — or expecting strategic transformation from a help-desk team.
The sequencing is not complicated. Honest readiness first, separate contracts second, measurement of outcomes throughout. The firms that follow that sequence are the ones whose AI spend actually compounds.