AI Services

This category has been superseded. Forrester now evaluates this market as two separate categories, AI Consulting Services and AI Technical Services, and current buyers should work from those evaluations. What follows is how one category became two.

In 2021 Forrester evaluated thirteen firms and called the category AI Consultancies. The question it answered was straightforward: who can help us work out what to do with AI.

By 2026 that single question had become two, and Forrester was publishing two separate Waves to answer them. One covers the advisory work of deciding what AI is for. The other covers the engineering work of building it.

The vendor lists show why the split had to happen.

What AI services meant

Forrester's framing during the category's single-market period described AI service providers as bringing a holistic approach across strategy, people, process, data, and technology, combining management consulting with technology consulting to drive business outcomes.

Every noun in that definition is doing work, and together they describe a scope no single team inside a client organisation owns. Strategy sits with executives. People and process sit with operations and HR. Data sits with the CDO. Technology sits with engineering.

That breadth was the original selling proposition. An enterprise with no internal AI capability needed one partner who could cover all of it, because coordinating four specialists across four internal owners was harder than the AI problem itself.

The breadth is also why the category eventually could not hold. A firm optimised for board-level strategy and a firm optimised for production data pipelines were being scored against the same criteria, and neither was really competing with the other.

The 2021 field

The Forrester Wave: AI Consultancies, Q1 2021, authored by Michele Goetz, scored thirteen providers against twenty six criteria and was addressed to CIOs and CTOs: Accenture, Boston Consulting Group, Capgemini, Cognizant, Deloitte, EY, HCL Technologies, Infosys, KPMG, McKinsey & Company, PwC, Tredence, and Wipro.

Read that as three groups sitting uncomfortably together. The strategy houses, McKinsey and BCG. The Big Four accounting and advisory firms. And the technology services firms with large offshore delivery capability, Cognizant, HCL, Infosys, and Wipro, plus the specialist Tredence.

Those three groups sell different things to different buyers with different economics. In 2021 they were all answering the same question because enterprises did not yet know enough about AI to ask a more specific one.

The regional editions

Forrester also published The Forrester Wave: AI Consultancies In Asia Pacific, Q2 2021, authored by Achim Granzen, scoring thirteen providers against the same twenty six criteria: Accenture, BCG, Capgemini, Cognizant, DXC Technology, EY, IBM, iSoftStone, KPMG, NEC, PwC, ST Engineering, and Tata Consultancy Services.

Same count, same criteria, and five names that do not appear in the global edition. iSoftStone, NEC, and ST Engineering are serious regional players with no meaningful presence in a global shortlist, and their absence from the worldwide Wave says nothing about their capability in their own markets.

That is a useful general lesson about analyst research. A global evaluation describes the firms competing for global engagements, not the firms best suited to a specific region, and buyers with regional footprints should check whether a regional edition exists before assuming the global list is complete.

The market matured

The AI Service Providers Landscape, Q3 2022 marked the point where Forrester started mapping a wider field rather than scoring a shortlist. The reason was adoption. Forrester's research at the time found sixty seven percent of data and analytics technology decision-makers reporting that their organisation had adopted AI, with the majority of them looking for partners to help scale to production and realise business results.

That phrasing captures the shift precisely. The 2021 question was whether to do AI. The 2022 question was how to get from a working model to something running in the business, which is a different problem requiring different skills.

The Landscape also reflected a market that had broadened well beyond thirteen large firms, with providers ranging from small to large and from generalist to domain-specific. Once a market contains both McKinsey and a twelve-person specialist in pharmaceutical clinical data, a single scored evaluation stops describing it.

The Forrester Wave: AI Services, Q2 2024

By 2024 the category had been renamed from consultancies to services, which was an accurate reflection of what buyers were purchasing. Consultancy implies advice. Services implies delivery, and by 2024 the work included building things.

Capgemini placed as a Leader in that edition. Its own framing at the time is a fair description of where the market sat, emphasising generative AI as a catalyst, ethical governance and organisational frameworks as foundations, and delivery at controlled cost.

That combination, generative AI plus governance plus cost control, marks a market past the experimentation phase. Nobody asks about controlled cost while running pilots.

The split, and how it sorted the field

For 2026, Forrester separated the market into AI Consulting Services and AI Technical Services, with distinct evaluations, distinct criteria, and distinct vendor sets.

The consulting evaluation covers the advisory work: use case identification, business case, operating model, governance, and value measurement. The technical evaluation covers data pipelines, model deployment, MLOps, and lifecycle management, with roughly eleven capabilities falling on that side of the line.

Compare the 2021 field against the 2026 AI Consulting Services field, which covers Accenture, Bain, BCG, Capgemini, Deloitte, EY, IBM, KPMG, McKinsey, and PwC.

Cognizant, HCL, Infosys, Wipro, and Tredence are not in it. Bain and IBM are.

The split sorted the field along exactly the fault line that was visible in the 2021 list. The strategy houses and Big Four firms stayed in consulting. The technology services firms with large delivery organisations moved to the side of the line where their strength actually is.

Nobody was demoted. The category was redrawn so that firms compete against comparable firms, which is more useful for a buyer than a single ranking that mixed advisory reputation with delivery capability.

What this means if you are buying now

Three things follow from the history.

First, check which evaluation you are reading. A Leader in AI Consulting Services is not necessarily strong in AI Technical Services, and vice versa. Vendors cite whichever placement flatters them, and both are legitimate.

Second, the handoff between the two is where programmes fail. Forrester's reference feedback across the 2026 consulting evaluation found that the most consistent request from customers was continuity of team from strategy through implementation. Not better strategy. Not cheaper delivery. The same people staying on the job across the boundary the category split created.

That is a slightly awkward result. The market divided because the two disciplines are genuinely different, and customers immediately asked for them not to feel divided. Both things are true, and the practical response is to make the handoff a contractual matter rather than a hope.

Third, older research is still useful for a specific purpose. The 2021 Waves and the 2022 Landscape are stale on scores and sound on structure. They tell you which firms have been serious about this market for five years rather than five quarters, which is worth knowing when every firm now has an AI practice.

Forrester does not endorse any provider named here, and tier placement should not be read as a recommendation to buy.