TL;DR:

  • 26 of the 38 firms in our Pricing Benchmark sell training. The 14 that also sell implementation sell a different product than the 12 that don't, under the same name.

  • Implementers sell training short and senior: one-day workshops, often for groups of 1-5, and nearly all training retainers are theirs.

  • Specialists sell training long and wide: multi-week programmes for classroom-sized cohorts, and they price accordingly.

  • A cheap workshop that converts into a £100k build is customer acquisition, not a training business. Know which one you're selling, and which one you're competing against.

Took the survey but didn't leave your email? Reply to this issue and we'll send you the full results. We work on the honour system: if you claim a copy you didn't earn, we'll leave it to karma, which in this industry usually arrives as an unusually thorough procurement officer on your next big deal.

-Daria

💰 Market Monitor (AI services M&A + Investment)

Notable M&A

(IN/US, AI implementation) NewVision Software has been acquired by SoftServe. The India-based firm, specialising in agentic assurance and intelligent managed services, adds 700+ engineers and a growing Global Capability Centre footprint to SoftServe's AI transformation practice. [Link]

(US, B2B marketing) Sperling has been acquired by Arketi Group. The Massachusetts-based digital agency adds AI capabilities - chatbots, AI agents and generative search optimisation - to Arketi's integrated B2B marketing and communications platform. [Link]

(US, technology advisory) OneConnect has been acquired by Amplix, its 15th acquisition since 2022. The Florida-based technology procurement specialist deepens Amplix's advisory and lifecycle management capabilities as it pushes into services- and AI-driven sales. [Link]

(US, AI marketing) AdSkate has been acquired by Brunner. The Pittsburgh-based AI creative analytics platform - which predicts ad performance before launch using synthetic audience modelling - bolts onto Brunner's integrated marketing stack. [Link]

Notable Investments

(DE, AI-native professional services) Skalar has raised a €12m pre-seed and seed led by Headline. Founded by Björn Goß (who sold Stocard to Klarna for €110m), Skalar is building a regulated tax and accounting firm with AI-first operations - targeting a model where one expert serves 100+ clients. The "AI-native professional services firm" archetype, in a regulated European context. [Link]

(US, business transformation) SaaS Consulting Group has received a strategic investment from Caltius Equity Partners. The mid-market business transformation consultancy will use the backing to expand its AI and data services practices and pursue acquisitions - the PE-buys-services-to-add-AI pattern at work. [Link]

Have we missed an interesting deal? Or do you want to correct details? Just hit reply to let us know.

🙌 Partner Program Updates (models + platforms)

OpenAI: The Partner Network is now officially live, opening a three-tier structure to SIs, MSPs and consultants, with Google Cloud's former partner chief Philip Larson hired to run it. [Link] Eliza, an AI-native consulting firm built on the forward-deployed engineer model, was named an Advanced Partner in the first visible tier allocation. [Link] PwC US launched agentic contact and service solutions with OpenAI, including a joint Center of Excellence. [Link]

Anthropic: Ode with Anthropic formally launched this week - the $1.5bn enterprise AI services firm first announced in May as a JV between Anthropic, Blackstone and Hellman & Friedman. [Link] UST is embedding Claude into its CarePath healthcare platform, including work with Optum, and training 20,000 developers. [Link] L&T Technology Services is integrating Claude across its engineering platforms spanning semiconductor, telecom and mobility. [Link] FIS extended its partnership and is an early user of Mythos 5 via Project Glasswing. [Link] Origin Digital joined the Claude Partner Network. [Link] Deloitte launched a Claude-powered software supply-chain security platform and unveiled agentic intelligence within its Omnia audit platform. [Link]

Mistral: Global AI deployment firm CI&T signed a multi-year partnership with Mistral to build agentic enterprise solutions, serving as Mistral's preferred partner in Latin America - sovereignty-first positioning travelling beyond Europe. [Link]

Databricks: Austin-based data strategy consultancy DX Foundation became an official Databricks partner, with certified implementation coverage across five service areas including Mosaic AI. [Link]

Elsewhere: TCS plans a force of up to 8,900 forward-deployed AI engineers embedded with clients, the largest disclosed FDE cohort in the industry, while actively evaluating AI acquisitions. [Link] Accenture launched Accenture Edge, a midmarket unit that will lean on acquisitions and run acquired firms as independent, entrepreneurial companies. [Link] Serviceplan's Plan.Net launched a specialist business unit combining AI strategy, implementation and operations for enterprise marketing transformation. [Link]

Did we miss a partner program update from any of the major players which you feel is important?

The operator essay: How much you charge says a lot about who you sell to

Two firms both sell "AI training". One charges £200 a head and the room holds thirty people. The other charges £1,000 or more a head and the room holds four, usually the exec team. Same word on the invoice, different product, different buyer, different business model. And in our survey data, you can tell which room a firm sells into from one fact about them: whether they also sell implementation.

This is the first data drop from our 2026 AI Enablement Services Pricing Benchmark, a survey of AI Enablement services firms founders. The full results are available to those who participated. Limited drops like this one will be published here in AI Enablement Insider. Want the full report next time? Subscribe and keep your eyes out for the announcement of the next wave.

Of the 38 firms who have participated, 26 offer training. Fourteen of them also offer implementation; let's call them full-stack firms. Twelve do not; let's call them specialists, offering training and consulting only. Nearly everything about how these two groups sell training flips depending on which side of that line they sit: the format, the audience, the pricing model and the price.

The workshop and the programme

Full-stack firms sell training short. Nine of fourteen run workshops of a day or less, six run two-to-five-day courses, and only four offer a multi-week programme. Specialists sell it deep: seven of twelve run multi-week programmes and three go multi-month.

The sharpest tell is "ongoing continuous access". Four full-stack firms offer it. Zero specialists do. When you also sell implementation, training becomes a persistent line item that sits alongside the build. When training is the whole engagement, it has a start and an end.

The formats themselves are getting more inventive. Three firms in our survey run build-your-own-agent sessions where participants leave with a working assistant rather than pdfs, and two run train-the-trainer programmes explicitly designed to make themselves redundant. We see founders of AI training firms continuously innovating on format, creating products that simply did not exist in corporate training before. It is part of what makes this category exciting to be part of.

"Participants build their own personal AI assistant during the session ... attendees walk away with a working agent they use immediately."

— Founder, AI training & development co

Who's in the room

Six of fourteen full-stack firms typically train groups of one to five people. Exactly one specialist does. Specialists cluster at classroom size: six of twelve typically train groups of 13 to 30, five train groups of 6 to 12.

A group of four is an executive briefing, not a training cohort. The implementer's training is often about getting the leadership team comfortable and winning the build that follows. The specialist's training is a classroom, because the upskilled workforce is the deliverable. This is what the price per head is really telling you: it is a proxy for who is in the room, and who is in the room tells you what the firm is actually selling.

"Managers specifically leave able to lead AI adoption in their own teams rather than delegate it upward."

— Founder, AI adoption training co

How it is priced

Seven of the 26 training firms sell training on a monthly retainer or subscription. Six of the seven also sell implementation; among specialists, exactly one does. Across the wider survey, training is the least recurring service line: 27% of training firms use retainers, against 38% in consulting and 39% in implementation. Unless you are an implementer, in which case training-as-subscription is the glue that keeps the relationship warm between builds.

Per-learner pricing flows the other way: half of specialists price per head, against just over a third of full-stack firms. Per-head is product pricing. You use it when training is the product.

The specialist premium

Among the eleven firms that price per learner, four of the six specialists charge £1,000 or more per head. None of the five full-stack firms does, and two of them sit at £100-300. The premium follows depth: three of the four £1,000+ firms run multi-week or multi-month programmes, while sub-£300 prices concentrate in one-day and short-course formats. The fourth £1,000+ firm runs one-day sessions only, which sharpens the rule rather than breaking it: price follows depth, except at the top, where it can follow seniority instead.

The same story plays out at cohort level. The only £15k+ per-cohort prices in the survey belong to specialists, and the £2,000+ training day rates cluster with them too. Firms that treat training as a funnel price it like marketing. Firms that treat training as the product price it like a product, at roughly three times more per head.

Why this happens

The economics are rational on both sides. A £5k workshop that upsells into a £100k implementation needs conversion more than margin. Underpricing training is customer acquisition cost. The specialist has no downstream build to subsidise the classroom, so the training must carry the firm, and that is why it gets depth, duration and a real price.

Services are by nature not recurring, but the best companies find a way to make them "re-occurring". That means having more than one offering and being able to upsell an existing customer, improving the ratio of acquisition cost to lifetime value. We can see pure-play training firms adding new offerings: large levy apprenticeship providers like Multiverse (we covered them here) are adding shorter courses and moving into non-levy-funded corporate training. While every training product offered by specialists needs to cover its cost and be profitable, those who offer additional training products will have more margin to spare to invest back into their business.

What to do with this

If you are full-stack: the danger is not the cheap workshop, it is letting funnel pricing leak into your flagship training. Separate the SKUs: a priced exec briefing you are happy to discount strategically, and a per-head programme priced against the specialists, not against your own lead-gen.

If you are a specialist: when you lose a training deal to a £5k workshop, you are not losing to a competitor, you are losing to another firm's marketing budget. Their workshop does not have to make money; yours does. So do not compete on their terms. Put the question in your buyer's mouth: ask them what happens after the workshop ends. Sell the thing funnel-priced training structurally cannot deliver, which is capability that is still there when the vendor leaves. And if the economics of the exec briefing genuinely do not work for you, stop fighting for it: being the training partner implementers subcontract beats being the one they undercut.

Your buyers are starting to work this distinction out. It's better if they hear it from you.

The usual caveats: 26 firms, split 14 and 12; the per-learner comparison covers 11 firms. Read these as patterns in our respondent set, not market statistics. The sample skews UK and sub-50 headcount.

The full 2026 AI Enablement Services Pricing Benchmark, 35 pages covering pricing bands across training, consulting and implementation, pricing discipline and exploratory cross-cuts, has gone to everyone who took part. The next wave will be announced here in AI Enablement Insider. Participation earns the report.

Reply

Avatar

or to participate

Keep Reading