The market for AI-native consulting: We've been doing it for two years. Now the chase is on.
Also published on LinkedIn
For the past two years we have been running AI-native consulting with our clients: mapping their work task by task, letting agents do the heavy lifting, putting a consultant's signature on every finding, and measuring what lands in hours. We have watched the shift happen from inside real transformations, long before it had a name. This month the rest of the market said it out loud. The smart firms are already moving, and we are seeing it first-hand: consultancies are now co-leading live deals with us, running their practice on our platform rather than building their own.
This is what two years in the field has taught us, why October matters, and where it leaves everyone who does this work.
What we have seen change
When we started, buyers asked for an AI strategy. Now they ask a harder question: show me where the work changes, and prove it landed. Four things have shifted in the rooms we work in.
The diagnostic stopped being the product. Clients no longer want a twelve-week study and a deck. They want a model of how their organisation actually works that is still accurate next quarter. In the US, 3.3% of the workforce separates every month, so a third of roles change hands in a year. A snapshot is stale before it is presented.
Pilots stall until the work is mapped. Almost every stalled AI programme we have walked into had the same root cause: nobody could say which tasks existed, who did them or what they took. Gartner puts organisations with reliable skills data at 8%. Kyndryl's 2026 People Readiness Report has leaders who say their workforce is ready for AI at 23%, down six points in a year.
Buyers want hours, not outputs. Self-reported time savings no longer survive a CFO. The conversation has moved to capacity redeployed, measured from live systems against a baseline set before rollout.
The consultant's job moved to the gate. Agents now do the ingestion, task decomposition and scenario modelling in weeks. The consultant's value sits where it always should have: judgement, the decision with leadership, and the signature on the finding.
October made it official
This month the biggest names in enterprise software and consulting confirmed the same direction.
SAP agreed to acquire TechWolf, whose product is a continuously updated model of the work inside an organisation: the tasks inside each role, the skills people apply and the labour market around them. SAP described it as a grounding layer for its workforce agents. A week later Workday launched Work Intelligence, which breaks roles into tasks and ties them to skills. Deel has turned its internal agent platform into a product, and Remote has opened its workforce data to AI agents.
The consultancies are restructuring around the same insight. Accenture has folded its service lines into a single Reinvention Services unit. BCG says more of its largest AI work is now paid on results. IBM calls its model services-as-software, and its consulting team describes the new delivery model as forward-deployed. Kyndryl's 2026 State of Modernization Report finds 44% of organisations now treat modernisation as continuous, against 16% taking a big-bang approach. And in an HFS and IBM study, 65% of buyers said traditional consulting fails, with headcount-based contracts expected to fall from 49% to 16% within two years.
Meanwhile Google and Meta both shipped new frontier models in the same fortnight. The models keep improving and keep converging. What separates outcomes is not the model. It is how well an organisation knows its own work.
None of this tells you who wins. It tells you the category is real: the record of the work is now recognised as strategic, and delivering transformation without one is the old way.
What AI-native consulting is
It is not a consultancy with copilots. Every firm has those. From what we have built and delivered, it passes four tests.
1. The map is the deliverable, and the client keeps it. A verified, task-level model of how the organisation works, covering roles, tasks, processes, tools and vendors, with every fact carrying where it came from. The engagement ends. The map does not.
2. Agents do the lifting. Consultants hold the gate. A consultant signs every finding, and committing a change is a human decision, recorded with its rationale. The human gate is not a nice-to-have. It is the product.
3. Continuous, not a snapshot. A quarterly impact assessment on a live model of the work replaces a job architecture refresh every two years.
4. Measured in hours that land. Capacity redeployed, against a baseline set before rollout. McKinsey finds AI high performers are nearly three times as likely as everyone else to have fundamentally redesigned their workflows. Redesign, then measure.
The smart firms are already moving
Here is the part we did not expect to see so early. The consultancies leaning in are not trying to build all of this themselves. They are partnering.
Our partner programme is simple. We bring the platform and the deal. Partner firms bring the consulting. Regional consultancies and boutique firms co-lead live deals in one of two named roles:
- Transformation Partner, who owns the decision: senior operators who have run a transformation and been accountable for it. They place the organisation, run every decision gate with leadership and sign every finding.
- Work Architect, who owns the build: process redesigners measured on whether it worked. They verify, redesign, build and track hours returned every quarter.
The client sees one team. Our partners run the room, we run the platform. Discovery happens together, and from the decision gates on, it is the partner's name on the work. They invoice their consulting directly, at their own rates, and they never have to hire an engineering team to compete with firms a hundred times their size.
That is why the smart ones are moving now. The work is shifting from billed hours to reusable assets and measured outcomes, and a boutique cannot build a work graph, an agent layer and a measurement engine on the side of a consulting practice. They can run on one.
What to do this quarter
If you lead transformation inside an organisation:
- Map one function task by task before you buy another agent for it.
- Set the baseline in hours now, before the next rollout, so your CFO has something real to measure against.
- Rewrite your next consulting brief: the deliverable is the map and the measurement, not the deck.
If you run a consulting practice:
- Decide what you sell when the data gathering is free: judgement, decisions and accountability.
- Stop rebuilding the client picture on every engagement. Run your practice on one that compounds.
- Talk to us about becoming a Transformation Partner or Work Architect in your region.
We have spent two years learning what this looks like in practice. The market has caught up with the idea. The firms and organisations that move now will be the ones with two years of learning in hand when everyone else starts.
Sources: SAP News, SAP to acquire TechWolf, 6 October 2026; Workday, Workday introduces Work Intelligence, 13 October 2026; Google, Gemini API release notes, 13 October 2026; TechCrunch on Meta Muse Wonder, 12 October 2026; Deel and CXOToday on Akai, 2026; Remote, PR Newswire, 14 May 2026; MIT Sloan Management Review Middle East on Accenture Reinvention Services; BCG, AI is changing consulting and pricing models, 2026; CIO Dive on IBM Consulting, 2026; Kyndryl, 2026 State of Modernization Report and 2026 People Readiness Report; HFS Research and IBM, November 2025; Gartner, September 2024; US Bureau of Labor Statistics, JOLTS 2024; McKinsey, The State of AI, 2026.