IT Consulting Trends for 2027: 5 Shifts Already Underway
For 30 years, an IT consultancy sold time, by the day or by the head. In 2026, the largest firms started pricing something else, and their clients started asking for it before their contracts came up for renewal.
The 5 shifts below already show in signed contracts and in hiring plans, and one of them is written into EU law. Each reaches data projects early, because data work is where AI has made the build fastest to produce and easiest to check. For each shift, what it means for a consultancy that delivers data work.
1. The day stops being the unit of sale
In November 2025, a McKinsey partner said that about a quarter of the firm's global fees came from outcome-based arrangements. In August 2026, the chief executive of TCS said that about 80% of the contracts in its finance, HR and other business services were now outcome-based.
Oliver Wyman put the arithmetic plainly in November 2026: under time and materials, a 25% cut in effort can turn into a 25% smaller invoice. The firm expects staff augmentation to fall from about 28% of the IT services market in 2025 to about 20% in 2030.
For a data consultancy, the change starts with the build. When a connector that took 8 days takes an afternoon, cutting the quote by the same 80% prices the work below cost, because most of those days were never spent typing. What remains billable is the qualification of the source and the acceptance of the result, and both can be sold as fixed deliverables. The day will survive as an internal measure of effort, while the daily-rate model loses its place on the invoice.
2. Clients reopen contracts before they end
In June 2026, HFS reported that clients were renegotiating IT services contracts within 24 months of signing them, instead of waiting for renewal, and moving them towards outcome-based pricing and gain-sharing.
In August, Reuters described the same pressure on Indian IT services firms. The chief executive of Persistent said clients wanted the same work for 25% to 30% less. HCLTech agreed to go unpaid for the first year of a cloud contract with E.ON, with payment from the second year tied to efficiency gains, according to people familiar with the deal. Infosys told analysts it had walked away from contracts that no longer made economic sense.
The message for a consultancy is that the client already counts the AI saving as its own. The only way to keep a share of it is to show where it came from, and that requires a baseline: how long a source took to deliver before, and what its rework cost. A team that has estimated ingestion work in units rather than in tasks has that baseline. A team that billed days has only its invoices.
3. The pyramid loses its base
In August 2026, PwC's US leader said the firm had hired fewer consultants and accountants than 2 years earlier, and expected the trend to continue as it hired more data scientists and engineers. The same Oliver Wyman analysis shows the market buying fewer people by the head.
The work at the base of the pyramid is the work AI took first. In a data programme, that was the junior engineer writing the 40th connector, learning the trade one source at a time.
Margin moves to the people who can say what a source really contains, and who can sign that a deliverable is right when nobody on the team wrote it. That changes how a consultancy staffs a project, and it raises a question few firms have answered yet: if juniors no longer learn on connectors, where will they learn?
4. Reusable assets become the margin
Oliver Wyman names reusable assets as one of the ways a provider keeps the value of AI: an asset can carry its own fee, whatever the hours behind it. In August 2026, Erik Brynjolfsson and Georgios Petropoulos of the Stanford Digital Economy Lab published a pricing framework for consulting in the age of AI agents, and they expect AI to push firms towards hybrid and asset-based models. This is the asset side of the AI disruption in consulting.
In a data consultancy, the asset is rarely a product. It is the qualification grid that turns a client's description of a source into a fixed price, and the acceptance tests that decide when a data pipeline is production-ready.
These assets compound only when the consultancy keeps them in a form it controls. Built inside one vendor's interface, they follow that vendor's roadmap. Kept as code and configuration in repositories, they travel from one client to the next, and from one platform to the next.
5. Portability becomes billable work
From 12 January 2027, the EU Data Act stops providers of cloud and data processing services from charging their customers for switching to another provider. In financial services, DORA already requires exit plans for critical ICT services, and those plans must be "sufficiently tested". Meanwhile, the platforms clients depend on keep changing hands: within a year, Informatica, Osmos, Confluent and Dremio were acquired, and Fivetran and dbt Labs merged.
The law removes the exit fee and leaves the work of leaving where it was, in the pipelines, because vendor lock-in is built one connector at a time. Clients will ask their consultancy to prove they could leave: inventory what runs where, then run a pipeline outside its platform and show that it still passes its tests.
That work goes to whoever the client trusts to be neutral. As the platforms consolidate, vendor neutrality becomes scarce, and a consultancy that delivers pipelines its client owns and can move is the natural candidate.
What the 5 have in common
Each shift moves the invoice away from time spent and towards something the client can check: an outcome, a measured saving, a senior judgement, an asset, an exit.
For a consultancy that delivers data work, the thing the client checks is the pipeline, with its tests and the record of who accepted it. The firms that keep those in a form the client owns will find the 2027 contracts easier to write.
The same year, seen from the data team's side, is in our data engineering trends for 2027.
PrettyWhale.ai is an AI solution for data engineering. It generates complete data pipelines, with their tests and documentation, so the build can be quoted against a qualified source and the client keeps what it paid for.
You have any questions on PrettyWhale.ai ?