Every consulting practice I have built sold the same thing underneath the offerings: expert hours. AI changes what an hour is worth. The work that used to take a team of five a month now takes two people and a set of agents a week, and clients know it. A practice that keeps billing the hours it no longer needs will lose the work to one that does not.
Here is how I am running a Data & AI practice through that shift, and what I think firm leaders should expect of their practice leaders.
Sell the outcome, not the effort
The unit of sale has to change from the hour to the result. That does not mean guessing at value and hoping; it means naming the business outcome, agreeing how it will be measured, and pricing against it. At Sikich the largest program in the practice, an AI-driven modernization of an insurer’s core applications, is priced on results: change failure rate and recovery time down 30% in the first year, features reaching production faster against the backlog. The client pays for what moved. We carry the risk of delivering it, and we are better at it than they are, which is the whole point of hiring a consultancy.
Time and materials becomes the exception, for work whose scope genuinely cannot be known in advance. Everything else is a fixed-fee offering with a defined outcome or an outcome-priced program.
Productize what you know
Outcome pricing only works when the practice knows, precisely, what it will do. That is what named offerings are for. Eight of ours are strategic, fixed-scope, and fixed-fee: a six-week Data & AI Strategy, an Insights Framework and Data Solution Blueprint, an AI Operating Model and Governance engagement, a Data Governance Operating Model, an MDM Strategy. Each has a defined start, a defined deliverable, and a price the client can approve without a procurement cycle. The productized offering is where the practice’s method lives; the outcome-priced program is where it earns its margin.
When we identify a use case worthy of building, we tie it to a business outcome and we price it based on the value it brings. If we say we can reduce product costs by 15% with predictive machine learning models in the supply chain, then we price it on that result - not the hours needed to build it.
The rule I hold the practice to: offerings before headcount. A named, priced offering earns every hire.
Let agents into the delivery model
If the client’s work is being redesigned around agents, the consultancy’s work must be too. Status reporting, data profiling, documentation, the first draft of a data model: these are agent work now, supervised by a senior practitioner. The pyramid that consulting was built on, many juniors under a few seniors, flattens into a smaller team of experts running agents. The economics improve, and so does the work, because the people on the engagement are the ones who have done it before.
This is also where the talent model changes. The junior years used to be spent doing the work agents now do. A practice that wants senior people in three years has to develop them deliberately: future-leader programs, apprenticeship on real pursuits, promotion on demonstrated judgment rather than hours logged. Experts who grow the next experts.
What firm leaders should ask their practice leaders
Four questions tell you whether a practice is ready for consulting after AI. What share of revenue is fixed-fee or outcome-priced, and is it rising? Which offerings are named, priced, and repeatable? Where are agents doing delivery work today, and who supervises them? And what is the plan for developing the senior practitioners the new pyramid depends on?
A practice leader who cannot answer the four is still selling hours. The clients have already noticed.