AI will take hours out of most professional services projects, but will shrink professional services (PS) only for organizations that keep selling those hours. Where pricing follows deliverables or outcomes, the same efficiency becomes margin.

This post is about professional services inside technology companies: the teams that implement, integrate, and optimize what the company sells.

Below, we look at where AI removes hours, why pricing decides what happens next, where new demand is coming from, and what PS leaders should do this year.

Key Takeaways

• The definition matters: In technology companies, professional services teams implement, integrate, and optimize the company’s products for its customers.
• Hours are falling: AI cuts the labor in many PS projects, which cuts revenue wherever work is priced by the hour.
• Demand is shifting: AI products also create new services work, from readiness and governance to forward-deployed engineering and ongoing optimization.
• Pricing decides the outcome: Whether AI shrinks a PS organization depends mostly on how it charges for its work.
• PS 2.0 is the response: TSIA’s PS 2.0 research describes how PS organizations are changing delivery and pricing for this shift.

Benchmark Your PS Organization

What AI Does to the Hours in a PS Project

Broader professional services industry trends show how fast this is moving. Thomson Reuters’ 2026 AI in Professional Services report found that organization-wide AI adoption nearly doubled, from 22% to 40%. While research primarily covers legal, tax, and accounting firms, for technology companies, the more pertinent question is what AI in professional services does inside a typical implementation.

A standard deployment follows a consistent pattern:

• Discovery and requirements: AI can summarize existing documentation, draft current-state assessments, and turn workshop notes into requirements. The time-consuming first pass shrinks.
• Configuration and code: Generated configuration, scripts, and integration code reduce build time, especially for repeatable patterns.
• Data migration: Mapping, cleansing, and validation scripts take less manual effort, though data quality problems still need people.
• Testing and documentation: Test cases, test data, and user documentation can be drafted from the build itself.

The hours that remain are concentrated in thought work: scoping decisions, design trade-offs, stakeholder alignment, and fixing breaks in the customer’s environment.

This has a direct effect on billable utilization. If the same project needs fewer hours and nothing replaces them, either utilization falls or the team has to shrink to match. TSIA’s work on forward-deployed engineering adds a further complication. Some of the most valuable work in AI projects, such as turning a custom solution into a reusable blueprint, doesn’t look efficient under a utilization model.

The Solution Depends on How You Price

AI’s effect on PS revenue runs through the pricing model.

• Hourly (time and materials): Every hour AI removes is revenue removed. Delivery gets faster, the invoice gets smaller, and the PS organization shrinks in direct proportion to its own productivity gains.
• Fixed fee: The provider keeps the savings, for a while. At the next negotiation, buyers who know AI shortened the work push the price down.
• Outcome-aligned: The fee follows the result. When AI makes delivery cheaper, the difference becomes margin, and the provider has a reason to keep investing in efficiency.

That makes the professional services business model the real variable. Few PS organizations will move straight to pricing on outcomes, and they shouldn’t try. In TSIA’s webinar with Kantata on outcome-based pricing, the starting point was that most services organizations are not jumping to pure outcome-based pricing. They’re adopting hybrid models first, with a guaranteed base and a variable share tied to results. The same session stressed that delivery has to be standardized and measurable before outcomes can be priced.

For more on the mechanics of building those offers, see our guide to outcome-based pricing.

Where New PS Demand Is Coming From

The shrinking-hours story is only half the picture. AI products create services work that didn’t exist before, and much of it falls to PS.

The reason for this is the gap between deploying AI and getting value from it. AI that works in a demo still has to function inside each customer’s data, systems, and workflows. TSIA calls this the post-deployment value gap, and closing it takes skilled people. That’s why demand for AI professional services is growing, and buyers increasingly search for AI implementation services and expect their technology vendor to provide them.

TSIA’s research groups the new demand into three phases:

• Readiness and governance before deployment. Paid engagements that assess data, infrastructure, and governance, and qualify where AI can deliver value. Without this work, AI initiatives stall before they begin.
• Forward-deployed engineering during activation. Engineers who embed in the customer environment to get a first workflow working and show a visible result early. TSIA describes this as both a job role and a commercial offer.
• Optimization after go-live. Ongoing services that tune models, adjust workflows, and expand impact, often tied to consumption or shared-value pricing.

These map to TSIA’s DARE model of the AI customer lifecycle (Design, Activate, Realize, Evolve). Services carry more of the customer’s success in that model than they did in the SaaS era.

How Delivery Models Are Changing

The professional services delivery model is changing in three ways, and each one puts PS further from selling effort.

From Open-Ended Projects to Productized Offers

Open-ended statements of work give way to repeatable packages with defined scope and defined outcomes. TSIA’s PS 2.0 research describes this shift as moving from effort-based PS 1.0 toward a value-based model.

It treats services engineering - designing scalable, fixed-price offers - as a prerequisite for adopting AI. Productized offers matter for AI in two ways: they give AI a repeatable process to accelerate, and they allow providers to price the result rather than the hours.

From Staffing Pyramids to Smaller Senior Teams

The classic PS staffing model relies on many junior consultants per senior lead. Much of the junior work (documentation, configuration, testing, first-draft analysis) is exactly what AI accelerates. That points toward smaller teams weighted toward senior people who scope, design, and own customer relationships. The shift is likely to be gradual, and it raises a harder question about how future senior talent gets developed.

From Go-Live to Ongoing Ownership

AI systems aren’t finished at go-live. Models drift, data changes, and workflows need adjusting. This keeps PS teams engaged after deployment, often through managed, optimization, or subscription services.

This is the Evolve phase of TSIA’s DARE model, and it lines up with the PS-as-a-service models in TSIA’s PS 2.0 research. For PS leaders, it turns one-time project revenue into a recurring relationship.

What PS Leaders Should Do in the Next Year

The future of professional services in technology companies depends less on AI and more on how quickly PS organizations adjust what they’re selling. PS leaders should plan to make four moves this year:

1. Measure where AI already cuts hours in your projects. You can’t reprice or replan what you haven’t measured, and the savings are probably already showing up unevenly across your portfolio.
2. Reprice one offer away from pure time and materials. A single fixed-fee or hybrid offer gives you real data on margin before you commit the whole portfolio.
3. Build a small forward-deployed capability. New AI demand goes to providers who can get customer deployments working, and that capability takes time to build.
4. Set a margin target that assumes fewer hours. Plans that assume today’s effort levels will overstate revenue and hide the pricing problem until it shows up in the numbers.

Among the professional services trends TSIA is tracking, the shift from selling effort to selling results is the one most likely to decide which PS organizations grow. PS 2.0: The Year-One Scoreboard shows how members are progressing. The TSW workshop “From Outcome Talk to Outcome Operating Model” covers how to build the operating model behind it.

Talk to TSIA’s PS Research Team

Sources:

• Thomson Reuters 2026 AI in Professional Services Report
• TSIA: Professional Services 2.0
• TSIA: PS 2.0 Transformation Playbook
• TSIA: Services Engineering in the PS 2.0 Era
• TSIA: What Is Forward Deployed Engineering?