AWS Consulting Leader: AI is Reshaping Consulting’s Billable Hour Business Model

TLDR

• AI is breaking the link between consulting revenue and headcount.
• Consulting firms are turning expertise into reusable AI products and agents.
• Outcome-based pricing is growing, but billable hours remain entrenched.

AI is beginning to change the consulting industry’s oldest pricing structure: the billable hour.

As AI takes over parts of software engineering and business processes, consulting firms are starting to package expertise into products, deploy smaller teams and experiment with contracts tied more closely to results.

That is forcing firms built around selling human expertise by the hour to rethink how they make money.

“The old model of revenue being tied to their people, FTEs (full-time equivalents), is basically just a cost plus margin, and that’s what you charge,” Brian Bohan, newly minted director of partner strategy and growth at AWS, told The AI Innovator. “And now that’s completely decoupling.”

AI enables consulting firms to complete work faster and turn knowledge that once resided primarily in consultants into reusable software, agents and intellectual property.

This represents a fundamental change in economics. Instead of making money primarily by putting large teams on projects for months or years, firms increasingly can combine human expertise with products and AI systems that can be reused across customers.

Reusable assets accelerate timelines

Bohan said consulting companies have built reusable assets and intellectual property for years, but traditionally treated them as accelerators inside larger consulting engagements rather than products in their own right. AI is accelerating efforts to monetize those assets directly.

He pointed to work AWS has done with Accenture around know-your-customer (KYC) and anti-fraud processes in banking. Traditionally, a consultancy might develop reusable technology after doing several similar projects, then use it to reduce the time and cost of the next engagement.

Now, Bohan said, an agentic KYC and anti-fraud solution has been packaged as a product that customers can purchase through AWS Marketplace. The consulting firm can then provide the remaining integration and customization.

The result is a hybrid model: part software sale, part consulting engagement.

“Monetization (in this manner) has gotten really serious in the last year,” Bohan said. Customers increasingly do not want firms to, for example, “just charge me for 20 people times 40 hours times 50 weeks.”

“Give me some value add,” he said clients are now requesting. “We’re seeing this productization as a way to provide that value and recapture that value.”

‘Productization’ strategy

There are signs that the ‘productization’ strategy is beginning to generate meaningful revenue.

Revenue generated by consulting partners through AWS Marketplace has more than doubled from a year earlier, the cloud giant said.

Since 2024, revenue has increased more than fivefold while the number of participating consulting partners has remained roughly flat, suggesting the growth is coming primarily from customers spending more with individual partners rather than simply from additional firms joining the marketplace.

Accenture, for example, recently launched an AI Products capability within its AWS business group, offering reusable AI products through AWS Marketplace. Its initial offerings include an Autonomous Supply Chain product built with more than 30 enterprise AI agents using Amazon Bedrock and AgentCore.

In a related move, Accenture Edge and AWS launched a collaboration bringing six new AI offerings to market aimed at mid-market organizations through a curated AWS Marketplace storefront – another sign of consulting firms packaging expertise as products rather than selling it by the hour.

Outsourcing business faces its own reckoning

A second front of disruption is in the outsourcing market. The traditional model – handing back-office processes like finance, accounting or HR to an outside company – is being upended as customers now expect AI to “agentify” those processes and drive savings directly.

AWS’s answer is what Bohan called the Agentic Process Transformation Initiative. He pointed to work with a U.K. public-sector outsourcer that, unlike some global competitors, cannot send work overseas and must manage everything from the U.K.

AWS and the outsourcer agentified back-end processes it manages for the U.K., saving the government money – a playbook AWS says it has since rolled out globally across industries and process types.

Notably, Bohan framed this as self-disruption: the outsourcing firms are agentifying their own processes before someone does it to them. And while he acknowledged “a reduction in headcount, reduction of people-based” work, he said the firms’ public posture is reassignment to higher-value roles rather than cuts.

Genpact, for example, has built what he described as an autonomous IT capability – self-healing IT services powered by agents, skills and ontologies – and AWS is backing that shift with a managed services initiative and competency that invests in partners demonstrating that kind of innovation.

Smaller teams, faster projects

The consulting industry’s monetization shift also changes how the work is delivered.

Bohan said AWS is seeing greater use of forward-deployed engineering, or FDE, strategies in which small groups of highly skilled engineers work directly with customers for a defined period and toward a specific outcome.

That resembles traditional consulting in one respect: Firms have always sent experts into clients. But the economics and speed are different when those experts are augmented by AI and reusable technology.

Bohan described the emerging model as “small groups of engineering experts being dropped into customers for a fixed amount of time at a fixed priced, driving towards an outcome.”

AWS is putting substantial capital behind that approach. The company recently announced a $1 billion investment in forward-deployed engineering and is also extending FDE capabilities to consulting partners through a credentialing program.

AWS is piloting a partner-led FDE program with about a dozen large global systems integrators. Partners are building what Bohan called a “harness” containing ontologies, skills and agents on top of Amazon Bedrock. The idea is to capture what is learned during one project so the next similar engagement does not begin from scratch.

For example, after completing one claims-processing transformation, Bohan said a consulting team might arrive at the next client already “20%, 30% down the road” because knowledge from the earlier project has been encoded into the system.

That can compress project timelines. Bohan cited Deloitte’s work with Tower Insurance, saying the consultancy moved the insurer from proof of concept to production in 10 weeks. The deployment improved call handling by 26%, accelerated email handling by 18% and freed 8,200 hours of agent time.

Those are the kinds of operational measures AWS is trying to make central to AI deployments through its Business Value Realization initiative. The program includes more than 1,000 industry benchmarks covering different use cases, along with agents and skills designed to help measure outcomes.

“When you’re talking about AI, selling the AI is just the starting line,” Bohan said.

Companies must take projects into production, get the technology broadly adopted and make sure “it’s hitting the metrics that were initially laid out in the business value case in the beginning,” he said.

The billable hour is not dead

Despite the industry’s rhetoric around outcome-based consulting, Bohan cautioned against declaring the billable hour dead.

“Outcomes are becoming increasingly important,” he said. “If you look at the data, though, on how many contracts are in fact outcome-based, it’s still fairly low.”

One obstacle is measurement. Broad business objectives such as customer churn, Net Promoter Scores (NPS) or new-product launch cycles may take months to materialize and can be affected by factors outside the consultant’s control.

More immediate operational measures can be easier to incorporate into contracts. For example, companies can measure how many customer applications move through a process without human intervention and how quickly applicants are processed.

“These kinds of metrics get measured and increasingly baked into contracts,” Bohan said. “But again, it’s still early days for that.”

The new consulting cost equation

AI also introduces a new complication into the consulting industry’s economics: the technology doing the work isn’t free.

Model inference, token consumption, cloud infrastructure and data processing can become significant costs as companies deploy AI at scale. Bohan compared the challenge with the early years of cloud computing, when companies sometimes discovered that moving workloads to the cloud without proper controls could produce unexpected bills.

“We’re seeing the same thing now with AI,” he said. “It’s just that we’re very early.”

Bohan said companies need observability into AI usage, the ability to route work among models based partly on economics and governance systems capable of enforcing cost policies. Different steps in an AI workflow, for example, do not necessarily require the same model, allowing companies to choose models based in part on the economics of individual tasks.

The longer-term change may be what happens to the expertise consulting firms traditionally sold by the hour.

Bohan said consulting remains valuable precisely because businesses are struggling to absorb rapid technological change and determine how to respond. But he expects more of consultants’ accumulated knowledge to migrate into AI systems.

“I think it is going to get increasingly encoded in the harness, in agents, in ontologies, in skills,” he said.

Consider a consulting team working on a specialized utility process. The first engagement might require an expert embedded directly in the project. After the second or third implementation, Bohan said, perhaps 80% of that knowledge could be encoded into a reusable harness or product.

The expert can then oversee multiple engagements rather than personally implementing each one.

That creates a different scaling model for an industry that historically scaled largely by hiring more people.

The consulting firm of the future may still sell expertise. But increasingly, it may sell that expertise in two buckets: once through its people and again after their knowledge has been turned into software.