AI is pushing consulting away from billable hours and toward productized, learning operating models

Will Hinde, president of consulting services at Addison Group, is blunt about where AI pressure is landing first: the billing model. On Consulting Success’s August 2026 Podcast #386, Hinde argued that generative AI and emerging agent-like tools will make portions of traditional consulting delivery so fast and repeatable that selling time becomes harder to defend, pushing firms toward productized offerings and proprietary intellectual property.

McKinsey Global Institute Director Tanguy Catlin made the same underlying point from the buyer side in McKinsey’s July 9, 2026 episode, “The real AI advantage.” Catlin said the first wave of AI adoption has centered on efficiency, but as competitors access the same tools, advantage shifts to leaders who redesign customer experiences, rebuild how work gets done, and create operating models that learn faster than peers.

Ubiquitous tools are flattening “efficiency” as a differentiator

McKinsey frames a “productivity trap”: when a general-purpose technology becomes widely available, most companies initially stack it onto existing processes to do the same work cheaper and faster. Catlin’s point on McKinsey’s podcast is that once competitors do that too, the gains stop being a differentiator, and surplus value tends to get competed away or passed downstream.

That matters for enterprise operators because it changes what success looks like in an AI program. If the KPI stays stuck on labor hours eliminated or documents produced per analyst, leadership teams can miss the harder, more durable win: eliminating friction in the end-to-end workflow and changing the offering itself, which is what McKinsey describes as AI-powered reinvention.

Consulting delivery is being re-specified around IP and productization

Hinde’s comments on Consulting Success land at the commercial layer. He described a market moving toward productization and IP-based delivery, where firms package repeatable methods, data assets, templates, and tools into offerings that can be deployed consistently. The argument is partly defensive, if AI compresses delivery time, a pure time-and-materials model compresses revenue too. It’s also operational: productized work is easier to scale across teams and clients because the unit of delivery becomes a defined artifact and a repeatable implementation path.

Hinde’s vantage point is worth noting for procurement teams trying to interpret supplier claims. Consulting Success described him as having more than 25 years in management and technology consulting and over $900 million in consulting revenue generated during his career, and as leading growth across six specialized consulting brands within Addison Group. That kind of portfolio view tends to reveal which offerings can be systematized and sold repeatedly, and which truly remain bespoke.

What changes for CIOs, ops leaders, and procurement: the contract has to price learning, not labor

Put the two podcasts together and a practical requirement emerges: enterprises are going to need contracting constructs that match AI-shaped delivery. When McKinsey says the next wave of AI value flows to organizations that “learn faster,” that learning has to be made legible in governance. Otherwise, suppliers will optimize for what’s easy to bill and hard to verify.

For IT and operations leaders, this shows up in the messy middle of implementation. AI-enabled work often spans data access, systems integration, workflow redesign, and ongoing tuning as users adopt and edge cases surface. A billable-hour SOW can fund activity, but it doesn’t automatically force outcomes like reduced order-to-cash cycle time, fewer touches per claim, higher first-time-right rates, or lower cost-to-serve. If the supplier is also shipping reusable assets, like playbooks, configuration patterns, prompts, or proprietary accelerators, those should be spelled out as deliverables so the buyer can reuse them internally or across business units.

There’s also a benchmarking implication. Once a service line becomes “software-like,” operators can compare performance and cost per transaction or per workflow, not per role. That comparison is harder when the primary commercial instrument is a blended rate card.

Where this matters most in 2026 planning

This shift will matter most for enterprises buying large volumes of repeatable knowledge work, whether that’s finance transformation, customer operations, IT service management, analytics, or supply-chain planning, where AI can automate analysis and drafting and where delivery patterns can be reused. It’s less acute for truly one-off programs with heavy physical constraints, but even there, the planning, documentation, and reporting layers are increasingly AI-assisted.

The open question is timing: how quickly AI-enabled delivery becomes standard enough that buyers can demand outcome-based commercial terms without paying a novelty premium. McKinsey’s framing suggests the premium disappears as tools become ubiquitous. Hinde’s framing suggests suppliers will respond by packaging differentiated IP, making “what’s included” the battleground.

Questions to bake into your next AI services RFP and renewal

• Which deliverables remain after the engagement: reusable workflows, prompts, configurations, playbooks, training assets, and documentation, and what rights does the enterprise have to reuse them across sites and business units?
• How will the provider measure “learning faster”: what telemetry will be instrumented (cycle-time, rework, escalation rates, adoption), what is the cadence for improvement releases, and who approves changes in regulated processes?
• What is the unit price that actually matches the work: per transaction, per workflow, per site, per release, or per KPI improvement, and how does the provider avoid double-charging when AI reduces labor hours materially?
• If agentic AI is used, what are the guardrails: approval steps, audit logs, data access boundaries, and incident response responsibilities, and are those controls priced into the managed service or treated as change orders?