PwC's 7-Day AI Carve-Out Assessment: What Independent Consultants Can Steal

The fire doesn't wait for a study

PwC and SNP announced a carve-out intelligence tool that cuts corporate divestiture assessments from 4-6 weeks down to 7 days. Harry Schuster, Partner at PwC's Deals Practice, presented the system at Transformation World 2026 in Heidelberg. The mechanism: AI data extraction paired with a rule-based engine, replacing the manual document review that used to eat a month of analyst time.

I ran nuclear reactors on a submarine before I ran a firm. When something catches fire in the engine room, you do not convene a working group. You do not schedule a study. You run the procedure. You assess the damage. You act, in minutes, because the boat does not care about your calendar.

PwC just applied that same doctrine to deal consulting. A carve-out assessment is a damage control assessment for corporate structure: what depends on what, what breaks if you cut it loose, what it costs to stand alone. For decades, consulting firms have run that assessment like a research thesis. Weeks of interviews. Weeks of document review. A report nobody reads until the deal is already closing.

PwC compressed it to a week. That is not a productivity tweak. That is a different operating doctrine, and every independent consultant competing against firms ten times their size needs to understand why it works and how to run the same play solo.

Why the 6-week assessment was always theater

A carve-out assessment does one job: figure out what a business unit needs to survive after separation. IT systems, shared services, vendor contracts, headcount dependencies, financial reporting lines. None of that is a mystery. It exists in documents, in ERP systems, in contracts, in org charts. The 6-week timeline was never about the difficulty of the analysis. It was about the manual labor of extracting structured answers from unstructured sources.

That is the same bottleneck that eats every consulting engagement, not just M&A work. Independent consultants report spending 40-60% of project time on non-billable interview transcription and data-wrangling, work that produces zero client-facing value on its own. You are not paid to type notes. You are paid for the judgment you apply after the notes exist.

PwC's system attacks that exact bottleneck. AI extraction reads the source material. A rule-based engine applies the firm's carve-out doctrine, the accumulated pattern library of what matters and what doesn't. The partner's job compresses to reviewing outputs and making judgment calls on the ambiguous 10%. That's watchstanding, not manual labor: you're monitoring the system and intervening on exceptions, not doing the work yourself.

This isn't an isolated case. Link Consulting built an AI-accelerated discovery process that compresses a 2-week discovery phase into a 4-hour session. Vitelis runs assessments that used to take 4 weeks in a single workshop. Different firms, same math: the interview-and-synthesize model of consulting has a fixed cost that AI extraction removes almost entirely.

The doctrine: systems beat slogans

Here's what separates PwC's move from a marketing claim. They didn't tell clients to "work smarter" or "prioritize efficiency." They built a system: AI extraction plus a rules engine plus a defined output format. The system does the same thing every time, regardless of which partner runs it or how tired the analyst is at hour 80 of the engagement.

That's the whole game for independent consultants. You don't have PwC's balance sheet. You don't have hundreds of analysts to throw at document review. What you have is the ability to build your own system, once, and run it on every engagement afterward. A slogan is something you say to a client. A system is something you can point to and say: this is how we get from raw material to deliverable, on a schedule, every time.

Most solo consultants never build the system. They run each engagement as a bespoke project: new interview guide, new synthesis process, new deck template, reinvented from scratch because there was never time to compartmentalize the repeatable parts from the client-specific parts. That's the operator-dependent trap. Every hour you save on this project doesn't compound into the next one.

The tactical build: your own 7-day protocol

You don't need PwC's engineering budget to run this play. You need three components, built once and reused on every engagement.

Component 1: Structured intake. Before any client call, define exactly what data you need and in what format. Financial statements, org charts, contracts, whatever your practice area requires. Send a structured request instead of an open-ended "send me what you have." This is the equivalent of PwC's document intake step: it only works if the inputs are consistent enough for a system to process.

Component 2: AI-assisted extraction. Feed the structured intake into a tool built to pull specific data points, not summarize generally. Tools like AiDiscover are built specifically to automate the discovery phase, turning raw interview and document material into presentation-ready findings in minutes instead of days. The point isn't to replace your judgment. It's to replace the transcription and first-pass synthesis that used to consume the 40-60% of project time mentioned above.

Component 3: A rules layer. This is the part most consultants skip, and it's the highest-leverage part. Write down your own pattern library: the recurring risks, red flags, and decision points you've learned to spot across engagements. That's your accumulated expertise, encoded as a checklist or a set of prompts instead of trapped in your head. When you apply that rules layer to the extracted data, you get PwC's rule-based engine at solo-practice scale. This is the asset. The extraction tool is a commodity. Your pattern library is not.

Run those three components in sequence and a discovery phase that used to take two weeks collapses to a few days. That's not a hypothetical. It's what Link Consulting and Vitelis are already reporting in production engagements.

What this does to your economics

Speed isn't the real prize here. Speed is the visible symptom. The real prize is what speed does to your unit economics.

Run the math. If discovery used to consume 40% of a 4-week engagement and now consumes 10%, you've freed up three weeks of billable capacity per project without hiring anyone. That capacity either becomes more clients per quarter or more depth per client. Either way, it's pure margin, because your fixed cost (you) didn't change.

There's a second-order effect that matters more for anyone thinking about their firm as an asset rather than a job. An engagement that depends entirely on your personal ability to conduct interviews and synthesize findings by hand is operator-dependent. It doesn't scale, and it isn't sellable, because a buyer is purchasing your calendar, not a system. An engagement built on a documented intake process, an extraction tool, and a rules engine you wrote down is a system. Systems transfer. Systems are what make a consulting practice acquirable instead of just a well-paid job.

That's the receipts test for any AI tool you're evaluating: does it make you faster this week, or does it make the practice less dependent on you existing? PwC built the second kind. So should you.

Why the fastest firms move first, not the biggest

There's a pattern in every one of these AI-accelerated consulting plays worth naming directly. It's not the largest firms making the first move by size. It's whichever team has the shortest distance between a decision and an implemented system. PwC has the balance sheet, but 400,000 employees don't turn a battleship fast. What let this ship in the first place was a partner willing to co-build with a technology vendor instead of waiting for an internal committee to bless a multi-year platform build. Schuster's own framing at Transformation World 2026 was co-innovation with SNP, not a top-down mandate. That's a founder-operator move happening inside a partnership structure.

Independent consultants have that same structural advantage, magnified. You have no committee. You have no procurement cycle. You can evaluate a discovery automation tool on Monday and be running it on a live engagement by Friday. The firms this threatens most are not the solo operators. It's the mid-size consultancies stuck between PwC's R&D budget and your speed, running the old 6-week playbook because nobody with authority has greenlit the new one.

That's the actual competitive window right now. It won't stay open. Once AI-accelerated discovery becomes table stakes across the consulting market the way PwC's own SAP Insider coverage suggests is already underway, clients will expect the compressed timeline as a baseline, not a differentiator. The consultants who build the system now get to charge a premium for speed while it's still rare. The ones who wait get to match the new baseline just to stay in the bidding.

The casualty drill mindset applied to client work

Going back to the submarine for a second, because the metaphor holds all the way through. A casualty drill isn't just about speed. It's about having a rehearsed procedure so that when the real casualty happens, nobody is improvising under pressure. The procedure was built in calm water, drilled repeatedly, and then executed without hesitation when the alarm sounded.

That's what PwC did with the carve-out tool. The 7-day timeline isn't magic. It's the output of a procedure built once, in calm conditions, and then run on every subsequent engagement. The independent consultant who builds their own intake-extraction-rules protocol is doing the same thing: building the procedure now, in the between-engagement calm, so that when the next client fire hits, you're not improvising a discovery process from a blank page. You're running the drill.

Due diligence on your own practice means asking: if I got hit with three engagements next month, does my current process scale, or does it break because it all runs through my personal bandwidth? If the answer is the latter, you have a bottleneck, and PwC just showed you what removing it looks like at enterprise scale. Your version doesn't need to be enterprise scale. It needs to exist.

FAQ

Does this mean AI is replacing consultants? No. It's replacing the non-billable transcription and synthesis labor that consultants have always resented doing. The judgment layer, deciding what the extracted data means and what to recommend, stays human. PwC didn't fire its Deals Practice partners. It gave them a faster engine room.

What tools should a solo consultant use to build this? Start with whatever AI extraction tool fits your practice area's documents. Purpose-built discovery automation tools exist for consultants specifically, turning interview and document material into structured output. The tool matters less than the discipline of building a repeatable intake-extraction-rules sequence around it.

Is 7 days realistic for a solo practitioner, or just for a firm with PwC's resources? The absolute timeline will vary by engagement complexity. What transfers regardless of firm size is the ratio: if AI extraction removes most of the manual document and interview processing, a solo consultant sees the same proportional compression PwC saw. A 4-week engagement can plausibly become a 1-week engagement for well-scoped work.

What's the risk of moving this fast? The same risk as any casualty drill: skipping steps under time pressure. The rules engine exists precisely to prevent that. If you encode your pattern library into a checklist before you need it, speed doesn't cost you thoroughness. If you try to build the checklist while running the engagement, it will cost you both.

How does this affect what my practice is worth if I want to sell it eventually? Directly. A practice built on your personal interviewing skill is a job. A practice built on a documented, tool-assisted discovery system is an asset with a transferable process, which is what buyers pay a multiple for. Building the system PwC just validated is also, incidentally, the first step toward making your practice sellable.