A Case Interview for Your Knowledge Team: How Much Is Your Knowledge Worth?

Key takeaway: The knowledge tax is the compound cost a consulting firm pays — in duplicated work, slower delivery, weaker AI tools, and lost competitive advantage — because confidentiality constraints prevent systematic reuse of past deliverables. For a 2,000-person firm, it runs into tens of millions per year. The only way to reduce it is systematic, automated sanitisation of the content that's sitting unused.

Last updated: March 2026

The case

Here's one for your KM team's next offsite. Frame it exactly the way you'd frame a market sizing in a case interview:

"Your firm has 2,000 consultants. They produce roughly 50 new slide decks per week across all engagements. Over the past five years, that's approximately 13,000 decks. Ninety-five per cent of them sit in project archives, never reused. What is the annual cost of that locked-up knowledge?"

If a candidate gave you this case in an interview, you'd expect a structured approach. Direct costs, indirect costs, opportunity costs. A build-up from the bottom. Sensible assumptions, clearly stated.

So let's do exactly that.

Sizing the direct costs: recreated work and duplicated research

The most tangible cost is work that gets done twice. A team in London builds a market entry framework from scratch while an identical one sits in a Singapore project folder they can't access. A new hire spends a week developing a benchmarking model that three other teams have already built.

This isn't hypothetical. Consulting firms' best slide decks are locked away behind confidentiality constraints. The knowledge exists. People often know it exists. But there's no practical mechanism to get to it safely.

Let's size it. APQC research consistently finds that knowledge workers spend roughly 2 hours per week recreating work that already exists somewhere in their organisation. In consulting — where a single deliverable might take 20-40 hours to build — that's conservative. But let's use it.

The arithmetic

Not every consultant is duplicating work every week. Let's be conservative and say 30% of consultants lose an hour per week to rework that could have been avoided with better access to past deliverables.

• 600 consultants x 1 hour/week x 48 working weeks = 28,800 hours/year of duplicated effort
• At a blended billing rate of $350/hour, that's $10 million per year in lost billable capacity — time that could have been spent on client work instead of rebuilding something that already exists

And that's just the duplication you can measure. The work that never gets started — the proposal that doesn't include a relevant case study, the pitch that uses a generic framework instead of a proven one — doesn't show up in anyone's timesheet.

Sizing the indirect costs: slower proposals, weaker pitches, longer ramp-up

Direct duplication is the easy part of the case. The indirect costs are where the numbers get uncomfortable.

Slower proposal turnaround

When a partner needs to respond to an RFP, the team's first question is: "Have we done something like this before?" If the answer is "probably, but we can't find it or can't use it because it's confidential," the team starts from a blank page. That adds days to a process where speed often determines who wins.

A partner's time costs $500-800/hour. A manager's runs $200-400/hour. If a proposal takes an extra day because the team can't access relevant past work, you're looking at $5,000-10,000 in additional cost per proposal — before you account for the competitive disadvantage of being slower.

Weaker business development

The best business development tool a consulting firm has is its past work. "Here's what we did for a client in your exact situation" is worth more than any capability deck. But showing that work requires sanitising it first — removing the client's name, financials, strategic details, anything that could identify them.

Manual sanitisation takes 4-8 hours per deck. For a 30-60 slide strategy presentation, that's a full day of a knowledge manager's time. So most firms don't bother for most decks. The BD team works with a handful of pre-approved case studies instead of the full breadth of the firm's experience.

Junior consultant development

New consultants learn by studying how senior teams structured their thinking. The apprenticeship model depends on access to real work. When 95% of that work is locked behind confidentiality, juniors learn from the 5% that's been sanitised — which is typically the most generic, least instructive material.

The ramp-up cost is real. An extra month to full productivity for a new consultant at a blended cost of $15,000/month, across 200 new hires per year, is $3 million. And that's a modest estimate.

Sizing the AI opportunity cost: the most expensive line item

This is the part of the case where the numbers change by an order of magnitude.

Every major consulting firm is investing heavily in AI. McKinsey's Lilli reached 72% adoption among 45,000 employees, drawing on over 100,000 documents. Deloitte, BCG, Bain, and the Big Four — who collectively employ 1.5 million+ people — are all building or deploying internal AI knowledge tools.

The industry has poured north of $10 billion into AI infrastructure. But here's the problem: the content pipelines feeding these systems remain blocked.

If your AI knowledge assistant runs on 5% of your firm's actual deliverables — the thin layer of sanitised case studies and published thought leadership — it's operating on scraps. The detailed analyses, the proprietary frameworks, the sector-specific models, the engagement-tested methodologies: all sitting in confidential project folders, untouchable.

Failed RAG deployments

I've spoken with firms that spent millions building RAG (Retrieval-Augmented Generation) pipelines, only to discover that the system keeps returning the same generic content. The AI is technically working. The knowledge base is just empty of anything useful.

The problem isn't the AI architecture. It's that nobody solved the upstream content problem: how do you get confidential deliverables into a state where AI systems can safely ingest them?

Some firms have tried various approaches to making confidential decks reusable — document-level permissions, metadata-only indexing, summary extraction. None of them work at the scale AI systems need. The AI needs the actual content. The content is confidential. Without a systematic way to sanitise documents at scale, the AI investment sits there underperforming.

The opportunity cost calculation

A 2,000-person firm might spend $5-10 million per year on AI knowledge infrastructure — platform licences, engineering, data pipelines, change management. If that infrastructure operates at 20% of its potential because the knowledge base is starved of content, the opportunity cost is $4-8 million per year in unrealised value.

And it compounds. Each year of AI investment with a blocked content pipeline means another year of suboptimal returns. The sunk cost grows. The gap between what the AI could do and what it actually does widens.

Introducing the knowledge tax

Add it up. For a 2,000-person consulting firm:

Cost category	Annual estimate
Lost billable capacity (duplicated work)	$8-12M
Slower proposals and weaker BD	$3-6M
Extended junior ramp-up	$2-4M
AI opportunity cost	$4-8M
Total knowledge tax	$17-30M

That's the knowledge tax.

The knowledge tax is the compound cost a consulting firm pays — in duplicated work, slower delivery, weaker AI tools, and lost competitive advantage — because confidentiality constraints prevent systematic reuse of past deliverables.

It's not a technology failure. It's not a people failure. It's a structural problem: consulting firms produce high-value knowledge under confidentiality constraints, and until recently, there was no practical way to separate the insight from the sensitivity at scale.

The tax is invisible because it shows up as "how things have always been." Nobody gets a line item on their P&L for "knowledge we produced but couldn't reuse." It's buried in longer project timelines, in proposals that took three days instead of one, in AI tools that return mediocre results, in junior consultants who take six months to become productive instead of three.

Why the knowledge tax keeps growing

Here's what makes this urgent rather than merely interesting: the tax is compounding.

AI without sanitisation creates compliance risk

AI spend is skyrocketing across consulting. But without sanitisation, every GenAI tool the firm deploys — RAG pipelines, internal copilots, enterprise search — risks exposing confidential client information. The more content these systems can access, the more useful they are. The more useful they are, the more people use them. The more people use them, the higher the probability that confidential details surface where they shouldn't. The knowledge tax isn't just a cost problem any more. It's a compliance problem that scales with AI adoption.

Knowledge creation is outpacing sanitisation

AI is exponentially accelerating content generation. Consultants are producing more slides, more analyses, more deliverables than ever before. But the sanitisation process is still manual — which means it can only scale by adding headcount, bloating costs. The gap between what's being created and what's being made reusable widens every quarter. A firm that has accumulated 13,000 confidential decks over five years will have 20,000 in another three. The backlog grows faster than any manual team can clear it.

Margins are tightening: sanitisation is getting costlier

Staffing models across consulting are shifting. Leaner pyramids mean fewer junior staff and more work pushed up to experienced consultants and managers. That shifts the sanitisation burden — whether formal or informal — to more expensive people. The same manual review that used to happen at $80/hour loaded cost is now happening at $150-200/hour, if it happens at all. Most of the time, it doesn't. The deck just stays locked away, and the knowledge tax goes up.

The path to reducing the knowledge tax: systematic sanitisation

So how does the case resolve? What's the recommendation slide?

The knowledge tax exists because of a specific bottleneck: confidential content can't be reused without being sanitised first, and manual sanitisation doesn't scale. A knowledge manager processing 4-8 hours per deck, across thousands of decks, produces single-digit coverage. The maths doesn't work.

The only way to materially reduce the knowledge tax is to make sanitisation systematic and automated. Not manual review at heroic scale. Not metadata-only workarounds. Not "just use better search." Actual content-level sanitisation that produces clean, editable, reusable deliverables at the volume consulting firms need.

What systematic sanitisation requires

This isn't simple find-and-replace. Consulting redaction requires specialised treatment because the sensitive content isn't personal data — it's business information, strategic insight, and contextual details woven through every slide.

A systematic approach needs to handle the full sensitivity landscape in consulting content:

• Direct identifiers: Content that names or unmasks the client — much of which is not a literal keyword match
• Inference risk: Combinations of individually harmless details that narrow to a single client across a multi-section deck
• Non-public information: Content that's confidential even without identifying the client
• Visual content: Charts, diagrams, org structures, and images that carry sensitivity human reviewers often miss

And it needs to work natively in PowerPoint — the format consulting firms actually use — producing editable PPTX output, not flattened PDFs. The confidentiality wall between consulting content and AI won't come down through document conversion workarounds.

The ROI case writes itself

If systematic sanitisation can move a firm from 5% content reuse to even 30-40%, the knowledge tax drops dramatically:

• Duplicated work falls as consultants find and reuse existing frameworks instead of rebuilding them
• Proposals get faster as teams pull from a library of sanitised past work
• AI knowledge tools finally have enough content to be genuinely useful
• Junior consultants learn from the full breadth of the firm's experience, not just the generic fraction

Against a knowledge tax of $17-30 million for a 2,000-person firm, even a partial reduction represents millions in recovered value. The investment required to achieve it is a rounding error by comparison.

So, how much is your institutional knowledge worth?

Back to the original case question. The answer depends on your firm's size, but the structure is the same everywhere:

Take the number of deliverables your firm produces. Multiply by the percentage that never gets reused. Apply the cost of duplication, the cost of slower delivery, and the opportunity cost of AI systems running on a fraction of available knowledge. That's your knowledge tax.

For most mid-to-large consulting firms, it's tens of millions per year. For the Big Four, with their hundreds of thousands of employees and proportionally vast content libraries, the figures scale accordingly.

The institutional knowledge is worth exactly what you're paying to not use it.

The question isn't whether the knowledge tax exists. Every KM leader I speak with recognises it immediately — they've been living it. The question is whether your firm decides to keep paying it.