90% of Corporate Knowledge Lives in Heads – and Leaves With Them

01 /The Concept

"Undocumented knowledge is hidden organizational risk," declares Knoa. This isn't just about code that engineers refuse to document – it applies to any kind of institutional knowledge about how work actually gets done.

The startup estimates that 90% of a company's knowledge lives exclusively in its employees' heads. That sounds extreme, but it's probably close to true.

The consequences pile up fast:

- When a long-tenured employee leaves, the company loses years of accumulated expertise. And recovering that knowledge is rarely free – reconfiguring a legacy system that someone built and maintained alone can consume three months of a new developer's time.

- When consultants come in, they spend weeks or months just running interviews, trying to extract enough information about how the company actually works to make any advice worth anything. Without that groundwork, their recommendations are useless.

Knoa's target customers are companies, consultants, and technology implementation specialists – anyone who needs to systematically pull knowledge out of people.

The most common use cases inside companies: capturing knowledge from departing employees, transferring institutional context to successors, and documenting existing workflows before automating them.

For consultants and implementation teams, the obvious applications are initial discovery interviews and requirements gathering – both of which are recurring, time-consuming bottlenecks in every engagement.

Knoa's workflow is straightforward. You define the interview goal in plain language – describe what you need to learn, with as much relevant detail as you want – and the AI generates an internal interview structure from that description. Then you send the generated interview link to whoever needs to be interviewed – one person or many.

When someone opens the link, they're connected with an AI interviewer they can talk to by voice or in text chat. Critically, the AI doesn't just ask preset questions and record answers – it actually converses: probing deeper, asking follow-ups, clarifying, and redirecting back on topic. Its goal is to extract everything relevant to the original objective.

Afterward, the AI produces a clean, structured PDF report plus a Markdown insights file – the latter ready to be fed directly into Claude or another AI for further processing.

Interview goals and report formats can be saved as templates, so the same type of session doesn't have to be rebuilt from scratch every time.

One particularly useful feature: the platform can convert verbal explanations into flowcharts – helpful when employees are describing something naturally structured, like a business process, a decision sequence, or a system architecture.

Knoa reports that companies using the platform can onboard new employees three times faster, cut the time spent on process documentation by 97%, and fully preserve organizational knowledge that previously existed only in people's heads.

Knoa launched this week, with the announcement going live on Product Hunt.

02 /Why It Matters

The loss of institutional knowledge – through employee departures or poor internal transfer – turns out to be a staggeringly large problem:

- Fortune 500 companies collectively lose $31.5 billion per year to knowledge gaps.

- Large companies outside that cohort lose an average of $47 million per year.

- A company with 1,000 employees loses roughly $2.4 million annually.

- The departure of a single employee costs an average of $430,000, accounting for the productivity lag of their replacement and potential direct costs of reconstructing lost expertise.

In total, poor knowledge management can cost a company up to 25% of its annual revenue. Platforms like Knoa exist to claw that back.

Because the problem is genuinely massive, other startups have entered the space.

Sensay ([covered previously](/review/malozametnaja-problema-kotoraja-skoro-prevratitsja-v-katastrofu)) raised $3.4M for a platform that extracts departing employee knowledge through a similar AI interviewing approach. Its key differentiator is HR system integration – the platform can learn when an employee has signed their exit paperwork, then initiate a targeted interview while knowing exactly what role they held, so the AI knows which questions to ask.

Viven ([related review](/review/ii-dvojniki-jeto-milliardnaja-tema)) raised $35M in its Series A last October for a more ambitious take: full digital twins of employees. Unlike Sensay's exit-interview model, Viven's twin accumulates knowledge continuously while the employee is still on the job – no dedicated session required. The second goal is more novel: colleagues can query an employee's digital twin when that person is unavailable – on vacation, traveling, or simply in another time zone.

And here's an adjacent application of the same AI interviewing model, pointed at a completely different goal:

Your360.ai ([covered here](/review/vtykaj-ii-mezhdu-ljudmi)) built a platform that collects peer feedback about an employee's performance – using an AI interviewer to gather input from colleagues, then synthesizing it into a structured personal growth plan that maps strengths, development areas, and specific next actions – backed by anonymous quotes from the interviews.

03 /Opportunities

AI interviewers are a genuinely new category – one that only became viable at scale recently, as voice AI hit a threshold of quality and affordability. Expect significantly more startups to deploy this capability across a widening range of domains.

The broadest opportunity is building AI interviewers for any context where pulling structured insight out of people is valuable. The use cases are enormous – startups are already applying this technology to capture family memories, for instance.

But the more direct path to revenue runs through enterprise use cases. The most obvious ones have already been covered here. That doesn't mean they're off the table – the addressable market is large enough that a few early entrants won't saturate it.

What less obvious or more specialized corporate AI interviewing applications can you imagine from your own professional experience? The window to build them is open right now.