A $63 Million Bet on a Very Familiar Problem

A Menlo Park-based startup called Skan AI just pulled in $63 million in Series C funding — bringing its total raised to around $120 million — to solve a problem that should sound very familiar to any business owner: AI tools keep failing because nobody actually knows how work gets done inside a company.

According to VentureBeat's coverage of the raise, Skan's entire thesis is that the AI models themselves aren't the problem. The real gap is operational context — the messy, undocumented reality of how employees actually move through their work versus how leadership assumes they do. That gap is what kills AI projects before they ever deliver value.

What Skan Actually Does (In Plain English)

Skan observes how employees interact with software systems in real time, then builds what it calls a 'context graph of work' — essentially a detailed map of actual workflows. That map becomes the foundation for AI agents that can understand, model, and eventually automate those workflows. The platform runs on NVIDIA AI Enterprise and NIM microservices and has processed over 25 billion work signals.

Alongside the funding announcement, Skan launched two new products — Skan AI Blueprint and Skan AI Agents — which combine with their existing Skan AI Intelligence platform to create a full pipeline: discover how work happens, model it, then automate it. The platform is used by seven of the ten largest U.S. banks and one-quarter of the Fortune 50.

This is enterprise-grade software aimed at large organizations. But the underlying problem it's solving? That one hits home for businesses of every size.

The Stat You Should Sit With

Here's the number worth pausing on: Gartner research cited in the article found that only 8% of enterprises have AI agents actually running in production, and a staggering 95% of early implementations need a complete redesign. At the enterprise level, with unlimited budgets and dedicated AI teams, the failure rate is nearly 100% on the first try.

If billion-dollar companies are struggling this badly, it tells you something important about AI implementation in general — and it should recalibrate your expectations for your own AI experiments.

What This Means for Your Business Right Now

You're probably not buying Skan's platform. It's built for organizations with thousands of employees and complex software ecosystems. But the strategic insight behind it applies directly to you.

  • AI needs context you probably haven't documented. Before layering any AI tool onto your operations, you need a clear, honest picture of how work actually flows — not how you think it flows. Walk the process. Watch someone do the job. The gap between the org chart and reality is where AI implementations go to die.
  • Start with process mapping, not tool shopping. The instinct for most owners is to find a great AI tool and plug it in. Reverse that. Map your highest-volume, most repetitive workflows first. Then find tools that fit the actual process.
  • Small failures now are cheap tuition. The 95% redesign rate at the enterprise level reflects one universal truth: your first AI implementation will teach you more than any vendor demo. Build that expectation in from the start so a stumble doesn't become an excuse to abandon the whole effort.
  • Your advantage is speed and proximity. A 20-person company can do in two weeks what takes a large enterprise six months — observe workflows, identify friction, test a tool, and adjust. Use that advantage aggressively.

The Bigger Trend to Watch

The round was co-led by Cathay Innovation and Dell Technologies Capital, with Citi Ventures, Bloomberg Beta, State Farm Ventures, and Wipro Ventures also participating. That's not a random collection of names — it spans financial services, enterprise tech, and venture. They're collectively betting that 'operational context' becomes the critical layer that makes AI actually work inside real organizations. That thinking is going to filter down into SMB-focused tools over the next 12 to 24 months.

Watch for AI tools that ask more questions about your workflows before they claim to automate them. That's the signal that a vendor has internalized this lesson. Be skeptical of any tool that promises automation without first understanding the process it's replacing.

The Bottom Line

Skan's $63 million raise is a large-scale validation of a very practical reality: AI fails when it's disconnected from how work actually happens. You don't need a nine-figure platform to act on that insight. You need honest process documentation, realistic expectations, and a willingness to iterate fast. Start there, and you'll be ahead of most enterprises spending millions to learn the same lesson.

Source: VentureBeat — Skan AI raises $63 million

This review was written in response to: Skan AI raises $63 million betting that watching how employees actually work is the missing layer of enterprise AI