The Knowledge in Their Hands: How One Startup Is Betting AI Can Capture What Veteran Field Workers Know

When a senior utility technician retires, they take something with them that no system has successfully captured: the judgment that lives in muscle memory, the instinct for which poles to eye differently in high wind, the knowledge of which shims have settled wrong and which junctions no longer match any drawing in the system. Cloneable, a two-year-old startup, has raised $4.6 million — according to Crunchbase News published today, a figure type0 could not independently verify — to capture exactly that. Its system watches how technicians perform specialized tasks and deploys AI agents to replicate those workflows in the Computer-Aided Design and asset-management programs utilities already run, according to Cloneable's product page. What Cloneable has not yet demonstrated is that it can do this at scale: the evidence is one anonymized customer quote and a product integration, not a documented deployment with verified results.

The question the expertise-commodification pitch raises — and the story does not answer — is what happens to the experts themselves when their knowledge is extracted and deployed as AI agents. Do they become more valuable, their judgment supervising a larger workforce? Do they get redeployed, or does the system simply make them redundant? Cloneable did not respond to a request for comment on this question.

The Katapult Pro integration lets field crews collect measurements on mobile devices and feed them directly into the asset-certification programs utilities use to approve new pole attachments, working offline in areas where cellular coverage is unreliable. Prior attempts to digitize field workflows in utility corridors have failed for exactly this reason: the translation layer between field data and engineering software is where the work actually happens, and that is where AI model quality stops being the constraint. An independent 2025 analysis by CableLabs, the cable industry's research consortium, corroborates that the translation-layer bottleneck is the real problem. It does not corroborate that Cloneable has solved it.

The 40% capacity improvement cited on Cloneable's website comes from a single anonymized telecom infrastructure provider, not an independent source. Cloneable's prior funding was a $750,000 pre-seed reported by FinSMEs in 2023. Whether the current funding figure is real is a separate question the evidence cannot answer. The next six months of actual customer deployments will determine whether Cloneable has achieved genuine generalization or is a well-marketed wrapper around existing workflow automation. The funding, if real, buys time to find out.