April 23, 2026AgentsFunding-Seed

Cloneable turns tradesmen into agents

Raleigh-based Cloneable raised $4.6M seed yesterday, led by Congruent Ventures with First In, Overline, Bull City, and Texas telecoms investor St. Elmo Venture Capital participating. Total funding hits $5.35M since 2023. Founders are Lia Reich (CEO), Tyler Collins (CTO), and Patrick Lohman (CRO). Customer list already includes American Electric Power, Southern California Edison, Burns & McDonnell, and Perdue — that is a genuinely stacked utilities roster for a seed.

The mechanic is the interesting part. Cloneable sends its system to shadow a human expert — think a senior lineman inspecting transmission towers or a rail safety inspector walking a track — capturing video, audio, and documentation of how they actually do the job. That trajectory gets cloned into an autonomous AI agent deployable on drones, mobile apps, and field hardware for the next generation of workers. They claim 100x ARR growth between February and end of 2025.

This is agent training at the trajectory level, not the prompt level. Most enterprise agent startups are wrappers around GPT or Claude — Cloneable is building a data-collection apparatus specifically for tacit knowledge in industries where the experts are retiring faster than they are being replaced. Energy, oil and gas, public utilities, vegetation management, rail, mining, agriculture. Every one of these is losing a generation of workers whose expertise was never documented.

$4.6M is small for the ambition, but this is the right shape of bet for vertical agents. Horizontal models get better every six months; proprietary trajectory data from expert workers in regulated industries is durable. Whoever captures the cloning step for utilities first gets a moat nobody can rebuild from a bigger model. The category to watch next is construction and healthcare, where the same knowledge gap exists but the compliance hurdles are higher.

https://news.crunchbase.com/venture/cloneable-cloning-expert-worker-knowledge-ai-infrastructure/
← Previous
Era raises $11M to be the AI layer for gadgets
Next →
context-mode cuts agent context by 98 percent
← Back to all articles

Comments

Loading...
>_