This AI Isn't Here to Assist You
On July 14, a company called InstaLILY raised a $60 million Series B. Led by Energize Capital, with Home Depot Ventures and United Rentals piling in as strategic investors.
Total funding sits near $100 million, and revenue grew 5x over the past year.
But the money isn't the story. The thing they built is — it's called Lily, billed as the world's first AI Forward Deployed Engineer.
It's not a copilot that helps you work. It walks into a business, learns how it runs, builds the software the work needs, and gets it live.
What That Actually Means
Most AI assistants help you do the work. Lily's logic is build and run the software that does the work.
Example: a national distributor whose sales team couldn't manually cover the entire customer base. Lily went in, built a piece of software that hunts for revenue opportunities across every customer, and generated over $200 million in new annual sales.
Or an industrial supplier where pricing complex RFQs used to be slow, manual grind. Now quotes come back fast, and the commercial team gets hundreds of thousands of hours back per year.
A field-service company had technicians spending 15 minutes diagnosing each problem. Now under 10 seconds. Cost-to-serve per call dropped 98%.
A logistics operator that planned routes in 15 minutes now does it in 3.
These aren't lab numbers. They're from businesses actively running.
How It Pulls That Off
InstaLILY partnered with Google DeepMind on a hybrid architecture: a large model for hard reasoning, a small model for fast and sensitive routine work, with each task auto-routed to whichever fits.
The small model runs on their own Small Data Center, built with NVIDIA tech — cloud, on-prem, or edge. Customer data stays private, latency stays low, cost stays down, energy too.
Different from most AI companies' shove-everything-into-one-giant-model approach. This one's pragmatic — small when it can be, big only when it has to be.
A Solid Client List
Already live with: SRS Distribution (part of The Home Depot family), United Rentals, ShipStation Global, PartsTown, Radwell, Henry Schein, PartsSource, Kedrion Biopharma.
All heavy, traditional industries — construction, industrial distribution, logistics, healthcare. Same pain point across all of them: too much work that only they know how to do, work too specific for off-the-shelf software, running on legacy systems and manual workarounds. That's Lily's lane.
But Let's Be Real
This is a Series B funding announcement. The numbers skew optimistic, for sure. $200 million in new sales — take it with salt. How much is Lily's work versus momentum that was already there? Hard to cleanly separate.
And AI forward deployed engineer sounds impressive, but the hard part isn't tech. It's that enterprise environments are constantly in motion — systems, data, regulations, even the models themselves shift. Lily sells continuous deployment, constant adaptation. Whether that promise holds up under long-term real-world wear, we'll see.
Industrial capital like Home Depot Ventures coming in means real deployments are running. But AI replacing forward deployed engineers? Still a long way off.
The direction is genuinely interesting. While most AI is still competing on who chats better or writes cleaner code, someone's heading toward just take over the ugly grunt work of real businesses. That's a lane worth watching.