The New York company that stopped asking AI for advice and started handing it the job.
EXHIBIT A: The InstaLILY mark. Behind it, a quiet argument that the most exciting place for AI is the loading dock, not the demo stage.
Somewhere right now, inside a distributor's aging ERP, a claim is being read, a part is being matched to a machine, a quote is being pushed across a CRM. No human is doing it. There is no chat window open, no prompt being polished. The work is simply getting done. That worker has a name - an InstaWorker™ - and it belongs to InstaLILY AI.
InstaLILY sells something deceptively plain: AI Teammates for the physical goods economy. Not a chatbot. Not a copilot hovering over your shoulder with helpful suggestions. A teammate that opens the same software your staff opens, and finishes the same tasks your staff finishes - in distribution, supply chain, insurance, healthcare, and the other industries that keep the lights on and rarely get a keynote.
The company is small - around 110 people - and barely four years old. It is also, by the looks of a recent $25 million check, exactly the kind of unglamorous bet that serious investors have started to love.
Here is the uncomfortable truth the AI boom prefers not to mention: the most powerful models on earth are very good at telling you what to do, and almost useless at doing it. They draft the email. You still send it. They summarize the claim. You still key it in. The promise of automation arrived everywhere except the place it was promised - the actual work.
Distribution-heavy and regulated businesses felt this most. These are companies running on ERPs older than some of their analysts, with workflows stitched across a dozen tools that were never meant to talk. Traditional automation - the rules-and-bots variety - shattered the moment reality got messy. And reality, in a parts warehouse or a claims department, is always messy.
So the work stayed manual. Expensive. Slow. The kind of work nobody writes think-pieces about, which is precisely why nobody fixed it.
InstaLILY was founded in 2022 by Amit Shah, a former McKinsey executive, alongside co-founder and COO Sumantro Das. Their wager was contrarian for its time: don't ask companies to rip out their systems, and don't ask them to learn a new tool. Ask them to do something they already know how to do - hire a teammate.
The team they assembled reads like a recruiter's fever dream - engineers, operators, and researchers with stints at McKinsey, Meta, Google, Stanford, MIT, and Harvard. But the bet wasn't about pedigree. It was about a stubborn belief that vertical depth beats general cleverness. An AI trained on the specific grammar of distribution would outwork a brilliant generalist every single time.
It is a quietly radical idea. Most of the industry raced to build one model to rule them all. InstaLILY went the other way and built coworkers for industries that the one-model-to-rule-them-all crowd had never set foot in.
Amit Shah and Sumantro Das start InstaLILY in New York with a thesis: AI should do the work, not narrate it.
The team builds vertical AI Teammates that operate inside legacy ERPs, CRMs, and ticketing tools across distribution, insurance, and healthcare.
Insight Partners leads the round, with Perceptive Ventures and Marvin Ventures joining. The plan: more pre-trained InstaWorkers, deeper integrations.
InstaLILY lists its AI agents on the Google Cloud Marketplace, smoothing the path through enterprise procurement.
Strip away the branding and an InstaWorker is a domain-trained AI agent that logs into the systems you already run and executes real workflows - the sales follow-up, the service ticket, the claim evaluation - end to end. It reads the policy. It checks the SKU. It updates the record. When the stakes call for it, a human stays in the loop. When they don't, it just works.
The distinction InstaLILY keeps drawing is between suggesting and doing, and it matters more than it sounds. A suggestion still needs a person. Execution doesn't. That single shift - from advice to labor - is the entire company in one sentence.
Surface the right products, draft and push quotes, and keep the CRM honest - across thousands of SKUs no human can hold in memory.
Predict replacement parts, resolve tickets, and answer field questions inside the tools technicians already use.
Process claims, extract data from documents, and evaluate cases against coverage rules at high volume.
No rip-and-replace. InstaWorkers operate inside existing ERPs, CRMs, and ticketing platforms - the messy reality, not a clean slate.
Skepticism is healthy here - "AI that does the work" is a sentence every vendor on earth is currently saying. So look at what InstaLILY's customers report instead. A PE-backed insurance and healthcare services provider stood up an AI claims operations team to handle high-volume denials; the InstaWorkers extracted policy and claim data, evaluated it against coverage rules, and cut manual review time by 70%.
A construction supplier north of $10 billion put AI sales support behind 1,500-plus sales managers. An OEM equipment platform used AI service specialists to predict replacement parts across thousands of SKUs - the kind of needle-in-haystack matching that exhausts people and delights machines. Reported customers include names like Parts Town and SRS Distribution.
A 70% drop. The other 30% is where humans still get the final word - which, for denied health claims, is rather the point.
InstaLILY's stated mission is to build AI teammates that change how the physical goods economy runs. It is a deliberately unsexy frontier. The physical goods economy is forklifts and invoices and replacement gaskets - the plumbing of modern life that everyone depends on and almost no one studies.
The company's tagline puts it more plainly: Humans + InstaLILY AI = Superhumans. The framing is not about replacement but augmentation - giving the people who run distribution and service operations a teammate that absorbs the volume, so the humans can handle the judgment. With its Series A, InstaLILY plans to expand its catalog of pre-trained InstaWorkers across new verticals and deepen its integrations into common enterprise systems.
The bet under InstaLILY is bigger than any one warehouse. If AI's real value turns out to be labor rather than advice, then the winners won't be the companies with the cleverest chat interface. They'll be the ones that taught their agents the specific, tedious, high-stakes grammar of a real industry - and let them work.
That is the wager Insight Partners, Perceptive Ventures, and Marvin Ventures backed. It is also a useful test you can run yourself: ask any AI product whether it suggests or executes. The answer sorts the demos from the teammates.
So return to that distributor's ERP, the one humming quietly at the top of this page. A claim gets read. A part gets matched. A quote goes out. The screen looks exactly as it always did. The only thing that changed is who - or what - is doing the work. That, in the end, is the whole story InstaLILY is trying to tell: the most interesting place for artificial intelligence may not be the future. It may be the loading dock, today.