QueryPal reads the answers your company already wrote - docs, closed tickets, internal systems - and turns them into agentic AI that drafts, measures and closes customer support.
It's peak season somewhere. A support inbox that normally hums is suddenly up 62% in volume, the kind of surge that historically means overtime, temp agents, and a manager refreshing a dashboard at midnight. Except this time the queue is draining on its own. Draft replies appear pre-written and mostly correct. The chatbot on the website is closing issues instead of collecting them. The humans are handling the handful of tickets that actually need a human. Nobody is panicking. That, in a sentence, is what QueryPal is built to make ordinary.
QueryPal is a San Francisco company with a stubborn opinion: the answer to most customer questions already exists inside the company being asked. It is sitting in a help doc nobody can find, a Slack thread that scrolled away, a ticket that was solved and closed nine thousand times before. The problem was never a shortage of answers. It was retrieval. QueryPal's whole pitch rests on that distinction - and on a second one it repeats like a mantra: it aims to resolve, not merely deflect.
The difference matters more than it sounds. Deflection sends a customer away - to an FAQ, a form, a dead end - and calls it a win because a human didn't touch the ticket. Resolution actually answers the question. One number goes down on a cost report; the other goes up in a customer's memory. QueryPal decided to chase the harder one.
Built for resolution, not just deflection.- QueryPal's founding principle, printed on the tin
Dev Nag has done the enterprise-software tour: engineer at Google, PayPal and eBay, a founding engineer at trading platform GLMX (which touches trillions in daily balances), and then the one everybody cites - founder and CTO of Wavefront, the real-time monitoring company acquired by VMware, where he went on to help launch their AIOps product.
He kept seeing the same shape of problem. In observability it was signal drowning in noise. In enterprise support, he decided, it was the same story wearing a different badge: knowledge existed, but nobody could reach it fast enough to matter. QueryPal - which began life under the name CtrlStack - is the answer he built the second time around.
The company's ambition is not simply cheaper support. Nag frames it as moving support teams "from reactive responders to strategic contributors" - letting AI carry the repeatable load so people can do the part that needs a person.
The human knowledge layer for secure AI.- How QueryPal describes itself
QueryPal's suite reads like a three-act play - draft it, measure it, resolve it - all fed by the same underlying knowledge the company already owns.
Plugs into email and generates draft replies from your existing knowledge base - reaching 90%+ approval from the agents who ship them.
Enterprise analytics on tickets and customer sentiment, surfacing trends and driving cost per resolution down.
An agentic chatbot that autonomously deflects and resolves issues using company docs, websites and internal systems.
P oint QueryPal at the knowledge you already have and a few things start happening. Agents open a ticket to find a draft reply waiting - most of them good enough to send. Customers on your site get answered instead of routed. Leaders open Prism and see which issues are burning the most time and money. And when volume spikes, the team absorbs it without a hiring scramble.
The proof QueryPal points to is blunt: one customer, Simply Benefits, reportedly ripped out an incumbent AI - Freddy AI - in the first week, and clocked response times roughly ten times faster. Across deployments the company cites a 60% cut in cost per resolution within a month. Numbers to take with the usual grain of salt for a young company, but the direction is the point.
AI that acts intelligently - and securely.- The other half of the QueryPal promise
Company founded (as CtrlStack) around real-time troubleshooting.
Launches with $5.2M seed co-led by Sequoia Capital and Engineering Capital.
Rebrands to QueryPal, consolidating around Intercept, Prism and Concierge under a "secure AI" banner.
Founder-interview run across Pulse 2.0, AITech, Unite.AI and H.I.E.C. lays out the resolution-first thesis.
QueryPal is chasing a support-software market it sizes at $300-400 billion - as an 11-person team. The wager: a sharper answer beats a bigger org.
QueryPal started as CtrlStack, an observability startup - then pivoted into customer support.
Its X/Twitter handle is still @CtrlStackHQ, a fossil of that first identity.
Founder Dev Nag's resume runs Google, PayPal, eBay, GLMX and Wavefront.
The AI is built to cite its sources - a quiet jab at "trust me" chatbots.
One customer swapped out an incumbent AI vendor within a single week.
Return to that inbox at peak season - volume up, nobody panicking. What changed isn't that the questions got easier. They didn't. What changed is that the answers the company had written down for years finally started answering back, at the speed a customer expects and with a source attached.
That's the modest, almost unglamorous promise underneath QueryPal's agentic-AI language: put a company's own knowledge to work, close the ticket, and let the people spend their attention where a person is actually needed. Whether an 11-person team can hold that line against a field of well-funded rivals is the open question. But the queue that empties itself is a good place to start an argument.