David Lanstein Is Building the Search Bar His Old Team Needed

At PagerDuty, David Lanstein could find an internal answer in about 30 seconds. A new hire could spend two hours looking for the same thing. The pricing deck existed. The customer notes existed. So did the usage dashboard and the healthcare case study. They were simply scattered across Salesforce, Slack, Google Drive, Looker, Box and Confluence, each one waiting for somebody who remembered the route.

Lanstein remembered. He had joined early, helped build the sales organization and accumulated the kind of private map that looks like instinct from the outside. He knew which dashboard held a number, who carried a useful piece of competitive intelligence and where the last representative had left an account note. His fluency was valuable. It was also a warning: a growing company was depending on tenure as a search engine.

He tried the normal fixes. He made calendar routines for checking NPS responses. He kept Evernote pages full of links. He opened Slack channels around particular topics. The system worked for its author because its author had spent years building it. It could not be handed to the next person as easily as a laptop and a login.

“Search is a revenue blocker disguised as an IT ticket.”David Lanstein, reflecting on the PagerDuty years

That small piece of operator frustration became the seed of Atolio, the company Lanstein founded in 2019 with Gareth Watts and Mark Matta. It is tempting to describe the business as one search bar across all of a company's software. The more revealing description is a machine for giving a new employee some of the context an old employee carries around without thinking.

First, he learned the machinery

Lanstein's route to the chief executive's chair did not run through one discipline. He worked as a software engineer at Transitional Data Services, then spent more than three years as a senior software engineer at Splunk. A small artifact from that period still sits in public view: a Salesforce Python Toolkit package released in 2011 under his name. It is an unshowy clue to the shape of his career. Before he sold enterprise software, he built it.

Then the work moved closer to customers. He led major accounts at Loggly and joined PagerDuty in early 2013 to head major accounts and sales. PagerDuty's welcome announcement introduced the former Splunk coder with a joke about his skill at opening champagne. The more consequential pairing was engineering fluency with a commercial job. Lanstein could talk to a buyer about the result and still think in terms of systems underneath it.

PagerDuty supplied the operating education. Lanstein and Matta helped start its sales team when the company was small. As it expanded, the information problem became visible in the distance between insiders and arrivals. The business added people and applications. Each tool made one function better while making the whole company harder to see at once.

The missing signal was a person

A conventional search engine asks whether a document contains the right words. Atolio's thesis adds another question: right for whom? A salesperson and an engineer can enter the same query and reasonably need different results. Their projects differ. Their closest collaborators differ. The documents they are permitted to open differ. At work, relevance is partly social.

This is where Watts's infrastructure background and the founders' shared history become important. Atolio builds a collaboration graph from the activity around workplace information: who works with whom, which projects occupy them and which material is fresh or widely useful. That graph helps rank a corpus that would otherwise produce thousands of technically matching, practically useless results.

Before writing code, the company says its team interviewed 762 enterprise leaders, more than 100 of them from the Fortune 1000. The conversations reinforced a second constraint. For sensitive organizations, finding the right document is only half the work. The system must refuse every wrong document just as reliably.

The architecture is the promise

Atolio runs inside a customer's chosen environment, including AWS, Azure, Google Cloud, GovCloud, on-premise and air-gapped deployments. The index, connectors and model orchestration remain within that boundary. Document-level access controls from connected applications continue to decide what each employee can retrieve. Customers can bring their own model keys.

That design choice narrows the easy path. Software deployed inside somebody else's infrastructure is harder to operate than a conventional multi-tenant service. But for financial institutions, public agencies and other regulated buyers, the constraint is often the admission ticket. “Data cannot leave” is an operating condition, not copy for a security page.

In September 2025, Atolio announced that a Series A led by Translink Capital had brought its total funding to $24 million. IBM Ventures, Bloomberg Beta, Acorn Pacific Ventures and Parameter Ventures participated. The company said it had signed multiple seven-figure contracts over the preceding three quarters. That August, Atolio began work under a $1.25 million U.S. Air Force SBIR Direct-to-Phase II contract focused on secure organizational search and intelligence.

The contract is a useful test of the founding premise. In a defense setting, permission errors are not an irritating product defect. Information can be distributed across environments, yet access must remain specific to the person asking. The conditions make Atolio's central tradeoff unusually visible: broad discovery paired with narrow authorization.

One box, fewer decisions

Lanstein's public writing returns to usability as often as it returns to security. Many enterprise AI products divide the screen into search, questions, agents and knowledge tabs. Atolio deliberately presents one box. The user does not have to decide whether a sentence is a query or a question, select a model or learn the vendor's internal taxonomy before asking for help.

The restraint matches the original problem. A new employee already carries enough uncertainty. Which system contains the file? Which words did its author use? Which version is current? Who knows the history? Adding another choice at the search bar only converts product architecture into user homework.

“A company's knowledge is the work product of their people.”David Lanstein, on Atolio's 2025 funding announcement

His more recent argument stretches beyond human search. As companies deploy AI agents, those agents will need access to internal knowledge. Every retrieval call raises the same questions an employee search does: which source is current, which context matters and what may this identity see? In that world, enterprise search becomes less like a destination and more like a governed layer under other software.

Lanstein brings a measured skepticism to the present agent rush. In a 2026 HumanX conversation with former Fortune 50 innovation leader Doug Shean, the pair focused on the distance between corporate AI claims and systems that actually reach production. Their advice leaned toward specific bottom-up work, short pilots and retrieval that respects permissions. It sounded less like a futurist's manifesto than an operator inspecting the plumbing.

A map that can leave its maker

Atolio remains a relatively small company working on an old category. Search inside companies has disappointed buyers for decades because workplace information is unlike the open web. Links are weak signals. Permissions change. Identities differ across tools. Valuable material may sit in a private channel or a barely linked file. The answer for one colleague can be noise for another.

Lanstein's career gave him several views of that failure. The engineer saw data systems. The salesperson saw customer urgency. The early employee saw an organization accumulate knowledge faster than it could distribute context. The founder's move was to treat those as one problem.

There is something pleasingly circular in the result. His PagerDuty workaround was a graph before it had the name: a set of links, reminders, colleagues and remembered locations that routed him toward an answer. Atolio is an attempt to make that map dynamic, governed and available to the person who started last week.

The aspiration is larger now. Lanstein wants the same layer to serve employees and the agents beginning to work beside them. But the standard remains grounded in a scene from years ago: one veteran, one new hire, the same question and a gap of almost two hours. A useful system closes that gap without opening a door the new hire was never allowed to enter.