Danny Shteinberg Is Turning the Corporate Org Chart into a Map of Possibility

The question arrived before the company did. Danny Shteinberg was working at what he later called one of the biggest and brightest companies in the world, yet his career felt oddly stationary. A large organization should contain a large number of possible futures. From the employee’s chair, however, those futures can be harder to see than a vacancy advertised by a stranger.

That contradiction bothered him until it became, in his phrase, a “need-to-do-something-about-this” problem. It is an appealingly practical origin story. No thunderbolt. No garage mythology. Just the quiet irritation of knowing that opportunity must exist somewhere in the building, combined with the suspicion that the building has misplaced its map.

In 2015, Shteinberg joined engineers Ben Reuveni and Amichai Schreiber to start Workey. The first map pointed outward. Workey let employed people explore other jobs anonymously. A candidate’s profile stayed concealed until the candidate chose to reveal it, while software compared skills and career history with millions of possible paths. The founders pitched privacy as both courtesy and correction: employers could consider capability before race, gender, ethnicity or religion entered the frame.

A clever answer finds a larger question

By 2017, Workey had worked with more than 400 companies, including Yahoo, Amazon, Dell EMC and Oracle. An $8 million Series A funded a New York office and more hiring. Shteinberg, then the company’s marketing leader, argued that conventional recruitment was not merely clumsy. It encouraged people to jump from one job to another without a clear idea of whether the move served a longer career.

Workey’s promise was to make the leap less embarrassing and more informed. The product even avoided the word “applicant.” A person could express willingness to work somewhere; if interest was mutual, the system proposed an introduction. Rejection could occur without ceremony. It was online dating with fewer profile photographs and, one imagines, more JavaScript.

Yet the founders kept running into the question hidden inside Shteinberg’s original complaint. Why did ambitious employees have to look outside at all? A company with thousands of people might contain an open role in another region, a short project on another team, a mentor two floors away or a useful skill nobody had bothered to record. External job boards advertised possibility with neon lights. Internal opportunity often lived in a spreadsheet with permissions trouble.

So Workey turned inward and became Gloat. The company began building what came to be called an internal talent marketplace: software that could connect employees with full-time roles, projects, learning and mentorship inside their own organization. Shteinberg has described the route there as countless brainstorming sessions, conversations with executives, frustrating but useful setbacks and late-night takeout dinners. Enterprise software rarely gets a montage, but if it did, the hero would be holding a carton of noodles while arguing about skill taxonomies.

The product and the pitch grew up together

Shteinberg’s role at Gloat is unusual because it has straddled the making of the product and the making of its language. He served as chief marketing officer, later carried both marketing and product titles, and is now listed by Gloat as co-founder and chief product officer. The progression matters. Gloat has not simply shipped features; it has repeatedly tried to define the category those features belong to.

First came the talent marketplace. Then workforce agility, skills infrastructure and the Workforce Graph. Each phrase widened the aperture. A marketplace shows opportunities. A graph attempts to show the relationships among employees, skills, jobs, projects, mentors and courses. By 2025, Gloat had added work and technology to that model. Its Mosaic product could break an objective into tasks, identify the skills involved, and suggest a mixture of people and software to do the work.

This is where Shteinberg’s engineering background and marketing instincts meet. The technical problem is semantic: “Python” in one system must relate sensibly to an AI engineering role, a project brief and the experience described in a person’s work history. The communication problem is equally stubborn: employees do not wake up longing to normalize workforce data. They want to know what they can do next.

Gloat’s growth supplied evidence that large employers shared the concern. A $57 million Series C in 2021 was followed by a $90 million Series D in 2022. The latter announcement said the company’s customer base had grown two and a half times in a year, while annual recurring revenue tripled. Gloat cited deployments at companies including Unilever, Schneider Electric, Seagate, Nestlé, Standard Chartered and HSBC. At Unilever, its marketplace was credited with unlocking more than 750,000 hours across over 4,000 projects. The abstraction called “mobility” had acquired a timesheet.

Job codes in a trench coat

The arrival of generative AI gave Shteinberg a fresh target: software that looks intelligent because it speaks fluently while standing on thin data. In one public post, he mocked an agent that merely filters job codes, tenure, ratings and old skill tags as “a SQL query in a trench coat.” The line is funny because it identifies a serious product risk. Grammar can disguise retrieval. It cannot manufacture context.

He illustrated the distinction with a succession-planning demo. The prompt asked for a plan for a vice president of AI research and development. The response grouped candidates by readiness, identified individual gaps, suggested development paths, summarized the choices and proposed actions for an HR partner. Shteinberg said the demonstration took 20 seconds, compared with weeks of spreadsheet work.

The claim rests on what Gloat calls Loomra, its Workforce Context Engine. Instead of treating a person as a row in a human-capital database, the system is designed to connect work performed, projects led, skills demonstrated, career direction and organizational patterns. The agent is the conversational surface. The bet is buried underneath, in nearly a decade of modeling how skills and careers relate.

That distinction also explains Gloat’s 2026 move into Microsoft Teams and Microsoft 365 Copilot. Traditional HR software waits in its own portal. Employees visit when obliged, then leave with the speed of people exiting a dentist’s office. Gloat’s newer premise is that career guidance, redeployment and succession work should arrive inside the conversation where the question arises. The capability comes to the employee rather than demanding another login.

A career made of smaller pieces

Shteinberg’s public writing has become increasingly concerned with what happens when AI breaks a job into tasks. He has argued for fractionalizing work rather than people: let machines take the rote pieces while people keep judgment, relationships and strategy. When his eight-year-old asked whether robots would need a lunch break, Shteinberg reached for Lego. The child’s future, he suggested, may be less about a fixed title than about swapping blocks of work with human and machine teammates.

There is a tension here worth keeping. Decomposing jobs can create mobility, but it can also turn a career into a tray of interchangeable parts. Gloat’s stated answer is development: show adjacent skills, connect people to projects, recommend mentors and make the next move legible. Whether software can preserve agency while optimizing allocation is one of the defining questions in workplace AI. A clean graph does not absolve a company from making humane choices.

Shteinberg’s decade is interesting because he has followed the same discomfort through several product forms. Workey asked how a person could look elsewhere safely. Gloat asked how that person could move internally. Workforce intelligence asked what the organization actually knew about its people. Work orchestration asked how people and technology could share a task. Agentic HR now asks whether the system can act on that context at the moment a decision is made.

The labels have multiplied, as labels do when enterprise software meets a conference stage. The sturdy idea beneath them is simpler. People contain more possibility than their current job title records. Companies are generally poor at seeing it. Shteinberg has built his career around making that gap visible, then turning visibility into a next step.

An org chart will always have boxes. Bureaucracy adores a rectangle. The more useful map is the one drawn between them: a skill learned here, a mentor found there, a project that becomes a new career. Shteinberg’s wager is that those faint lines can be read by software without becoming less human. The work now is to prove that a map of possibility can guide people, not merely route them.