Matt Singer Keeps Choosing the Hard Categories

Matt Singer has spent much of his career selling software that arrives with a glossary. Applicant-tracking systems. Municipal budgeting platforms. Translation orchestration. These are not products that leap off a shelf. They enter organizations by committee, survive procurement, unsettle an old routine, and eventually become part of how work gets done. Marketing them requires patience with detail and a firm belief that detail need not be dull.

Singer’s answer is an unusual mixture of measurement and mischief. He calls himself a data geek, but holds a master’s degree in journalism and media studies. He talks about product-market fit and ideal customer profiles, then opens a job announcement by calling Britney Spears a poet. For the 2026 launch of LILT’s Model Context Protocol server, he borrowed a line from 1990s hip-hop and Al Pacino before explaining enterprise AI plumbing. The cultural references are not the point. They are the door handle.

Now based in San Diego, Singer is chief marketing officer at LILT AI, a company that builds translation technology for enterprises and government agencies. He arrived in 2025 after five years at OpenGov, where his tenure coincided with nearly tenfold company growth and ended in the year Cox Enterprises acquired the business at a $1.8 billion valuation. Before that came Jobvite, SuccessFactors, HumanConcepts, Jabra, and Cantor Fitzgerald. The industries changed. The stubborn question remained: how do you make a complicated system intelligible without sanding away what makes it valuable?

Cookies, lemonade, and a taste for the measurable

Singer’s unofficial commercial education began at eight, according to a biography published during his Jobvite years. He tried cookies, lemonade, and lawn mowing in his neighborhood. The childhood-enterprise anecdote is charming because the adult version is easy to recognize: find demand, make an offer, see what happens. It also explains the self-description that followed him through HR technology - a “data geek” interested in building brands and marketing machines that could scale.

His formal education broadened that instinct. At the University of California, Santa Cruz, he studied economics and environmental studies. At the University of Colorado Boulder, he earned a master’s in journalism and media studies. One discipline asks what the numbers reveal. The other asks how a person will understand them. Singer’s career sits comfortably between those questions.

Before Jobvite, he worked across sales and marketing roles at Cantor Fitzgerald, Jabra, HumanConcepts, and SuccessFactors. His SuccessFactors work focused on demand generation and marketing operations. Jobvite hired him as vice president of marketing in 2015, noting that he already had more than 15 years of sales and marketing experience. He later became executive vice president of marketing.

At Jobvite, Singer’s public writing treated recruiting as a cousin of marketing. Candidates were an audience. Employer brands were brands. Application friction mattered. Automation could take on scheduling, distribution, and screening, but the human implications remained central. In a 2018 interview, he argued that marketing needed a stronger place in product-development cycles and placed a premium on authenticity. Long before generative AI made “human in the loop” a boardroom phrase, his subject was already the border between systems and people.

Five years, fewer than five marketers, one public mission

Singer joined OpenGov in 2019. Government software offered a different difficulty from recruiting technology. A buyer might be a city, county, school district, or state agency. The purchase could touch budgeting, procurement, licensing, public assets, or the way residents find information. Sales cycles were complex because the work was complex. The emotional payoff, however, was plain: software could help a public servant answer a resident faster, make a budget easier to inspect, or keep an agency running when its offices closed.

During the pandemic, Singer described customers using OpenGov tools for virtual town halls, testing-site pages, and online budget access. The product story was no longer merely cloud migration. It was continuity. At a moment when municipal buildings were quiet, the work of government still had to reach the public.

His remit covered product marketing, demand generation, communications, and enablement. Yet when Singer announced his departure in late 2024, the detail he lingered over was the team. OpenGov had fewer than a handful of marketers when he arrived, he wrote, and the organization had been assembled “brick by brick.” It clarified product-market fit and its ideal customer profile, launched campaigns, and built a recognizable position in government software.

“Calm seas don’t make good sailors.”Matt Singer, closing his OpenGov chapter

The nautical line suited someone who has described himself, with cheerful imprecision, as a “manphibian.” Outside work, Singer has long listed surfing, fishing, and abalone diving among his interests. The water also gives his career metaphor a bit more texture. He seems less attracted to calm categories than to markets in motion, where incumbents are large, the route is uncertain, and the product still needs a story.

OpenGov’s destination was significant. Cox Enterprises, already an investor, acquired the company in 2024 in a deal valuing it at $1.8 billion. Singer framed Cox as a long-term home for the mission rather than a short-term financial owner. He did not claim the transaction as a marketing trophy. In his farewell, pride in the valuation sat beside marriages, births, difficult debates, long travel, and colleagues who had grown with the organization. Companies scale in charts; teams experience the scale one week at a time.

The word after translation

After a break to settle into a new community and spend time with family, Singer joined LILT. He described the company as another insurgent with an unusually good product in a field of legacy incumbents. It was, in his phrase, “a marketer’s dream”: a category disruption in which the marketing could influence the company’s outcome.

Translation looks simple only from far away. A literal sentence can be technically accurate and commercially wrong. Product terminology must remain consistent. Legal language has little room for improvisation. A global campaign has to sound like the same brand without sounding as though every market lives in the same place. “Affordable premium” cannot casually become “cheap.” A British “sale” does not travel into French by swapping one dictionary entry for another.

Singer’s public case for LILT begins there. The company can build or adapt domain-specific models for a customer, then use corrections from human verifiers to improve them continuously. Different functions may need different models. Product copy, employee material, support text, and advertising each carry their own vocabulary and tolerance for error. The machine provides speed and scale. Human judgment protects the last, consequential inch.

In a 2026 conversation with Lenovo digital marketer Ashwin Naiksatam, the operational promise became concrete. Lenovo was running more than 60 domain-specific models across 39 languages. Reported delivery times had fallen by as much as 60 percent, with average annual translation costs down by more than half against its previous provider. The figures belonged to the customer story, not to Singer personally. His role was to make the mechanism legible: continuous learning, brand-specific context, and human review where it matters.

He also shifted the frame. Localization, in this telling, is not the back-office errand at the end of a campaign. It is part of customer experience from search to conversion. LILT’s connectors let teams submit work from tools including Adobe and Figma, avoiding the small errors and delays created by moving between systems. The less translation feels like a separate department, the more naturally multilingual work can enter daily operations.

A CMO with one eye on the interface

Singer’s next bet is that the interface itself will change. He expects enterprises to put company-specific AI assistants on employees’ desktops. Through Model Context Protocol integrations, a worker could ask for brand-aligned translation or expert verification inside that conversation, without opening a separate localization application. Translation becomes less like a destination and more like a capability waiting wherever the work begins.

It is a technically current idea with an old marketing concern underneath: meet people in the workflow they already inhabit. Singer made a similar argument at Jobvite when he wanted recruiting to behave more like marketing, and again at OpenGov when cloud tools brought public services online. Tools matter most when they reduce the distance between an intention and an action.

There is a tidy temptation to call Singer’s path a sequence of reinventions. The record suggests something more continuous. He has kept moving toward categories where language carries operational weight. A candidate decides whether to apply. A resident tries to understand a budget. A global customer decides whether a brand sounds credible in her own language. In each case, the words are attached to a system, and the system is attached to a human decision.

The little jokes help. So do the metrics. Neither works alone. A spreadsheet cannot give a category a pulse, and a clever line cannot rescue a product that fails its customer. Singer’s particular lane runs between the two - evidence on one side, narrative on the other, and difficult software waiting to be understood.