Larry Smith and the art of the shorter wait

Twenty minutes is an awkward amount of time. Too long to pretend you have not been waiting. Too short to make a proper afternoon of it. In 2017, Larry Smith offered a more productive use for the interval: a developer on his team, with no previous machine-learning experience, had built and deployed a model in less than twenty minutes. Putting that model to work inside Provenir’s platform took hours. Smith said the same undertaking would historically have occupied a team for a month.

The anecdote has the pleasing proportions of a founder’s demonstration: modest cast, compressed running time, visible result. Its subject, though, is revealing. Smith was discussing the distance between an analytical idea and something a lender could actually use. A prediction sitting on a developer’s screen cannot approve an application. It needs somewhere to go.

That gap is where Smith has built his working life. As Provenir’s founder and chief executive, he sells software for the machinery behind financial decisions: gathering information, applying models and rules, and moving an application toward an answer. The customer might never learn his name. Ideally, the customer barely notices the wait.

A useful place for an impatient idea

Provenir’s own product history places its first platform decisioning solution in 2004. On-premise and cloud-managed options were part of that early offering. The vocabulary has expanded since then, through predictive analytics, software as a service and artificial intelligence. The central job has remained recognizable: give financial institutions a way to make credit and risk decisions with less manual handling.

There is a particular kind of entrepreneurship in choosing this territory. The visible part of banking has obvious charms: the app, the card, the reassuring little notification. Smith’s business lives behind those surfaces. A lender needs to bring information together and decide what to do with it. The software that connects those steps rarely gets a glamorous introduction.

Smith’s public professional profile uses Lawrence Smith and lists New York University. His company operates from Parsippany, New Jersey, while his LinkedIn profile places him in New York. His career also includes founding Kitewheel, where he served as chairman. That combination places him across financial decisioning and customer engagement, two businesses concerned with what an organization should do when a customer appears.

The connection becomes more explicit in TruNarrative, described in its 2019 US expansion announcement as a collaboration between Smith and John Lord, formerly a group managing director at GBG. Its problem was the untidy business of connecting identity, fraud and compliance data. Different providers supplied information in different formats. Someone had to make the pieces cooperate.

Across those ventures, the recurring commercial problem is fragmentation. A company may possess the information and expertise it needs, yet struggle to connect them in a working process. The repetition is worth noticing. Smith has returned to businesses where an answer becomes valuable only when it can travel.

The twenty-minute test

The 2017 machine-learning example gives that idea a human scale. Provenir had introduced an adapter for Amazon Machine Learning. A predictive score could feed into its risk decisioning process, which could then execute a pass, fail or referral. The demonstration connected the clever part of analytics to the comparatively uncelebrated business of getting something done.

Smith emphasized the obstacle facing institutions that could not devote full-time resources to developing their own machine-learning capabilities. This was a practical sales argument. A financial business has customers to serve and products to run. It may welcome a better prediction while being considerably less enthusiastic about assembling a new specialist department.

“With no experience in machine learning” ... “less than 20 minutes”

Larry Smith, describing a team developer’s experiment, 2017

The time claim belonged to a specific example, rather than a promise that every institution could repeat the exercise on command. Its appeal was nevertheless clear. Smith was offering a picture of expertise becoming easier to use. The developer’s lack of previous experience was the point of the story.

Even the three possible outcomes deserve attention. Pass. Fail. Refer. The last word gives the system a little humility. Some applications need further consideration. Good automation has to make room for that possibility, rather than treating every uncertain case as an embarrassment to be hurried out of sight.

When the waiting became a business

By 2020, the demand for faster answers was showing up in Provenir’s operating figures. Its SaaS revenue rose 25 percent year over year, and its customer base increased 22 percent. The company reported processing more than two billion transactions across eleven data centers. These were business results for the company Smith led, rather than personal performance statistics, but they give his argument some weight.

In April 2021, he put the customer expectation plainly: “Customers are demanding immediacy in all interactions”. Provenir was expanding its presence in buy now, pay later and challenger banking, where a lengthy pause sits rather awkwardly beside the promise of a quick digital experience.

The company also said it invested a quarter of revenue in research and development during 2020. That choice belongs beside the growth figures. Keeping a decisioning platform useful requires work beyond winning the next customer: connecting new data, adapting workflows and making the technology usable as its buyers change their own businesses.

In 2021, SaaS revenue grew another 55 percent. The company reported a 194 percent increase in new customers, a separate measure from growth in the total customer base. Its workforce expanded 49 percent, and the markets it served approached fifty countries. A useful story about a developer had become a business operating across many more desks, currencies and time zones.

Smith described the platform around three components: data, AI and decisioning. Provenir introduced Cloud 2.0 and Provenir AI, while its marketplace added more than sixty partners. The idea was to bring the ingredients closer together. A financial institution should be able to move from information to analysis to an action without repeatedly starting the plumbing from scratch.

The geography of an answer

Expansion brought another test. In 2022, Provenir established a global technology and operations hub in India and introduced Provenir Brazil. Revenue increased 35 percent, while the customer base grew 24 percent. These figures describe different measures from the earlier SaaS numbers, so they deserve their own line rather than a conveniently dramatic growth curve.

The named customers added that year included Davivienda, Varo, AutoChek and AMU Leasing. The range matters to Smith’s story. Selling decisioning software across markets means dealing with institutions whose customers, products and available information differ. The software has to accommodate those differences while preserving a process its users can understand and operate.

Connections with other financial infrastructure businesses helped extend that reach. Mastercard included Provenir in its Engage network for digital-first solutions, assigning it the real-time risk decisioning role alongside businesses supplying other parts of the customer journey. Provenir also announced collaborations with TransUnion and participation in Visa Ready for BNPL during 2022.

This is a less theatrical version of the founder’s network. The connections are operational: a data provider, a payment network, a lender, a software platform. Their value lies in what can pass between them. For Smith, partnerships form part of the product’s working environment, rather than merely furnishing a handsome collection of logos.

People behind the automatic reply

Recognition arrived along the way. Smith was named among The Financial Technology Report’s Top 50 US Financial Technology CEOs for 2021. Provenir also appeared on the 2021 Inc. 5000, at number 4,430. The two distinctions measured different things: one recognized an executive; the other ranked a private company’s growth.

As the organization expanded, the personnel decisions became part of the narrative. Provenir appointed Lori Copeland chief people officer in January 2024, with a workforce then spanning twenty-five countries. Smith pointed to her organizational-development experience and the importance of supporting people operations as the business grew.

It is an instructive detail in a story about automation. There are people behind the system that replies automatically, and more people behind the systems that help it reply. A geographically dispersed workforce needs its own infrastructure. Software can remove a wait from a loan application without removing the work of running a company.

What happens after yes?

In February 2026, Provenir introduced a decision intelligence platform with agentic AI features. The company described an environment joining data, models, decisioning and optimization, with simulation tools and an embedded AI assistant. Human oversight, transparency and explainability formed part of the announced approach.

The change extends the story beyond reaching an answer quickly. It asks what happens after the answer: whether the decision produced the intended outcome, and what should change next time. Under Smith’s leadership, the product’s ambition has moved toward connecting execution with feedback. The commercial proposition is a continuing cycle of decisions and learning.

The developer’s twenty-minute experiment remains a useful way into that ambition. It begins with an ordinary obstacle: putting an analytical idea to work takes too long. Smith’s career connects him to several attempts to shorten that distance. At Provenir, the consequence is measured in the decisions financial institutions make every day. For the person waiting on the other side of the screen, the best evidence may be wonderfully uneventful: an answer arrives, and the day moves on.