Avi Korenblum arrived with a yellow sheet of paper. The idea on it was ambitious enough to make the stationery seem underqualified: take the sprawling, untidy conversation happening online and turn it into something people could use to make decisions. His audience was Ronald Cohen, a co-founder of Apax Partners. Cohen wanted to invest. Lloyd Dorfman, the founder of Travelex, came into the picture too. A company began to take shape around a problem most internet users encounter daily, then solve by closing another tab.
That company was Voyager Labs. Korenblum founded it in 2012 after two decades in Israel’s intelligence community. By its public debut in November 2016, it had raised more than $100 million. The contrast still catches the eye: an enterprise devoted to interpreting the information people leave in public had spent its first four years largely out of public view.
The yellow paper is an appealing opening. What followed is more complicated, and more revealing. Korenblum’s career at Voyager runs through the commercial possibilities of artificial intelligence, the effort to make software useful to investigators, and the argument over who gets to collect the digital traces of other people’s lives.
A founder who hired the technical answer
Korenblum’s starting asset was experience in intelligence rather than a technical background. He described his contribution as identifying a need and assembling people capable of solving difficult problems. It is a practical definition of founding: first understand what an answer must accomplish, then recruit the people who can build the machinery.
That division of labour matters to his story. Voyager’s early identity rested on the combination of a founder familiar with decisions under uncertainty and a research team working on artificial intelligence. Intelligence work supplied the question. Researchers and engineers had to turn it into a product that somebody outside that world would pay to use.
The company’s original pitch reached beyond government. Its launch described customers and applications across retail, consulting and financial services. The ambition was to make publicly available information usable across different markets. A company asking what customers might buy and an investigator asking how people are connected could, in this model, begin with similar raw material.
Company capital, not Korenblum’s personal wealth.
Four years behind the curtain
The investors were part of the story from the beginning. Alongside Cohen and Dorfman, Voyager’s backers included Horizons Ventures and OCAPAC Holding Company. The disclosed funding accumulated across rounds before the public launch. Reading the number as a single cheque would miss the time spent building the business beneath it.
By November 2016, Voyager reported more than 90 employees, research and development in Israel, and offices in New York, Washington and London. It emerged with customers and use cases as well as capital. The geography was already international; the challenge was to make a shared technical engine useful in settings with different customers, incentives and consequences.
Stealth supplied a neat headline, but it also makes the launch an unusual point in a founder’s biography. The public introduction came after years of hiring and development. Korenblum was presenting an organisation already in motion. The curtain rose on a working company, complete with the less theatrical obligations of serving customers.
The internet refuses to fill in the form
Voyager worked on unstructured data: information that arrives as posts, language, pictures and connections rather than obedient rows in a spreadsheet. Its proposition was to find useful patterns within that material. The difficulty is familiar. People write in fragments, switch subjects, imply things and assume their audience knows the context. The internet is an energetic correspondent and a dreadful filing clerk.
In a 2017 explanation of the products, Voyager’s executives described a process of collecting information, applying behavioural analysis and displaying connections. Voyager Analytics, Voyager Finance and Voyager eCommerce shared underlying technology. The intended result varied: an investigative lead, a view of creditworthiness, or a product recommendation.
Those are distinct decisions. A connection on a diagram does not carry the same consequences in a shopping recommendation as it does in an investigation. The commercial breadth of the early business therefore exposes a central question in Korenblum’s project: how should an interpretation of human behaviour become something another person can responsibly act on?
The interpreter takes the stage
At the December 2018 Mind the Data conference in Tel Aviv, Korenblum discussed the pressure created by accessible information arriving faster than people can assess it. His emphasis included the need to combine artificial intelligence with human work. Speed, in that account, was essential to getting a useful answer in time.
He demonstrated the argument through social posts concerning France’s Yellow Vest protests and Iran. He presented the software as capable of making inferences about affiliation, motivation and context. These were claims about interpretation, not simply retrieval. The demonstrations showed the kind of answer Korenblum wanted his company to sell; they did not independently establish the accuracy of every inference.
His appearance on the 2019 DMBI conference programme at Ben-Gurion University made the direction explicit. His keynote was titled “AI for Public Safety.” A founder whose launch covered finance and commerce was also explaining his work to a research audience through the demands of public safety.
Recognition, then distribution
During Korenblum’s tenure, Voyager collected recognition for specific applications. It was included in Gartner’s 2017 Cool Vendors report for AI in banking and investment services. VoyagerAnalytics also received Frost & Sullivan’s 2017 European award for new product innovation in AI-based social behaviour analytics.
These distinctions belonged to the company and its products. Korenblum’s response to the Frost & Sullivan award credited customer relationships and the collaboration of his team. That is consistent with the founder’s early account of his job: assembling the people who could convert a need into working technology.
A different kind of milestone came in March 2021. Voyager announced its investigation platforms on Microsoft Azure, including VoyagerInsights. It also described co-sell ready status in Microsoft’s partner programme. Distribution had joined invention as a practical concern. An investigation product must be deployed, supported and brought into the organisations expected to use it.
The promise meets the permission
Privacy was already part of Korenblum’s public position in 2017, when Voyager announced that it had joined the Future of Privacy Forum. He said: “Technology is always going to outpace regulation”. The company presented respect for privacy as a core principle and described its searches as restricted to public or consensually available data.
The distinction between available information and permitted collection later became consequential. In January 2023, Meta sued Voyager Labs, alleging that it used fake accounts and unauthorised automated scraping to collect data from Facebook and Instagram. Those were allegations against the company. They should not be recast as a finding of personal misconduct by its founder.
In December 2024, Meta announced a settlement involving a permanent injunction and a monetary payment. The episode belongs in Voyager’s history alongside its funding and awards. It puts a hard edge on the original business proposition: the ability to gather and interpret information also requires an answer about permission.
“Technology is always going to outpace regulation”
Avi Korenblum, July 2017
A partnership announced with Voyint in August 2022 showed what the company’s public-safety focus meant in practice. Voyager supplied investigation technology; Voyint brought research and analysis services for federal customers. Their announcement described a relationship built through fraud investigations and proposed combining software with trained investigative professionals. The work extended beyond the moment a connection appeared on a screen.
For Korenblum’s story, the arrangement is a useful detail. The business he had established to extract meaning from information was selling into organisations that also needed people, training and procurement options. Voyint would train analysts on Voyager’s suite and act as a reseller. By then, Bromley was leading Voyager’s US business and speaking for it in the partnership announcement. The next stage of the company was already taking shape around a different executive, with customer support and operational delivery prominent in the pitch.
Handing over the chief executive’s chair
The leadership transition preceded Meta’s lawsuit. Courtney Bromley joined Voyager to lead its US business in 2022, following a 32-year career at IBM. In October, Voyager announced that she would oversee global operations as CEO while its headquarters moved from Israel to the United States.
Korenblum endorsed the change, pointing to the US federal, state and local markets as important areas of growth. His title in that announcement was founder and former CEO. The wording marks a real boundary in the story: founding Voyager and running it day to day were now separate roles.
The corporate record supplies another dated endpoint. Korenblum’s directorship of the UK entity, Voyager Labs Limited, began in July 2012 and ended on November 30, 2023. Later company announcements therefore need their own attribution. An August 2025 update described an integrated interface under new executive management; it was a development at the business he founded.
What the yellow paper set in motion
Korenblum’s Voyager years trace a particular entrepreneurial problem: how to turn experience in asking questions into software other people can use. Capital, researchers, commercial applications and government customers each supplied part of the answer. The privacy dispute exposed another requirement, one that technical capability alone could not settle.
The yellow sheet remains a useful image because it catches the project before all those complications acquired names, products and contracts. A founder brought an investor an idea about making sense of an unruly world. Building the company meant discovering how many decisions sat between the information and the answer.