Heka Global

A fraud analyst that never sleeps, reading the entire open web

Somewhere right now, a bank is deciding whether the person on the other side of an application is real. A pension fund is hunting for a member who moved three addresses ago. An insurer is squinting at a thin file, trying to guess. Heka Global sits quietly behind all three, reading the public web and handing back an answer with the receipts attached.

Heka is an external customer intelligence engine for financial institutions. It takes the scattered, contradictory mess of the open internet - behavioral signals, online activity, relationships, risk indicators - and turns it into a structured, explainable profile of a single human being. Banks, insurers, payment processors and pension providers use it to make faster, audit-ready decisions. The unglamorous part, the part that matters, is that every result ships with a source link. You can see where the signal came from.

The financial system is running on records from another century

Credit bureaus were a clever invention - for 1899. They tell you whether someone paid back a loan. They are quiet on whether the person exists, whether the identity was assembled last Tuesday by a fraud ring, or whether the customer simply moved and forgot to mail a forwarding card. The world generates more verifiable signal about a person in a single afternoon online than a bureau collects in a year. Most institutions ignore all of it, mostly because it is messy.

And messy it is. The open web is the largest data source ever built and the least cooperative. It is full of duplicates, aliases, dead links and noise. Pulling a clean, defensible answer out of it is the kind of task that sounds simple and ruins quarters. That gap - oceans of signal, almost no structure - is the problem Heka exists to close. Everything below is, in one way or another, a consequence of it.

An intelligence officer, a trader and a lawyer walk into a fintech

The team is the kind of mix that sounds like the setup to a joke. Ishay Horowitz, a former senior officer in the Israeli intelligence community, knows how to find a person inside a haystack of open-source data. Rafael Berber, former Global Head of Equity Trading at Merrill Lynch, knows what the buyers - large financial institutions - actually need before they will trust a number. Idan Bar-Dov, a fintech and high-tech lawyer turned CEO, knows the regulatory tightrope all of this has to walk.

Their bet, made in 2021, was that the techniques intelligence agencies use to assemble a picture of a person from public traces could be made compliant, explainable and fast enough for a bank's underwriting queue. Not a black box. An analyst-grade engine that shows its work. It is a less thrilling promise than "AI will fix everything," which is precisely why institutions started signing.

One engine, pointed at four expensive problems

Heka's core is a single web intelligence engine that extracts and structures verifiable data points from the public web into a live digital identity. The clever part is restraint: it cuts through the noise rather than drowning you in it, and it tags every conclusion with traceable evidence. Around that core sit four packaged solutions.

Numbers, names and the boring certifications that close deals

Pitches are cheap; production is not. In one deployment with a global payment processor, Heka caught 65% of account-takeover losses while leaving legitimate activity alone - the rare fraud win that does not punish good customers. The customer roster reads like a who's-who of European and Swiss finance, and the company holds ISO 27001 and SOC 2, the unsexy badges that let a bank's procurement team say yes.

Then there are the partners who put their own names next to Heka's. Dalriada Trustees works with the company to trace pension fraud victims and lost members. Winmark's PensionChair Network brought Heka in as a technical partner. The client wall:

Make the open web legible - and accountable

Heka's stated mission is to transform scattered web signals into structured, explainable consumer profiles. Read it twice and the word that matters is explainable. Plenty of companies promise to point AI at the internet and produce a verdict. Far fewer promise to show you the trail behind that verdict, which is the only version a regulated bank can actually use. Heka is betting that traceability, not raw prediction, is the moat.

The longer goal is bolder: to become the default source of truth for digital reality in financial services. Replace the assumption with the observation. Replace the stale record with the live one. It is the sort of mission that is either slightly arrogant or exactly right, and the next few years will sort out which.

The fraud got smarter; the records did not

Synthetic identities, AI-assembled personas and account takeover are not slowing down - they are getting cheaper to manufacture. Static records were already losing that race. The institutions that win the next decade will be the ones that can read live behavior and act on it without locking out their real customers. That is a precise description of what Heka sells, which is either lucky timing or the whole point.

Go back to the three scenes from the top. The bank still has to decide if the applicant is real - but now it has a source-linked profile instead of a hunch. The pension fund still has to find its lost member - but the trail is fresh, not five years cold. The insurer still has a thin file - but it is no longer guessing. Same questions, different evidence. Heka did not change the questions financial institutions ask. It changed what they are allowed to know before they answer.