The first thing Idomoo tried to personalize was not a perfume ad, a sneaker drop or a synthetic actor delivering your name with eerie precision. It was the bill. A statement. The sort of document designed to make a customer squint, sigh and call support. The company’s founders saw an unusually practical opening: turn confusing account data into a short film that explains itself. If the amount, plan and next step could change for every recipient, video might do more than decorate a campaign. It could carry the message.
That was 2007, years before “AI video” became a category with a crowded demo reel. Brothers Yaron and Danny Kalish and engineering leader Assaf Fogel founded Idomoo around a harder, less photogenic problem. Ordinary rendering technology was too slow and rigid for video assembled from live customer data. Their first product decision was therefore an infrastructure decision: rebuild the rendering engine so a business could produce many versions quickly, not edit one master forever.
That framing still explains the company better than the AI label. Today an enterprise can connect CRM fields, a CSV or an API; decide which person sees which scenes, text, audio and images; generate the variations; distribute them over email, apps, SMS, ads or the web; and track what happened. A viewer can click, choose a product, supply information or branch the story. Live feeds can update inventory, location or other context after the campaign has launched.
Five million tiny mirrors
The cleanest example came from Ubisoft. Players received personalized recaps built from their own behavior. Idomoo says it generated five million of them and measured engagement 2.7 times higher than the comparison. Call of Duty used the format for a campaign that reached number four among Twitter’s trending hashtags. Allianz used a personalized policy overview and reported a 25% reduction in cancellations. Orangetheory Fitness turned members’ workouts into year-in-review films and reported 97% class attendance during launch week.
These are selected company case studies, not a promise that putting someone’s first name in an opening frame will perform miracles. The stronger pattern is that the data changes the substance. A player sees the games they played. A policyholder sees the coverage they bought. A member sees the classes they attended. The personal detail earns its place because it helps explain, remember or act.
The customers are accordingly large and operationally fussy: banks, insurers, telecoms, airlines, game publishers, fitness chains and service organizations. Public examples include Vodafone, Google, American Airlines, Aegon, PenFed Credit Union, Virgin Atlantic, Oracle, Activision and Ubisoft. Their problems are not merely “we need a video.” They need thirty markets, approved claims, accurate account fields, governed access, a measurable call to action and an answer when legal asks where the data went.
Then the prompt box arrived
Personalization worked, but the company has acknowledged what failed next: the project could demand too much heavy lifting. Specialists still had to prepare creative, data logic and production workflows. The technology belonged mostly to brands large enough to support the effort, and the viewer still watched a largely one-way message. Idomoo added interactivity, user customization and live data. The next change was more consequential. Generative AI offered a way to make the production process itself programmable.
Lucas, introduced as an AI video agent, accepts a prompt, webpage, document or image. It can write the script, assemble scenes, pick footage, make animation, add voiceover and music, and apply a company’s rules. This is where Idomoo wants to separate itself from avatar tools such as Synthesia and HeyGen, free-form clip generators such as Runway, and editors built around stock templates. Lucas is pitched as the operator of an entire enterprise workflow, not simply the face on screen or the model that produces a few seconds of footage.
Strata refuses to flatten the pancake
Most generative video systems produce pixels: a flat clip. If the words are wrong, the timing drags or the logo drifts, the user often regenerates and hopes for a luckier roll. Strata, Idomoo’s foundation model released in 2026, generates a composition. Text, footage, animation, actors and synchronized audio remain independent layers. Lucas learns what the company calls Brand DNA - design, narrative and approved assets - then uses Strata to build a blueprint for that specific piece of content.
“Every other AI video model generates pixels: a flat file you can’t meaningfully edit.”- Danny Kalish, co-founder and CTO
This is a genuinely different technical target. It also reveals the business model. Idomoo sells enterprise software on custom terms, with platform access, rendering volume and implementation or creative help where needed. There is no tidy public price card. The broader company was financed like an infrastructure bet: a $9 million Series A in 2013, a reported $3 million investment in 2018 and an $18 million Series B in 2019, followed by a Tel Aviv Stock Exchange listing in 2021. At least $27 million of the venture funding is consistently documented; broader databases count the intervening round differently.
What did that money buy? International reach, enterprise sales and nearly two decades of rendering, distribution, security and data plumbing. The company says its patented engine can render at up to 100 times real time. That claim matters more than an especially pretty demo when the assignment is five million individual films before the campaign goes stale.
It also bought Idomoo a peculiar kind of patience. The company entered the public market before the present AI-video rush, when its proposition was still described mostly as personalized-video SaaS. Since then, the competitive map has split in several directions. Avatar companies make a presenter inexpensive. Diffusion models make fresh footage on demand. Design suites make templates accessible. Personalized-video specialists automate variations. Idomoo is attempting to join those jobs without surrendering the operational controls that attracted regulated and global customers in the first place. That is a wider product, but also a more demanding sale: buyers need enough video volume, useful data and repeated communication moments to benefit from the machinery.
The latest product work shows where the bet is moving. In June 2026, Strata became a selectable style inside the enterprise creation flow, and Idomoo added a feature that turns a screen recording into a polished video. Those are modest-sounding releases with strategic intent. One brings the new model into the everyday interface; the other captures a common piece of raw workplace material and pushes it through the same production system. The goal is not a spectacular clip once. It is routine video output without a routine trip to the production queue.
The useful part to steal
A smaller company should not copy Idomoo by beginning with a foundation model. It should copy the shape of the problem. Pick a moment where a customer already has a reason to pay attention: onboarding, renewal, a complicated bill, a milestone or a recommendation based on actual use. Choose only the data that makes the explanation more useful. Give the viewer one obvious action. Then run a control group and measure a business outcome, not applause for the creative.
A four-part field test
Make one personal video before you make a million
- Start with a high-friction customer moment, not a vague brand campaign.
- Use trusted data to change the story itself, not merely the salutation.
- Put a visible next step inside or immediately beside the video.
- Compare conversion, retention or support volume against the old message.
The conditions matter. Personalized video is a poor fit when the underlying data is incomplete, consent is unclear, the audience is too small to justify integration, or the message is simple enough for three lines of text. A one-off artistic film may be better served by a director or a free-form generative model. A quick training presenter may be easier in an avatar tool. And no amount of rendering speed rescues a campaign whose offer is irrelevant.
Idomoo’s place in the market is the awkward middle that enterprises often need: more controlled than a generative toy, more flexible than a template library and more programmable than a production studio. Its history is also a tidy lesson in changed timing. The company built the plumbing when personalized video sounded niche. Now every software vendor wants an agent, every brand wants more video, and the least glamorous work - permissions, layers, live data, rendering and measurement - is suddenly the point.
The result is not a machine that makes one perfect movie. It is a system that makes the right version, keeps the pieces editable and knows whether anybody clicked. For enterprise video, that may be the more useful trick.