A machine that turns spreadsheets into advertising
Here is a problem that sounds boring and is actually enormous: a mid-sized e-commerce brand has 10,000 products and a creative team that can make good ads for maybe fifty of them.
The other 9,950 products get whatever's left over, which is usually a stock photo on a white background and a caption someone wrote in a hurry. Genus AI's entire reason to exist is the suspicion that those 9,950 products also deserve ads that look like a person made them - and that a computer can now do it, at a volume no design team could survive.
The mechanics are almost aggressively practical. You connect a product catalog - a feed URL, a file, or a Shopify store. The platform reads it, and then it starts generating: branded creative templates applied across the whole catalog, product videos churned out by the hundred, background removal and seasonal overlays and logos dropped in dynamically, seed audiences modeled from live order data, and copy to go with all of it. Then it distributes the whole package across Meta, Google, TikTok, Pinterest, X, Reddit, Snapchat and Criteo, and reports back on which specific creative choices actually moved money.
Genus AI describes the result, on its own homepage, as "product ads that look great and perform." This is a refreshingly small promise for an AI company in 2026. It is not claiming to reinvent marketing or achieve consciousness. It is claiming your dynamic product ads will look designed instead of auto-generated, which - if you have ever seen an auto-generated dynamic product ad - is a meaningful thing to promise.
"AI is transforming how brands grow and engage with their customers. This technology should be accessible to all, and we are working hard to open it up."
- Genus AI, on its founding thesisThe word doing the heavy lifting there is "accessible." Large brands have always had creative automation. They just called it an agency and paid it seven figures a year. Genus AI's actual product is not the AI - lots of people have AI now - it is the price of the AI. The same machinery, aimed at a company with two dozen employees. That is a less glamorous mission than "reinventing creativity," and a more defensible business.
The volume is the whole story
These numbers matter mostly in relation to each other. A team of about two dozen people, generating forty million pieces of creative a month, is not a company that scales by hiring designers. It is a company where the marginal cost of one more ad is approaching zero - and that curve is the entire investment thesis.
From ten million to forty million
In 2023 the company reported generating ten million product images across the whole year. By 2025 it was quoting forty million a month. When a capability gets close to free, usage does not grow in a straight line - it detonates. The relevant question for any brand reading this is uncomfortable: are you still hand-making the thing your competitors now produce by the million?
Three hubs: make it, target it, measure it
Genus AI splits its platform into three parts, which is less of a UI decision than a statement of belief about what marketing actually is. Marketing has always been the same loop - create something, put it in front of the right people, find out if it worked. Naming the loop out loud is the sort of clarity that most software avoids.
Creative Hub
Catalog creative automation, dynamic personalization, and hundreds of product videos generated in minutes from your feed.
Audience Hub
AI audience modeling and lifecycle segmentation that maps customers to stages using live order data.
Reporting Hub
Creative performance insights that show which specific visual elements are driving conversions.
One underrated feature hiding in the Creative Hub: custom AI image tagging. You can train the system to recognize your brand's specific visual features - a signature silhouette, a house color, a recurring motif - so the automation applies your visual DNA instead of a generic template. The catalog, in other words, becomes the creative brief. Most brands treat the product feed as a database and the ad as art. Genus AI treats them as the same object.
Two brothers and a neuroscience degree
Genus AI was founded in 2017 by Tadas and Viktoras Jucikas. The detail that makes the company make sense: Tadas, the CEO, holds a PhD in computational neuroscience from the University of Cambridge. If you are going to build software about capturing human attention, having someone who studied human attention at the neuron level is not the worst place to start.
PhD, computational neuroscience, University of Cambridge. Works at the intersection of AI, product and e-commerce.
Leads the machine-learning organization and technology direction from the company's Vilnius engineering hub.
The company itself is a study in deliberate geographic splitting. The commercial operation sits in Nashville, Tennessee - relocated, notably, from San Francisco. The engineering brain stays in Vilnius, Lithuania, where the team of machine-learning specialists lives. This is not an accident of remote work. It is talent where the talent is, market where the market is. Increasingly, that is what a real technology company looks like: distributed, specific, and assembled from wherever the right piece happened to be.
Eleven million dollars, in two acts
In September 2023, Genus AI closed a $6 million seed extension, bringing its total seed funding to $11 million. The earlier money came from Picus Capital, Transamerica Ventures, Maschmeyer Group Ventures and HDI. The extension pulled in Aleph Group along with a notably operator-heavy cast of angels: Kazuki Ohta of Treasure Data, Magnus Lundin of Heep Agency, and Tomas Slimas, co-founder of Oberlo - the app Shopify bought and that arguably defined a generation of dropshipping. When your angel list is full of people who built the previous decade's e-commerce tooling, it tells you who thinks you understand the problem.
The interesting number here is not the funding - it is the $2 billion in e-commerce revenue the company says it processed to train its models, which drove more than $250 million in client revenue on Meta in 2024. Data is the moat, and unlike most moats, this one gets deeper every time a customer runs a campaign. The model does not just work; using it makes it better. That is the compounding loop that clever architecture alone can never buy.
Sage meets the agents
In August 2025, Genus AI connected its Sage assistant to OpenAI using the Model Context Protocol, letting marketers talk to their campaigns in plain language rather than clicking through dashboards. It is early, and it is a signal worth watching. The interface to software is quietly shifting from clicking to asking, and a marketing platform is a natural place for that to happen first - the person who wants "a video campaign for the summer collection targeting past buyers" would very much rather say it than build it.
- AUG 2025Sage AI assistant connected to OpenAI via MCP - a first step toward agent-driven growth marketing.
- APR 2025Published research on AI predicting ad performance, part of an ongoing "future of marketing" series.
- SEP 2023Closed $6M seed extension; total seed funding reaches $11M.
- 2017Genus AI founded by Tadas and Viktoras Jucikas.
What to actually make of it
A fair reader should note the usual caveats. The eye-catching numbers - forty million images a month, two billion in processed GMV, a quarter-billion in Meta revenue - come largely from the company itself, and self-reported metrics deserve the pinch of salt that all self-reported metrics deserve. Genus AI is also not alone: it competes with AdCreative.ai, Pencil, Smartly.io and, most awkwardly, with Meta's own increasingly capable Advantage+ tooling, which gives away for free some of what Genus AI charges for. The bet is that "free but generic" loses to "paid but on-brand," and that bet is not yet settled.
But the underlying observation is hard to argue with. There is a real, unglamorous problem - most products in most catalogs never get decent advertising - and Genus AI has built a specific, working answer to it. No grand narrative required. Just a company that noticed the boring 9,950 products and decided they were worth the trouble.