The useful detail in Thomas McKenna's career is not that he arrived early to artificial intelligence. It is that he arrived late to founding. Before Ocula Technologies, there were years of learning how retailers make decisions, where their data gets stuck and why an elegant analysis can die somewhere between a meeting and a product page. The company he now runs is a consequence of that apprenticeship.
McKenna, usually called Tom, began in the machinery of retail personalization. At dunnhumby he spent seven years working across Europe and Asia, in a business known for turning baskets, loyalty cards and customer behavior into commercial decisions. His stated specialties would later read like a map of modern ecommerce: customer analytics, personalization, price optimization and loyalty. These are neat labels for untidy work. A customer changes channels. A price moves. A promotion interrupts the pattern. The data is abundant; the decision is still hard.
He moved to Deloitte, where he led BenchMarque globally and worked on analytics transformation projects for major retailers. Then came Bain & Company, where he led data, analytics and digital transformation engagements for consumer businesses across EMEA. The altitude changed. The recurring problem did not. Large organizations could collect more information than ever, but turning it into a prioritized action remained expensive and slow.
The last mile of analytics
Consulting rewards diagnosis. A software company has to survive the treatment. That distinction explains the shape of Ocula. McKenna and co-founder Dr Gregory Fletcher, a data scientist and technologist, did not set out to build a general-purpose AI oracle. They focused on the stubborn last mile between an insight and a changed ecommerce experience.
Consider a catalogue with thousands of products. Every page needs accurate copy, a consistent brand voice, useful details and enough structure to be understood by search engines. The same item may appear on a retailer's own site, Amazon, eBay and Google Shopping, each with different conventions. A small team cannot lovingly rewrite every description. A generic text generator can produce volume, but volume without context creates another quality-control job.
Ocula's answer has evolved through products for ecommerce analysis, pricing and content. Its software identifies what deserves attention, then helps execute the work. The emphasis on priority matters. A backlog of 10,000 possible improvements is not a plan. A ranked set of interventions, connected to traffic and revenue, begins to look like one.
Users consistently tell us that our generative AI is levelling the playing field with their most advanced competitors.Thomas McKenna, on Ocula's 2024 Series A
The phrase “levelling the playing field” carries McKenna's central argument. The largest retailers have long been able to hire teams of data scientists, engineers and content specialists. Smaller groups face the same complexity with fewer hands. Software can compress the gap, but only if it arrives in the workflow as a usable action. Another impressive demonstration is not enough.
A company between two cities
Ocula was formed in 2021 and built between London and Belfast. That second city was a deliberate operating choice. McKenna said the company selected Northern Ireland and “hasn't looked back,” pointing to local talent and support from Invest Northern Ireland. In 2023, Ocula announced an £11 million research and development investment there, with plans to grow its local technical staff from 10 to more than 50 over three years. For a young software business selling into the UK and North America, Belfast became more than a lower-cost engineering outpost. It was part of the company's technical identity.
There is a pattern in this decision. McKenna's career has crossed Europe, Asia and North America, but his operating language remains specific: people, product and measurable client work. Ocula's publicly stated values emphasize transparency, fairness and drive. When the company raised its Series A, McKenna described what it had built as special from both “a product and culture perspective.” The pairing is revealing. An AI product is not only a model. It is the group that chooses what the model should do, how its output is checked and which customer problem earns another month of attention.
The funding put more weight behind that choice. In July 2024, Ocula announced a £3.25 million Series A led by Praetura Ventures, with Castelnau Group and Lloyds Banking Group adding to their existing investments. Ocula said it would expand sales and technology teams, build products and accelerate in the UK and US. At the time it reported more than 25 major retail customers, including Boots, AO and Hornby.
Those names make the proposition concrete. Hornby manages distinct brands and thousands of SKUs across multiple markets and languages. AO sells products whose details matter to search discovery and purchase confidence. The Kansas City Chiefs, another reported customer, show that the underlying problem stretches beyond conventional retail. Every organization with a large digital catalogue faces some version of the same grind: content decays, prices move, competitors change and yesterday's optimized page quietly becomes today's missed opportunity.
The product page gets promoted
McKenna's current focus is arriving just as the product page is being promoted from supporting actor to storefront. A shopper may land directly from Google. Another may begin with a full question inside an AI assistant. A third may never visit the brand's homepage at all. The page has to attract, explain and convert without the old sequence of browsing a shop window, walking through categories and eventually reaching an item.
In his public talks, McKenna has argued that conversational search changes the language retailers need to publish. Shoppers increasingly express an occasion, constraint and preference in one sentence. Product information must contain enough trustworthy context to answer them. At the same time, content has to fit each channel rather than being copied indiscriminately from one marketplace to another.
This creates a more demanding job for ecommerce software. It must write, but it must also remember. It must produce enough variation for different channels while preserving the product facts and the brand's boundaries. Ocula has increasingly described this as agentic work: systems that can plan and carry out multi-step content tasks, rather than wait for a person to prompt a blank box each time.
The fashionable version of that story is autonomy. McKenna's version is more operational. Retail teams still need control, review and a reason to trust what reaches the customer. The value is in removing repetitive production while keeping judgment near the result. At Savant London in 2025, he appeared with Hornby's ecommerce leader Michael Manton, whose team used Ocula's copywriting tools across large volumes of products, brands, markets and languages. The benefit discussed was not synthetic prose for its own sake. It was recovered bandwidth.
A founder's earned advantage
McKenna's biography offers a useful counterweight to the cult of the sudden breakthrough. His founder story is cumulative. Personalization at dunnhumby taught him the customer-data problem. Deloitte added transformation and product-building at organizational scale. Bain put him in C-suite conversations about strategy and performance. Ocula required all of it at once, plus the unglamorous work of hiring, financing, selling and deciding what not to build.
That accumulated context is difficult to imitate. A general model can learn the vocabulary of ecommerce quickly. It cannot automatically know which recommendation a time-poor merchandising team will act on, which claim a brand will refuse to publish or where a retailer's systems create friction. These details live in the workflow. McKenna spent his career inside it.
We have created something special at Ocula from both a product and culture perspective.Thomas McKenna
The next test is expansion. Ocula has spoken about growing in the United States, serving more retailers and moving toward agentic tools that can manage larger parts of the content cycle. McKenna has also taken the company's argument onto larger stages, from NRF in New York to Shoptalk Europe, where he was listed to discuss AI and the new marketing playbook in June 2026. Visibility creates opportunity, and expectation.
The company still has to prove that automation can produce durable commercial gains across many clients, not only striking pilots. It must keep output accurate as catalogues and models change. It must help teams move faster without making every product page sound the same. Those are product questions, but they are also the kind of operating questions McKenna has spent years circling.
His career can be read as a steady shortening of the distance to the result. At first, he studied shoppers. Then he built analytics capabilities. Then he advised leaders. Now he owns the roadmap and the consequences. The title changed from consultant to founder, but the subject stayed remarkably consistent: how to turn a retailer's messy information into a better decision.
That consistency may be the most transferable part of the story. New technology rewards speed, but useful companies also require patience with an old problem. McKenna did not discover retail when generative AI became popular. He waited inside retail long enough to see what the technology could finally make practical.