AiFi uses Azure to bring real-time intelligence to the physical world

Mapping technology changed how we navigate the world. It made roads, traffic, and distance visible in real time — creating entirely new industries built on that understanding. Spatial intelligence company AiFi is setting out to do the same thing for physical spaces: grocery stores, stadiums, airports, clothing retailers. The behavioral data is already there. It just hasn’t been visible.

“Our mission is to bring that visibility,” says Joao Diogo Falcao, technical founder and CEO of AiFi.

AiFi’s platform digitizes physical spaces in real time, using computer vision and AI to turn the movement of people, objects, and activity into structured, queryable data. At the center of that platform is eddie — short for Event Driven Data Interpretation Engine — which runs on Azure and is powered in part by Microsoft Copilot.

A startup bet on the right partner

When AiFi set out to build its spatial intelligence platform, the company needed more than compute infrastructure. It needed credibility in the retail industry and the ability to get in front of major grocery chains and global retailers and be taken seriously. That partner was Microsoft.

"Microsoft's presence in the retail domain is very, very good," Falcao says. "The relationship that Microsoft has with all these retailers is stronger than any other cloud provider. That was a big indicator for us going into this industry."

That credibility mattered immediately and translated directly into deals. Data privacy and security are the first questions any client asks when trusting another company with their data. In AiFi's early days, on-premise deployments triggered exhaustive security questionnaires with hundreds of questions just to satisfy physical safety requirements. The Microsoft name collapsed those conversations into:  "Oh, you use Microsoft — okay." 

 The Azure GDPR compliance‑ready platform addressed the requirements for AiFi’s European clients, reducing a critical barrier to expansion by establishing trusted legal assurances, robust technical safeguards, and operational transparency enabling AiFi to confidently scale AI deployments across Europe.

In the platform's early days, the vision outpaced what was technically possible. Falcao wanted a single model capable of reasoning across an entire physical environment at once, not just isolated tasks. As large AI models matured and AiFi accumulated years of video data from hundreds of deployments, that vision became achievable — and so did the case for Azure. 

"As we go deeper into it — training eddie, training larger models, as AI became more and more prominent in our solutions — that just cemented the decision that Azure was the right player," Falcao says.

Today, eddie runs on the Azure high-performance NC-series GPU infrastructure, processing an enormous volume of video from hundreds of camera-equipped locations. Azure Databricks powers the data exploration and analytics layer that clients use to query their spaces. And Microsoft Copilot serves as the reasoning engine on top of eddie. When a retail manager asks a question about their store, it’s Copilot doing the reasoning, and eddie supplying the physical-world context that makes the answer possible. Running across hundreds of locations at that scale demands infrastructure AiFi can count on: predictable performance, enterprise-grade governance, and the confidence that what works in one deployment will work in a hundred more. 

From one store to a hundred

The results across AiFi's more than 370 deployments and 40 clients tell a clear story. In grocery, the most immediate application is frictionless checkout: shoppers pick up what they want and walk out, with eddie assembling a complete transaction record automatically. Accuracy is non-negotiable in high-volume grocery — thin margins mean even small error rates are unacceptable. Eddie changed that math. Clients who once piloted a single location are now rolling out 10, 20, 100. The platform delivers 25–40% reductions in measurable shrink, with some stores recovering significant revenue annually.

Loss prevention and inventory management tell a similar story — and what sets eddie apart is not just the data it surfaces, but how concretely it surfaces it. When one grocery client was told they were losing 1.7% to theft, the numbers alone weren't enough to prompt action. It was only when eddie surfaced video of exactly what was happening — a shopper scanning lower-value items at self-checkout to cover more expensive ones — that the picture became undeniable. On the inventory side, eddie monitors shelf availability in real time, flagging gaps and out-of-place products so staff can act before a missed restock becomes a missed sale.

Beyond grocery, the platform is transforming how operators understand their spaces. An apparel retailer running a Black Friday test used eddie to surface not just that sales were up, but exactly why — which displays drove engagement, which sections drew more foot traffic, which promotions converted. 

“In the past, finding that kind of insight would have taken data scientists weeks,” Falcao says. “Here we could query: Show me insights on this particular object or on this particular date.” AiFi analyzes more than 60 million human interactions every month.

At the Intuit Dome, home of the LA Clippers, AiFi powers frictionless concessions across the entire venue. Transaction times drop to as little as 10 seconds; throughput increases of up to 300% and labor cost reductions of up to 60% are possible for operators.

"Once it's there, everything just feels magical," Falcao says. "It feels like you're going to your pantry and getting food — the experience is so personalized."

AiFi is now expanding beyond retail — into airports and, Falcao envisions, virtually any large indoor space where understanding how people and objects move creates value. The map, it turns out, is just getting bigger. 

Discover more about AiFi on LinkedIn and YouTube.