Albertsons is working to scale its retail operations in merchandise intelligence. The grocery giant is directly partnering with buyers, who are store operators, to integrate multiple platforms into a unified decision space to improve insights on merchandise, pricing, promotions, and displays. So said Karthik Iyer, Albertsons' group vice president of merchandising transformation and AI.

Iyer told sister publication CIO Dive that the merchandise intelligence platform is built on Databricks. Albertsons is rolling the platform out to buyers, with a goal of full deployment by the end of 2026.

"We are changing the way buyers have used technology in the past, guiding them into a new thinking space," Iyer said. He reports to Albertsons' chief technology and transformation officer, Anuj Dhanda.

Merchandise intelligence is one of four key investments Albertsons has planned this year, as the company seeks to scale AI across the enterprise and maintain an edge in the fierce competition for digital leadership in grocery.

Building an AI-first platform

Building a platform with clean data, governance, and AI is critical, Iyer said.

Databricks' Lakehouse product hosts Albertsons' retail data, Iyer said. The next layer of the platform is a governance layer via Databricks' Unity Catalog and AI Gateway. At the top of the tech stack is Databricks' AI agent, Genie, which allows buyers to easily query the platform in natural language.

Integrating large language models into the platform, trained on clean retail transaction data in the Lakehouse—such as price points for green versus red apples over the past three years—enables the platform to provide clear, forward-looking diagnostics, Iyer said.

Genie allows buyers to mine the value of data within the platform, Iyer said.

"If there's a summer drought and not as much sun as the previous year, what does that mean for Tillamook ice cream sales? Should I shrink the display space or expand it? Add more flavors? Or reduce space to give it to other products?" Iyer said. "That's exactly the type of insight we can get through this comprehensive platform with Databricks."

Buyer adoption of the platform will be key to its success, so Iyer said he worked side-by-side with the retail business to build the system rather than buying an off-the-shelf tool.

"If our buyers can say, 'Hey, it's given us insights we never thought of, given us trade economics we didn't have the bandwidth to think about before,' that's what excites me the most—transforming our business for meaningful outcomes," Iyer said.

Albertsons realized it needed better real-time visibility into its business, which is why Genie plays a key role in the merchandise intelligence platform, said Ken Wong, senior director of product management at Databricks.

"They wanted to provide a real-time system that could answer any type of question to meet those needs," Wong said. "With advances in LLMs, building this kind of system became possible, but making the results reliable and trustworthy is actually a considerable effort."