Cognira makes the discount do its homework

A grocery discount is a tiny public announcement with a surprisingly large private life. Before the shopper sees the yellow ticket, somebody has to choose the product, predict demand, secure supplier support, coordinate the advertisement and decide whether the whole exercise earns its keep. Cognira has built a company around that backstage work. Its proposition is appealingly specific: put the people and the arithmetic in the same place.

The spreadsheet was already losing

Consider KeHE. In its September 2023 announcement with Cognira, the wholesale food distributor described promotion planning spread across legacy systems, cumbersome spreadsheets and fragmented email conversations. Its network then included 16 distribution centers, serving a business with roughly 30,000 retailer customers. A forecast had to travel through that organization. So did every revision.

KeHE selected PromoAI after evaluating several vendors. The stated attraction paired an AI forecasting engine with improved planning workflows. That combination matters. A clever prediction sitting in somebody’s attachment is still an attachment. Merchants, suppliers and supply-chain teams need an agreed plan they can use.

Cognira, founded in 2015, brought its AI Promotion Solution to public availability in January 2021. Founders Hatem Sellami and Bahadir Ustaoglu framed the product around integrated promotion decisions. Today, PromoAI joins calendars and campaigns to recommendations, performance analysis and vendor funding. Its expertise sits where retail merchandising meets data science.

From a signature to a working system

The customer announcements offer a useful progression. Schnucks selected PromoAI in December 2024 for a business then operating 114 stores. The scope covered fresh and center-store departments, weekly ads, emails, endcaps and online channels. Weis Markets followed in October 2025, announcing plans across 200 stores and digital channels. Both emphasized coordination across departments.

Brookshire Brothers supplies the more revealing sequence. The employee-owned grocer, with more than 100 locations, announced its selection in July 2025. On July 8, 2026, Cognira announced that implementation was complete and the platform was running across its promotional operations. The two announcements describe different accomplishments. Choosing software starts a project; getting teams to operate through it delivers a working system.

“They came in prepared, stayed engaged throughout implementation”Luke Anderson, Cognira VP of Delivery, on Brookshire Brothers

The delivered capabilities were concrete: centralized planning, financial forecasting, promotional optimization and post-event evaluation. Anderson’s comment credits the customer’s participation. That is a useful detail for anyone buying enterprise AI. The software arrives with models. The retailer must bring decisions, definitions and people prepared to work through them.

One price. Several kinds of intelligence.

Imagine a discounted product selling briskly. Has the promotion worked? That depends on what would have sold anyway, the margin surrendered, the supplier’s contribution and the effect on other products. Those questions explain why Cognira places vendor collaboration beside forecasting. A promotion’s economics include the funding agreement behind the ticket.

PromoAI’s vendor tools cover funding, negotiations and trade-spend analysis. Its optimization capabilities let planners compare scenarios before committing. Its effectiveness tools examine performance afterward. The appeal is a connected decision: plan an offer, test its likely consequences, account for funding and bring the results back into the next planning cycle.

Cognira also works with neighboring specialists. Its October 2024 Eversight by Instacart partnership connects promotion management with pricing and markdown capabilities. A Planalytics collaboration brings weather-driven demand analytics into the picture. A winter offer, after all, may meet a winter that behaves nothing like last year’s.

This places Cognira in enterprise retail planning, with a particular concentration on promotions. RELEX offers overlapping planning and optimization inside a broader suite. Cognira’s case rests on the depth and coordination of its promotion workflow, with integrations extending into adjacent decisions. Buyers should compare that fit against the systems their teams already use.

An assistant that knows what margin means

The newer development is Cora, Cognira’s AI agent embedded in PromoAI. The company describes an interface that helps planners query forecasts, understand recommendations, create planning material and identify conflicts using their actual promotion context. The interesting word is embedded. Cora works through the planning application and its business logic.

In an October 2026 essay, Ustaoglu argues that retail data requires client-specific meaning. Does margin include vendor funding? Which approvals permit a promotion to run? The answers live partly in configured applications and their rules. His argument favors AI working through those systems, preserving the logic that turns a number into a usable decision.

There is a cultural echo here. Cognira’s published values emphasize trust, teamwork and learning, alongside a blunt rule against brilliant jerks. For a company selling collaboration, that is a relevant aspiration. Its public office list spans five countries; its product asks equally varied retail functions to share one process.

Borrow the discipline

The buying process is enterprise sales, beginning with a requested demo and followed by implementation work. A retailer evaluating it should bring a real promotion: its forecast, funding terms, approvals and actual results. Ask the system to explain what changed between expectation and outcome. Then ask who can act on that explanation.

The same exercise is useful before any purchase. Agree on a baseline. Record the supplier contribution. Put the plan somewhere the relevant teams can see. Compare scenarios. Review the result. Reliable data and execution remain essential; Cognira itself qualifies outcomes by those conditions. A shared calendar cannot refill an empty shelf.

Cognira’s most transferable idea is this insistence on joining the calculation to the work. The shopper gets a price. Behind it, someone should be able to explain why that price was worth offering.