The Future of AI Pricing: From Human-in-the-Loop to Fully Automated Decisions with Alex Halkin

In this episode of Pricing Heroes, we speak with Alex Halkin, founder and CEO of Competera, returning to the show two years after his first appearance in 2024. In that initial conversation, Alex predicted a major shift toward AI-driven, contextual pricing in retail. Today, he revisits those predictions, assesses the industry's rapid evolution, and shares what’s coming next—from autonomous pricing decisions to four-dimensional optimization models and algorithmic trading-style speed.

Our conversation explores how retail leaders are shifting from skeptical AI oversight to fully autonomous decision-making, the move toward four-dimensional optimization to balance competing KPIs, and the transition of retail pricing into high-speed algorithmic trading. Alex explains why pricing is never just a mathematical calculation; it’s a reflection of organizational data maturity, a culture of experimentation, and the fundamental reality of market demand.

The Shift in Trust: Moving Beyond Human-in-the-Loop

Two years ago, the industry was still debating the role of human oversight in AI pricing. Today, Alex notes a tremendous shift in trust. Retailers are no longer just looking for predictions; they are requesting fully autonomous systems that operate without a human "review and confirm" step. “The acceleration is so hard that retailers are asking to really not even have people in the loop," Alex explains. This trust is driven by a global tailwind of AI adoption that has turned even the most conservative CIOs into AI advocates.
 

Closing the Data Gap

While data quality was the primary bottleneck in 2024, the landscape in 2026 is far more mature. Retailers have stopped deleting historical data, recognizing that context is hidden in these datasets. Alex points to companies like Snowflake and Databricks as catalysts that allowed retailers to move legacy data to the cloud and "speak" with it more accessibly. The conversation has shifted from "can you upload a CSV file to an FTP bucket?" to "can you deploy directly on our database?"
 

Is Contextual AI Right for You?

Alex is candid that advanced contextual AI is not a universal solution. For example, luxury goods sold on airplanes might not have the transactional data density required for these models to work effectively. He suggests two primary markers for executives to determine if they need a tool like Competera:

• The Margin Impact: If a 0.5% margin optimization represents a sizable number for your business, the "beauty of small price tuning" becomes essential.
• Price Sensitivity: If your customers are highly price-sensitive or you struggle with your brand's price image, contextual AI is a fit.

The Rise of Four-Dimensional Optimization

Retailers often struggle to balance competing priorities like revenue, margin, and inventory efficiency. Competera’s answer to this is Model Routing—where a computer orchestrates different AI models under the hood to achieve specific goals, such as driving sales for one assortment while protecting margins for another. Looking ahead, Alex introduces the concept of four-dimensional optimization. This new type of algorithm reviews all possible future variations while balancing four specific parameters simultaneously:

• Revenue
• Margin
• Sales Items (Volume)
• Profit Margin

Pricing as Algorithmic Trading

One of the most provocative insights from the conversation is the comparison of modern retail pricing to the finance sector. Due to sanctions, shifting supplier channels, and fluctuating gas prices, retail pricing has become a game of speed. "Pricing in retail right now is close to betting on the public markets, like algo trading," Alex says. Competera is currently experimenting with temporal neural networks and models typically used for high-frequency trading to help price managers calculate the "cost of doing nothing" in a volatile market.

When Pricing Can't Fix Demand

Alex offers a dose of reality regarding seasonal inventory: some overstock problems are not pricing problems. If a product has no real demand, even the most advanced AI cannot find an "optimal" price to move it. “You have to be honest—you do have some items with no real demand... demand optimization just can't find the optimal price position because no one needs the item," Alex notes. In these cases, retailers must look upstream to manufacturing and production decisions rather than expecting a pricing algorithm to perform "magic".