Featuring insights from Robert Bates, Data Science Specialist
AI, machine learning, and large language models (LLMs) are having a profound effect on the retail industry, touching everything from customer support to the supply chain. Amongst these innovations, one area is consistently underestimated: pricing.
In our latest episode of the Retail Pricing Insights Podcast, 7Learnings CEO Felix Hoffmann sat down with Robert Bates, former Head of Decision Sciences at Currys, pricing leader at Waitrose and John Lewis, and a seasoned expert with a PhD in physics. They explored how retailers can drive real business impact through pricing, and why it’s the best place to start with AI.
Why pricing should be your first AI use case
Most retailers are still operating on manual, rule-based pricing systems. That means they’re missing out on huge margin and profit potential. As Robert described:
“If your pricing strategy is right, you can drive a lot of value. If you’re slightly off, it’s easy to fall behind - either becoming uncompetitive or investing too heavily.”
Pricing affects every customer touchpoint, which is why it consistently outperforms other AI applications in retail. Even Gartner ranks it as the highest-impact AI use case, well above more hyped use cases like chatbots.
Why decision science matters more than just data science
Robert led a function called Decision Sciences, focused not just on building models, but on how to make better decisions.
“It’s not about building the perfect model. It’s about knowing where to modify a system to make it work better.”
While working at Currys, his team applied AI and machine learning to everything from pricing and promotions to sales forecasting and customer lifetime value. The philosophy was always the same: focus on practical, high-leverage decisions.
Humans + AI: The blended model that actually works
Retailers often ask: should we automate pricing entirely? Robert’s answer is no. The best approach is augmented decision-making, where AI provides recommendations, and merchandisers make the final call. Ultimately, the machines are acting on behalf of category teams to hit targets. If you don’t have their buy-in, the best tool in the world won’t help.
To build that trust, the AI must explain its logic clearly; showing why a price recommendation was made, how it fits into the market, and how it aligns with the category’s goals (sales, margin, or stock optimization).
How to measure pricing performance
Unlike A/B testing in marketing, pricing optimization is tricky to validate—especially offline. Robert and his team used digital twin models to simulate what would’ve happened with a different price strategy. You can’t always do a clean A/B test, but you can build confidence with simulation. Ask: What would the alternate scenario have delivered?
Both Robert and Felix agree: when you run controlled tests, like product-based splits online, you can see real, measurable gains. 7Learnings has seen up to 50% profit improvement for low-margin retailers switching from rules-based to predictive pricing.
Should you build or buy a pricing solution?
It depends. If your pricing process is highly custom or strategic, in-house might be the way to go. However, for most retailers, outsourcing parts of the stack (like competitive price tracking) makes sense.
“You've got to be clear; what is it that is unique to your business which is hard to replicate as well?”
The biggest mistake? Ignoring the human factors. Even the best technology fails without the right governance and team alignment.
Learn more in our Buyer’s Guide to Pricing Software
What’s next for retail pricing?
Thanks to cloud compute and evolving AI models, we’re now able to solve problems that were out of reach just a few years ago.
You've got a lot more AI based tools coming in, identifying where there are opportunities within the business, what's performing well, what isn't. On the one hand, that's really exciting. There's so many options. The slightly scary part is working out how do you get all of those interacting properly with the business?
He recommends starting with pricing because it’s scalable, controllable, and delivers fast results. We couldn’t agree more.
Key takeaways
- Pricing is the #1 AI use case for retail impact, beating out marketing, personalization, and chatbots.
- AI should augment human decision-making, not replace it.
- Success depends on clear logic, governance, and buy-in from merchandising teams.
- Digital twin simulations and controlled experiments help validate pricing strategies.
- Retailers must balance build vs. buy based on their capabilities and strategic needs.
Listen to the full episode:
“Building AI Pricing models that Retailers trust and use” featuring Robert Bates
