The John Lewis Partnership (JLP) is deliberately prioritising AI to address the “basics” and  improve business operations over customer-facing use cases because “that’s where the money is.”

Speaking at the Retail Technology Show, Barry Panayi, Chief Data & Insight Officer at JLP explained how AI is being used to address customer pain points by fine tuning back-end operations, such as supply chain organisation, routing and stock management.  Or, as Panayi describes, it: “The unglamorous, lucrative and safer bits.”

Building a nose-to-tail AI team

Having joined John Lewis in 2001, Panayi has focused on building a “nose-to-tail” team integrating data engineers, analysts and researchers to ensure truly triangulated data.  By combining transactional data with qualitative feedback, his team can address business needs by providing insights identifying measurable improvements. 

One such practical application is the creation of a Tableau dashboard, developed by the Waitrose Retail Performance team, and accessed more than one millions times.  The dashboard is used by partners to access information, using natural language, to answer customer queries, demonstrating the positive impact of accessible data tools. 

JLP’s momentum is back

Panayi joined the retailer, which stage chair and Customer Whisperer, Kate Hardcastle MBE, described as “one of the nation’s best loved” retail brands, when the company was experiencing significant losses.  He explained that, while the business had accessible data, culturally there hadn’t been a focus on creating a single, unified view of this data and insight.  When asked about recent turbulence across the business, which have included financial and profitability issues, staff reductions and the controversial cutting of its annual partner bonus, Panayi was upbeat”after four years, it feels exciting – momentum is back,” he said.

One AI watchout that Panayi was clear on is the need to avoid the common pitfall of data teams trying to dictate customer understanding to partners and business leaders.  He advises leaving the job of retailing to the experts, in this case JLP’s partners, with data and AI playing a supporting role and “letting the partnership do what it does well and enabling that – the tail isn’t wagging the dog.”  

Operational AI will come before customer-facing projects

Currently, JLP works with third-party AI solution providers, including Snowflake and Google’s Vertex AI.  Its foremost priority will be ‘nailing’ its machine-led operational optimisation as this will be the starting point before it moves onto more ambitious, customer-facing projects.  Personalisation is earmarked as the next focus area for AI application across the business. This could potentially include improvements such as improving website search functionality by implementing computer vision to correctly catagorise items.  Panayi also didn’t rule out the possibility of leveraging third-party expertise to improve loyalty programmes and personalisation initiatives. 

Panayi describes JLP as adopting a conservative approach to data security and privacy, exceeding minimum regulatory requirements, because JLP’s key differentiator is trust. 

“I don’t think it makes us slow, it makes us more deliberate,” continued Panayi.

Looking ahead, Panayi anticipates developments in natural language processing, enabling seamless interactions between customers, colleagues and data systems to improve both customer and employee experience.