AI-research | Romollo

Designing Decathlon's AI-Powered Post-Purchase Experience

My Role

As Global Director of Product Design for Lower Funnel & Payments, I led the end-to-end strategy and design direction for the initiative. My responsibilities included:

Defining the AI vision and product strategy for the Lower Funnel organization.
Prioritizing post-purchase as the initial focus area based on customer research, business impact, and organizational readiness.
Leading discovery by analysing customer feedback, CRC contact data, and benchmarking 22 leading retailers and logistics companies.
Facilitating cross-functional co-creation workshops with Product, Engineering, Research, Content Design, Data, and Customer Relationship Centre stakeholders.
Guiding the team from research to concept development, resulting in a portfolio of AI-powered customer experiences.
Establishing the strategic principle that "AI should amplify clarity, not replace it," and defining the long-term roadmap from assistive AI toward agentic commerce.
Aligning stakeholders around a research-led AI strategy and ensuring proposed concepts were evaluated against business objectives before entering delivery.

Executive Summery

Artificial Intelligence had rapidly become one of the company's highest strategic priorities. Across Decathlon there was significant excitement about AI, but also growing concern.Everyone wanted to build AI.

Very few teams could clearly articulate which customer problems AI should solve. Within the Lower Funnel organisation, responsible for Purchase, Payments, Delivery and Post Purchase,we recognised a significant risk: We could easily spend months building technically impressive AI features that customers neither wanted nor trusted.

Internally this was often referred to as avoiding "AI slop"building AI simply because it was fashionable rather than because it solved meaningful problems.Rather than starting with technology, I proposed that we begin with customers.

Objectives

The programme was designed around four strategic objectives:

Improve the payment experience while supporting migration to the new payment platform.
Design a reusable, product-agnostic payment component that could serve multiple travel verticals.
Build a foundation for future payment capabilities, including wallets, vouchers, instalments, credits and flexible payment timing.
Balance customer needs, technical constraints and business priorities through an evidence-based design process.

The Challenge

Through Medallia insights, customer feedback and Customer Relationship Centre (CRC) data, one pattern became impossible to ignore.Customers weren't frustrated because Decathlon lacked support.They were frustrated because information was fragmented.

The same questions repeatedly generated customer contacts:

Where is my order?
Why hasn't my refund arrived?
How do I assemble my bike?
Which spare part do I need?
How should I maintain my equipment?
When should I replace my running shoes?

Every answer existed somewhere.None existed where customers needed them.Meanwhile, the business faced significant operational pressure:

Rising CRC operational costs
Multiple disconnected support platforms
Fragmented ownership across domains
Limited engineering capacity
Increasing pressure to demonstrate AI value quickly

The opportunity wasn't simply to build another chatbot. It was to rethink the entire post-purchase experience.

Objective

I established a simple principle for the team:

AI should amplify clarity,not replace it.

Our goal was not to automate customer support. Our goal was to anticipate customer needs before customers even had to ask.

Ultimately, we wanted Decathlon to evolve from a retailer that reacted to problems into a proactive sports companion.

My Approach

Step 1 — Prioritising the Right Problem

The Lower Funnel organisation had ambitious business objectives but extremely limited capacity.

Rather than attempting AI across every touchpoint, I deliberately narrowed our scope.

We focused exclusively on the post-purchase journey because it demonstrated:

highest customer pain
highest business impact
greatest organisational readiness
quickest opportunity to validate AI value

This decision protected the team from becoming another feature factory while increasing the likelihood of meaningful outcomes.

Step 2 — Research Before Building

Instead of ideating immediately, I led a comprehensive discovery phase.

We analysed:

Medallia customer feedback
Customer Relationship Centre contact logs
delivery issues
returns behaviour
refund questions
ownership pain points

Alongside internal research, I commissioned a competitive benchmark across 22 leading retailers and logistics companies, including Amazon, Nike, IKEA, Zara, FedEx and Walmart.

Rather than comparing AI features, we asked:

What customer problems are they solving?
Which experiences genuinely reduce support demand?
Which AI implementations actually improve trust?
Where are competitors failing?

The research identified several common success patterns:

proactive delivery communication
accurate ETA predictions
intelligent self-service
AI-assisted returns
AI copilots supporting customer service agents

It also highlighted where AI damaged customer trust:

opaque automation
hallucinated information
hidden policies
chatbots with no human escape path

These findings directly informed our design principles and concept development.

Step 3 — Cross-functional Co-Creation

Rather than handing research to design, I facilitated collaborative workshops bringing together:

Product Managers
Designers
Researchers
Content Designers
Data specialists
CRC representatives

We used:

Journey Mapping
Crazy 8's
Storyboarding
Opportunity Mapping
Concept prioritisation

This transformed customer insights into tangible AI concepts that could be validated before engineering investment.

Designing the Future Experience

The workshops produced a portfolio of AI concepts centred on proactive customer support rather than reactive customer service.

Product Assistant

An AI companion that understands a customer’s purchase history to provide personalised product guidance, setup support, training and nutrition recommendations, and complementary equipment suggestions.