Self serve returns v.1

Product designer | 3 months

Reducing customer contact by enabling a seamless, on-platform return initiation experience.

Background

Before launching self-serve returns, customers were unable to initiate returns independently on platform. This limitation placed a heavy burden on the customer experience team, who had to manually troubleshoot each return and provide personalised support. With a clear goal to reduce this pain point, the team set out to empower users to create and manage their returns autonomously.

The lack of a self serve superience on platform resulted in high contact contact rate and low NPS score for return related queries.

The goal

The goal was to implement self-service workflows that will allow users initiate, track and complete returns independently with limited customer experience contact.

What we delivered

We launched self serve returns using a dual-pathway approach that empowered customers while supporting our internal teams. Customer could initiate returns directly on platform and are routed to an enhanced AI chat workflow for efficient resolution.

0.1 Returns initiation from My account

We enabled customers to initiate returns from their account section choose return reasons and select a resolution method before continuing to an enhanced AI chat workflow.

Autonomous return initiation from the account section empowering customers to self serve

0.2 Optimised AI powered workflows

We redesigned the returns support experience by building structured workflows inside FIN that collected essential information upfront, such as whether customers had already tried troubleshooting steps or if their kit was showing specific error codes.

This meant that by the time a Customer Experience advisor picked up the return, they already had the full context needed to resolve issues much more quickly. We also integrated Intercom’s AI agent, FIN, by exposing MPB’s internal knowledge bases, allowing FIN to troubleshoot common issues independently and often resolve cases without needing to escalate to a human advisor.

Impact and scale

Launched a self-serve returns experience used by 1M+ customers across 13 markets, with support for 7+ localised languages.
Over 3,400 customers initiated returns via the self serve flow within the first 12 weeks with 80% progression rate.
61% of returns resolved without an agent's intervention.
Customer experience team workload reduced, freeing time to focus on complex and high value issues.
CSAT remained at 86% for all returns conversations, matching the benchmark and proving customers didn't feel abandoned by automation.
Time to resolve and close a support ticket reduced from 11+ hours to 8-9 hours on average, with peak efficiency at 2-3 hours during optimal weeks.
Average replies per ticket dropped to 2.4 replies (vs. traditional 5+ reply conversations), meaning the Customer Experience team received enough information upfront regarding the customer's return.
Returns volume and values remained within expected historical ranges (~8-9% of GMV, 248 returns weekly), indicating no operational disruption or spike in user returns.

Challenges

We had to deliver value sooner. To keep momentum and start improving the returns experience quickly, we split the work into Phase 1 and Phase 2. Phase 1 focused on delivering the most impactful self-serve capabilities early, while Phase 2 layered on more complex workflows and integrations.
We needed to preserve essential human touchpoints. For some return reasons, such as faulty items, we created a hybrid process where automated workflows collected detailed information upfront before escalating to a Customer Experience advisor for resolution.
We were working with legacy systems. The returns infrastructure was not initially built to support self-serve functionality, so we had to design a solution robust enough to integrate with old processes without breaking them, while still accommodating new features.