Sprinklr is the third-party tool Shipt customer support agents use to handle issues from members, shoppers, and drivers across voice, chat, and email.

Leadership replaced the previous platform and rolled out in December. I joined in March toward April. After months of letting agents adapt, adoption and results still were not where they expected for a change this big. The vendor had promised cost savings and efficiency, but those only matter if agents use Sprinklr correctly so the platform features can run in production.

Success was real adoption, not forced logins.

Handle time drives agent metrics and pay, so agents optimized for speed. Admin stayed open in a second tab because that is where member context lived, not in Sprinklr.

I started with agents, not a mandate to force the tool.

Two other designers and I interviewed more than 10 agents and shadowed live sessions to see how they used Sprinklr. We asked why they were unhappy, why they skipped features, and why they relied on Admin when the data could live inside the platform.

Full Admin inside Sprinklr would have taken multiple quarters of engineering for a tool that serves many teams, not just support. With my PM and engineering partners, we met leadership halfway: surface the smallest, highest-value slices of Admin first.

After shadow sessions we presented findings to directors and VPs. The adoption matrix landed hardest: where agents are today, where we need them to be, immediate pain we could fix, and later capabilities to close the gap. That replaced force adoption with a phased integration plan.

Sprinklr already identified the caller. We used a broker to link that identity to Admin and embed the right profile in the active case view. When multiple accounts matched, the widget listed them so agents could verify with the caller without a separate search flow.

This was the first roadmap item and the proof that Shipt could integrate inside Sprinklr's constraints. We could not use Shipt's design system, so we led with information architecture: what to surface, in what order, and how much load each screen could carry, with the next integration in mind.

The discovery work didn't stop at the widget. Wrap-up codes led to an idea I brought to engineering: transcript and chat data already existed, and behavioral pattern detection could pre-select the most likely wrap-up codes so agents confirmed rather than filled from scratch. That became Shipt's first experimentation with AI on the X-Team, before the company had a formal AI feature strategy.

Canned responses, wrap-up optimization, and several other findings became officially roadmapped features for the X-Team product. A year and a half later, that team still references the original research when building new features. The widget also opened the door to chatbot and IVR parity: I designed the chatbot experience and the IVR team followed our interaction model for consistency between text and voice. Items 02 through 05 on the roadmap trace back to the same research cycle; the widget proved integrations could work inside Sprinklr, and everything else built on that credibility.