<— Back

Connecting customers to support agents

At Google, I worked on the Help Experiences team and part of my time was spent designing the support escalation experience — or how customers could connect with human support agents.

When customers run into problems using Google products, most try to fix them on their own. Sometimes they're unable to find a resolution and may need to escalate the issue to an agent for more help.

As Escalation UX Lead from 2020-2025, I executed designs for 30+ launches ranging from smaller features to entire flow overhauls. This work ultimately resulted in higher customer satisfaction, faster resolution times, and $XXM in cost savings, among other successful outcomes.

Here, I briefly cover one of the many escalation projects I worked on but if you're interested to learn more, please reach out.

Key project: Bringing escalation into AI chat

In 2023, the Help Experiences team had launched an AI chat on Google Help Centers and escalation was a key capability we wanted to add to that experience. This would mean customers using the Help AI could connect to an agent while in the chat, as opposed to being kicked out of it to another page, just to have to start a new chat.

I worked on integrating escalation capabilities into Google Help's generative AI chat experience from Q2 2023 to Q4 2024 in collaboration with a product manager, UX researcher, and engineering team of five. During this time, we launched an MVP, three major version updates, and many smaller experiments in between.

This was a complex project for a few reasons:


  • Generative AI was still very new and rapidly evolving across the industry. The LLMs massively improved every few weeks and customer attitudes towards AI shifted just as much. This required constant iteration in design.


  • The escalation space has many technical, legal, and business operations requirements to determine agent routing and eligibility and all of those same rules applied here too.


  • Escalation majorly expanded the capabilities of the Help AI chat. I had to define and validate many new components and interaction patterns for this nascent modality.


  • Customers have a different mental model when interacting with an AI chat compared to a traditional form UI. Adapting the deterministic escalation experience to fit seamlessly into a generative AI conversation wasn't a straightforward exercise.


As UX lead, I…


Developed a deep understanding of the problem space:


  • Built and maintained an experience map as a source of truth of all inputs, edge cases, end states, and technical/legal requirements


  • Moderated UX research sessions to understand user expectations


  • Stayed updated on industry-wide conversational AI design patterns as they continuously evolved


  • Analyzed chat transcripts and experiment data to understand where designs were and weren't working


Executed effective, scalable designs:


  • Adapted escalation features from a traditional form interface to fit within the new AI chat modality, validating more complex components with UX research


  • Expanded the team's AI chat design system to accommodate all these new escalation components


  • Continuously iterated on flow and component designs as the underlying technology rapidly improved


Crafted clear, cohesive copy:


  • Worked closely with conversation designers and model prompts to ensure all escalation messages in the chat were consistent with the LLM's voice and structure


  • Planned and moderated user research sessions to test various copy treatments of sensitive messaging and entry point language


Collaborated often:


  • Frequently consulted with engineers to find user-centered solutions around technical constraints


  • Worked closely with product managers to define phasing, requirements, and priorities of features for launches


  • Was a point person for escalation-related efforts across multiple AI chat projects at Google


  • Presented designs and experiment updates to multiple levels of leadership for review and awareness


Outcomes:


We saw higher customer satisfaction, increased solution acceptance, and $XM in cost savings when we launched these escalation capabilities for consumer products leveraging the Help AI chat.

Details have been omitted and abstracted due to confidentiality requirements. Please reach out to learn more — thanks for understanding!

Want to learn more? Reach out at