Chatbot improvements

Chatbot improvements

Chatbot improvements

Delivered 4 prioritized use cases and UI improvements that will be used to expand the chatbot functionality.

Delivered 4 prioritized use cases and UI improvements that will be used to expand the chatbot functionality.

PROJECT LENGTH

5 months (ongoing)

5 months (ongoing)

TEAM

2 SR. UXD, 1 SWE

2 SR. UXD, 1 SWE

COMPANY

Google : AdMob

Google : AdMob

Context

Project background

  • In the first launch of the chatbot, the goal was to support natural language queries for AdMob reporting.

  • Since launch, the team (1 Eng and 2 designers) is looking to further expand the use cases and workflows supported by the chatbot. 

  • At the AdMob Strategy Summit 2025, I designed and presented a North Star Vision detailing how AI can be seamlessly integrated across the platform to provide intelligent and contextually-aware recommendations.

Problem

  1. The team does not understand which use cases and workflows publishers would find beneficial in using a chatbot.

  2. The team is unsure how to build upon the launch of the chatbot.

Goal

  1. The primary goal is to understand and prioritize which use cases we should tackle in the next iteration of the chatbot.

  2. The secondary goal is to expand the design library for chatbot components.

Project milestones

01

USE CASE ANALYSIS

Using various data sources, identify use cases for the chatbot specifically for AdMob.

02

RESEARCH VALIDATION

Stack ranking will help us understand which important workflows can AI affect.

03

IMPROVE CHATBOT UI/UX

Work with Eng to improve the UI of the chatbot by designing richer response formats.

01 - Use case analysis

Leveraging existing data sources

To understand which publisher needs can be helped by a chatbot, I began looking at which data sources publishers currently use to help solve their issues. I landed on the following:

Scoring data sources to highlight areas of importance for publishers

I then established a prioritization framework by quantifying the frequency of specific topics/CUJs across all support channels (Help Center, Developer Docs, and Listnr feedback).

Finally, I assigned weighted scores (Low, Mid, High), which allowed me to identify and prioritize the top use cases.

02 - Research validation

Create the research plan for survey

The team needs to validate these use cases so I led the research discussions and wrote the initial survey research plan for 300+ publishers, aimed at strategically aligning their current workflow priorities with the potential application and demand for chatbot assistance.

03 - Improve chatbot UI/UX

Map out use cases to expected prompts and ideal responses with richer components

My next step is to map likely prompts based on priority use cases so that I can understand which likely outputs and components the team needs to build and support.

First milestone complete

Short-term UI updates

While the larger team validates the key use cases and key stakeholders align on the focus of the next iteration of the chatbot, the Eng and UX team can work in parallel on UI updates to the chatbot.

San francisco based UX designer
Let's connect!
© 2026 - Virg Leynes Designs
San francisco based UX designer
Let's connect!
© 2026 - Virg Leynes Designs
San francisco based UX designer
Let's connect!
© 2026 - Virg Leynes Designs