Context

To understand this project, you have to understand 2 things: Conversations and AiroHQ.

Conversations is a GoDaddy product offering that serves as a one-stop inbox for all customer communications for SMBs. It aims to replace all of the individual inboxes a business owner might have (email, web chat, social media, etc.) and connect them into one place to better manage customer messages. 

AiroHQ is GoDaddy's AI-powered hub. A user can land in AiroHQ and seamlessly summon any of the backend agents powering it to complete a variety of tasks. This spans from an agent who will validate your business idea based on market demand, to an agent that will help you create a marketing plan once your business is up and running.

​

We launched the first version of a Conversations agent in AiroHQ during a time-limited surge. It introduced:

• Message summarization

• Basic prioritization

• Opportunities to train chatbots to help respond for the user

This was the simplest version we could ship to get a Conversations agent into AiroHQ. From the beginning, we always knew the opportunity was much larger than this MVP approach.

The problem

Small business owners don’t just have an inbox problem. They have a business operations problem disguised as an inbox.

They’re:

• Managing messages across channels

• Repeating the same responses

• Missing opportunities (and sometimes revenue)

• Spending time reacting instead of growing

Instead of a Conversations agent that simply acts as an accessory that can slightly speed up inbox operations, we wanted to think of a future where a Conversations agent acts as a true business partner, anticipating problems before they occur, giving best industry practices, and taking the burden of sifting through the inbox away from the user. ​

However, product roadmaps had already moved on from AiroHQ and there was no immediate intention of looking at it again. So, as the Conversations lead designer, I pushed forward on a design-led exploration, completely separate from product priorities.

The goals

We defined success in terms of outcomes, not features.

For customers, our goals were to allow them to:

• Save time

• Convert more leads

• Look professional without increasing overhead

For the business:

• Increase retention

• Drive conversions

• Build trust in AI-powered workflows

And our design goals:

• Efficient

  • Streamline processes to minimize time and effort spent by users

  • Design is responsive and optimized for various devices and platforms

  • Minimize the number of steps required to complete a task

  • Utilizes resources effectively, balancing performance and aesthetics

  • Intuitive and easy to understand and navigate

• Effective

  • Address the core needs and goals of the target users

  • Adaptable to changing user needs and evolving business requirements

  • Users able to complete tasks with minimal errors or confusion

  • Supports accessibility and inclusion for all users

• Enjoyable

  • Design is aesthetically pleasing and visually engaging

  • Foster a sense of trust, credibility, and connection

  • Uses language and tone that resonates with users

Now that we had the goals in mind, I needed to figure out how users wanted AI to help in their inbox. I ran a user survey with SMBs to:

1. Gauge overall interest in using an AI assistant for managing customer messages.

2. Identify hesitations or concerns SMB owners have about using AI in this context.

3. Discover the most desired use cases—what specific inbox tasks SMBs would want help with.

4. Understand boundaries—what tasks they would or wouldn’t want AI involved in.

5. Inform post-MVP scoping based on what features are most useful and where user control is expected.

From this research, we mapped out what an ideal system could do:

• Smarter prioritization

• Auto-categorization and tagging

• AI-generated responses in the user’s tone

• Bulk replies for repeated questions

• Message summarization

• Task generation and follow-ups

• Trend detection across conversations