Buildium AI | Ramsey Gunter

Role: Content Designer

Duration: 2 weeks

Team: Product Designer

Project overview

Brief

Buildium is a cloud-based property management software company based in Boston, Massachusetts that recently implemented an AI-first roadmap and needs scalable AI design guidance for its international UX team. As the Content Designer on this project, I developed tone of voice and prompt design guidance as part of the broader AI framework while collaborating closely with a Product Designer to leverage our experience on the leading AI scrum team.

Project

Use user-centered design thinking to develop instructional and informative AI guidance to not only executives, but the entire UX department.

Problem

There was a shift toward AI-first development and the UX executives had no user-centered AI guidance to help grow the business.

Solution

Create tone of voice and prompt design guidance within broader AI guidance to promote consistent and scalable AI content that centers users while creating a foundation that supports business goals.

Learning opportunity

While I have years of experience growing existing content guidance, this would be my first time putting together guidance from scratch. Therefore, the Product Designer and I invited lots of stakeholder feedback and prepared ourselves for many iterations to ensure the guidance effectively supported its target audience.

Project overview

Understanding the reader

Starting point

Although the Product Designer and I were crafting this guidance from scratch, we had a strong foundation to build on. We had both been on the leading AI scrum team and had gained hands on experience designing AI features and collaborating with engineers across various projects. During these projects, we had also collected feedback and questions through our presentations to the Design Review Board and Product Review Board. These insights helped us define our starting place and identify key themes to address in the guidance.

Interviews

We also held interviews with UX executives to understand what they were hoping to learn from this guidance and what they felt would best support the team.

Pain points

Through our experience and interviews with executives, we identified the following pain points designers would have when designing AI features for the first time.

Constraints

Simplifying and connecting

Analogy

Writing for designers with a foundational understanding, I opened the content guidance with a conversational tone and an analogy to frame prompt design in a familiar, approachable way.

Soft coding vs. hard coding

Then, I explained the difference between AI-generated content written with prompts versus human-written content, and provided examples to illustrate the distinction.

Establishing rules

Bridging the gap

While some designers had computer science backgrounds, others had never touched backend code before. So, I wanted to frame the best practices in a way that helped bridge that gap and emphasize the need for precise instructions when working with large language models (LLMs).

I also provided a set of tone of voice base prompts to use across all AI features, helping to keep a consistent and cohesive voice throughout the product.

Engineering collaboration

I also worked with the engineers on my scrum team to develop a set of guidelines and questions to facilitate efficient communication and collaboration between UX and engineers when prompt designing as a team.

Embedding AI

Integration

It is important to strike a balance between highlighting new AI features and not interrupting the existing and familiar user experience. So, I explained how to use point of view to emphasize or blend AI features into the existing interface, depending on the objective. I worked closely with the Product Designer to ensure we also had guidance on how this coincides with design patterns and components

Marketing collaboration

Additionally, I collaborated with Product Marketing to create in-app, brand-specific guidance and crafted guidance for how designers should work with Product Marketing to ensure a unified voice and cohesive brand identity around AI.

Thinking globally

Nuance

I encouraged the designers to think about nuance too, since there would be infinite edge cases and cultural variations in the designs.

Reviewing and iterating

Comprehensive guidance

The Product Designer and I combined our AI guidance and strategically organized it into 10 sections to help readers seamlessly navigate and absorb the information.

Presentations

We presented this guidance to the Design Review Board and expanded on certain areas within the guidance where feedback suggested confusion or need for additional guidance. After this iteration, the Product Designer and I did a 3 hour presentation over the course of 2 days to the broader UX team and executives. We noted their questions and feedback to iterate on the guidance.

Next steps

I set up 4 hours of weekly AI Content Office Hours to monitor for recurring questions or themes that might need to be added to the guidance.
I joined the Design Review Board as a board member to provide prompt design guidance and monitor for gaps in AI knowledge in the team.
We plan to send out a survey to figure out where people are struggling , and we will continue to send out surveys as needed.
We will continue to grow the AI Guidance from additional feedback.

Conclusion

Results

This guidance informed the development of over 110+ new features within the first month of its release and laid the groundwork for future AI projects impacting over 17,000 users.

Takeaways

AI might be changing parts of the content design process, but at the core, centering users and collaborating efficiently within a team are essential human skills that will continue to co-exist alongside AI. I am glad this content could serve as an introduction to AI design for so many, and I look forward to continuing to support my team as we adapt to an AI-first model.