Campaign HUB - An AI case study
Overview
With the AI Exploration Team, we were looking for ways to leverage AI to create real value for our customers. From user interviews and funnel data, we knew that many customers struggled with the transition from a functional website to one with real reach. This is exactly where we wanted to step in—using artificial intelligence to generate social media campaign suggestions based on the customer's business information.
My Role
Product designer
Ideation, Wireframing, Interaction design, UI design
The Process
I organized an ideation workshop with the entire team, consisting of two developers and a product manager, where we developed a more detailed idea, learning objectives, and a problem statement.
In close collaboration with the product manager, I created initial user flows for various use cases. After these were reviewed with the developers for feasibility, I did start working on the wireframes.
The Idea
The goal of the case study was to build a functional prototype that we could test with Jimdo users. In an AI chat conversation, relevant data about the user's business is collected. Based on this data, we use AI to generate a social media campaign consisting of post content and image descriptions.
The AI should also have access to contextual information about the users and performance data from ongoing campaigns, allowing it to reassess and adjust the campaign texts accordingly.
Wireframes and UI design
I created the first wireframes with Miro to make the collaboration and feedback loops with the team in a remote setup as easy as possible.
The UI design were made with Figma which allows the developer directly inspect the designs to create a first functional prototype.
Learnings & Takeaways
After completing a functional prototype, we invited existing customers to a user interview and prototype test to explore our qualitative learning objectives. The user feedback was positive and promising. Many users felt motivated and empowered by the AI-generated content suggestions to start their own social media campaigns. They were surprised at how well the campaign posts matched their business, even though all participants made their own edits and adjustments.
The chat-based interaction model worked well for collecting business information for the initial content creation. However, when it came to editing the content, users preferred traditional text formatting tools.
Even though we were eager to further develop the prototype into a product, we documented our work at this stage and handed it over to a product team, as we had achieved most of our learning objectives.
Some key learnings
- AI technology can genuinely help users overcome the barrier of creating their own content for social media. Users are more open to AI-generated text than expected.
- Aligning on learning objectives upfront is very useful, especially in the exploration environment we were working in. Sometimes it can be difficult to stop working on a project because you cross the line between exploration and product development. Defining clear learning objectives helps in that case.
- The chat was invented for conversation. If users know what to change, the classic and learned text manipulation interactions are more intuitive