It's a great example of copilot really being just alongside the person, the human, all the way through the process...helping them be more effective and get their job done better and faster.
For marketers to improve productivity and campaign ROI with an AI-first experience
Led the design vision and strategy for the project, shaping the end-to-end user experience and defining the core mental model that improved clarity and usability. Established scalable design systems and frameworks, enabling cross-team alignment and ensuring consistency across the product. Set the foundation for other designers to follow, driving a high level of execution and quality.
Lack of a clear mental model in the earlier version made it hard for users to know what needed to be done and when they were finished.
...it looks really cool and everything is just a more seamless experience of building a campaign vs. going into each of the different entities and doing it one by one. I think this is a really exciting tool that will be useful to the team.
One of the core ideas was ensuring that the human was in the loop, making final decisions and shaping the campaign. The AI could quickly give suggestions. This would build user trust with the tool.
Our initial focus was on creating an intuitive flow, which users clearly appreciated. The additional feedback highlighted key opportunities for further enhancement, confirming the direction we need to pursue to refine and expand the experience.
Navigation is intuitive, and users find the steps clear, except for the last step where they struggle to review emails and publish the journey.
Users preferred to upload a brief instead of providing a prompt as it’s more structured.
Users want clarity on the information source of the generated segments, as it is currently unclear to them.
Expected ability to preview emails across different platforms like mobile and desktop to ensure consistency.
Expected to have A/B testing capabilities in the tool before sending the email journey.
Desired a collaborative space for approvals and collaboration.
It was important that the transition from curating to editing was clearly signaled to the user. The previous iteration didn't make that transition clear and it was part of our hypothesis that it was a large reason why it was unsuccessful. I believed that showing the navigation from the curation phase, then hiding it so the user could have a full canvas to work with, could help. I iterated on different ways this transition could happen.
I like this one a lot more...it is very clear on what I should do next and very digestible!
What demos well isn't what customers necessarily want
The previous one shot experience sacrificed users sense of control. We had to build customer trust with the AI. Giving users a sense of control could help with establishing trust with AI.
Developing Future-Ready Solutions Requires Discipline
It was tempting to fall back on the default text prompt as a quick solution, but we resisted. Sticking to our core principles was challenging at times, yet it allowed us to build a foundation that goes beyond just another text-based AI interface. Because we held firm, we now have a solution we can refine and iterate on—one that pushes the experience forward rather than defaulting to the familiar.
Context is everything even in AI
An empty input waiting for a prompt can be anxiety inducing. I was doing a lot of work with AI and Copilot, leveraging different ways for users to start engaging with Copilot outside of an input. But many of the journeys started from a blank prompt with little guidance other than some instructional text. What was clear was we needed more contextual controls to engage with AI other than inputs. With this project more context led to better results. And in a way this was easy because its an enterprise product with a very narrow focus. With that in mind how can we ensure we are getting enough context to better help our users?
Share so everyone can grow
I led the team to share our work with the larger design team of 200, hoping others could build on it and push their own ideas further. The goal wasn’t just to showcase what we did, but to create something useful—maybe even something I could learn from and use down the line.
