CRM system with Ai/ML integrations

UX project (internship)

This project was made during my internship at nvrmind, me and another intern together designed and UX improved a lovable prototype. This project explored how a modern CRM system can be designed to support a more efficient and structured sales process. The focus was on creating a solution that not only manages customer data but also helps users prioritize, make decisions, and work more strategically throughout the entire sales flow. A central part of the work was integrating AI/ML-based features that could provide insights and support in various parts of the process.

The work began by researching user needs and identifying challenges in existing CRM solutions, which formed the basis for both the information architecture and functional requirements. Wireframes were then developed to explore different flows and interaction models, before these were further developed into an interactive prototype in Figma.

Through user interviews and usability tests, the design was continuously evaluated and iterated. The goal was to create a CRM system that feels intuitive, reduces complexity in everyday tasks, and simultaneously provides better support for decision-making throughout the sales process.

Tools and techniques

Brainstorming, crazy 8, pain points, wireframing, Figma, interviews & usability test.

Duration

16th March - 10th april 2026

Ideation

We started of by mapping out the system and all key parts of it. Since neither of us had any prior experience with a CRM system alot of research was done here to really understand their layout and structure.

Pain points & key aspects

Here we wrote down the key aspects and pain points of the system.

Crazy 8

When we had a better understanding of the structure, we started ideating on different ways the data and system could be shown. This was done using the Crazy 8 method.

Decided on the frames of the app

After completing the Crazy 8s exercise, we had a clearer idea of the structure and key features. We moved to the whiteboard to map out initial frames and explore different directions.

We focused on two frames: the dashboard and detailed view (shown when one of the categories on the dashboard is clicked).

Translating to Wireframes

Once the flows were defined, we translated the whiteboard sketches into wireframes (using Frame0). These sketches where then sent to the client, awaiting feedback.

Iterating on the layout

After meeting with the client, we redid the wireframes to be even more intuitive.

The feedback we got was for it to be even simpler and show less data.

To achieve this another workshop was held using the crazy 8 method and the whiteboard.

Wireframes

The redesigned layout was then translated to new wireframes. These were sent to the client and we got a green light on the layout and structure of the system.

User testing

Parallell with the second iteration of wireframing, we started interviewing users of CRM systems and usability testing the prototype. During this stage we had our wireframes and the client's lovable prototype.

Three interviews and tests were held, during the span of a week. Each participant had a backround as CFO or Business Growth analysist.

"I think the data should be visualized in both graphs and key figures - because a preference here is mostly based on the viewer"

"I feel like the data presented isn't really intuitive - like for instance what does a 63% risk in pipeline mean?"

"I don't have an issue with AI in these types of systems really, but thorugh my own experience - all AI in CRM's is really dumb."

"I really like the structure and how minimal it is - it doesn't feel like an information overload."

Valuable insights

The data from each interview was then gathered and some key points were noted.

The data should we visualized in different types of graphs - since this is preference based on the user.
Clearer and more decpritive text was needed for the different key figures.
Established terms should be used.

Prototyping

Then a 2-day sprint was held, where everything we had gathered was translated into a Figma prototype. This also included the brand guidlines such as typeface and colour palette.

Wireframes

The redesigned layout was then translated to new wireframes. These were sent to the client and we got a green light on the layout and structure of the system.