Reimagining Atidot's AI Platform for Insurance Agents

I led both the design and product direction for this redesign. As the sole designer I owned the end-to-end experience, and I also took on the product work: defining the problem, prioritizing the roadmap against engineering capacity, aligning sales and leadership on what to build, and deciding what to cut.

Atidot had built genuinely powerful predictive technology. My job was to shape it into a product experience that matched.

Atidot is an AI-driven analytics platform for the insurance industry, helping carriers, brokers, and agents identify opportunities around policy retention, cross-sell, upsell, and customer engagement. The predictive models were genuinely powerful. Like many deep-tech products, the challenge was translating that technical strength into an experience users could act on.

The platform surfaced valuable predictions, but workflows were fragmented, and many features showed what was possible without building a clear path to doing it. The core problem was simple to say and expensive to ignore: users had to connect the dots themselves.

Users struggled to answer practical questions: Who should I call today? Why is this customer being flagged? What should I say to them? At the same time, insurance organizations were deeply attached to their existing CRMs and weren't looking to replace them. They wanted tools that fit into their current processes with minimal friction.

Rather than start from the interface, I started from the business. I sat in on prospect meetings, watched sales demos, ran client workshops, led onboarding calls, and presented the platform directly to brokers and agents. Two constraints shaped the strategy: engineering capacity was limited, so a full rebuild was off the table, and clients would not adopt anything that asked them to leave their existing systems. So instead of a redesign wish list, I set one prioritization principle and ran every decision through it.

None of the agents we spoke to wanted another CRM. These teams had years of processes, habits, and probably a few emotional-support spreadsheets built around the tools they already had.

Several flows were redundant or unfinished, adding confusion without driving action. With engineering capacity fixed, I made the call to cut before adding. I audited every flow against usage data and feedback, then aligned engineering on what to remove and what to protect, so our limited runway went only where it moved users toward action.

Out came the flows that created cognitive load or broke momentum, and the navigation was streamlined so users could focus on acting instead of hunting for features.

Sitting in sales calls, I kept hearing the same objection: clients did not trust "black box" predictions. I treated trust as a product problem, not a modeling one, and prioritized three changes across the platform.

The same gap surfaced in nearly every customer conversation: users trusted the insight but froze on the next step.

I prioritized an outreach composer that generates personalized outreach (email, SMS, or call script) directly from the platform's segmentation, drawing on the exact signals that flagged each policyholder. I scoped it deliberately: generate from data we already had, so we delivered real value without an integration lift engineering could not afford.

This wasn't about AI writing being trendy. It was the most direct way to make the product usable day to day.

One of the most requested capabilities was a way to act on a segment without a full CRM integration, which no client wanted to take on. I designed a campaign page where users view a segment of flagged policyholders and download or copy the list straight into their own marketing systems.

No integration, no new tool to learn, and users could go from insight to a running campaign in minutes.

I championed an AI chat assistant for the dashboard, built a prototype using Claude Code, and pushed for it internally. Then I tested the idea in client interviews, and the demand was not there. We were about to build the wrong feature.

My interviews caught it before we shipped, and I redirected our limited engineering and design capacity to the work that actually moved users from insight to action.

This was the first time I owned a product end to end, not just the design but the product decisions behind it. I wasn't handed a spec. I decided what to build, what to cut, and how the platform should evolve, and then I designed it. Owning the product and the design at once was a rare opportunity, and it changed how I understand my own role.

Honestly? The thing I'm proudest of isn't something I shipped. It's that I actually learned the business. Sitting in on sales calls, demos, and workshops taught me things no amount of staring at the interface ever could. Turns out eavesdropping on sales is a legit design method. Who knew.