By Maddie Witters, Machine Learning Engineer at Fenris.

Fenris recently attended the Scout InsurTech Conference in Columbus, Ohio, a two-day event that brought together founders, carriers, and industry leaders from across the insurance ecosystem.

As a Machine Learning Engineer new to both Fenris and the insurance industry, attending this conference was a particularly valuable experience. I went in wanting to get a clearer picture of how data science and analytics are being leveraged across the ecosystem today, and to understand how carriers, startups, and agents are actually keeping up with the wave of new technology reshaping the industry.

As AI continues to accelerate, what new pressures does that create for the industry, and how are companies like Fenris positioned to help navigate them?

Here are some insights from the conference that stood out:

Climate Risk Has Moved from Buzzword to Business Imperative

The conference opened with a showcase spotlighting companies focused on climate, and the depth and breadth of startup activity in this space was striking. Climate-driven startups are building tools to model and predict climate-driven risk at a granular level: wildfire exposure, flood likelihood, drought frequency, and more.

The message from InnSure Executive Director Charlie Sidoti, who moderated the showcase, and later led the “Climate, Communities and the Total Cost of Risk” forum, was direct: communities are facing protection gaps that the industry hasn’t fully caught up to yet, and intentional resilience planning is no longer optional.

For carriers and MGAs, this means that geographic risk intelligence is quickly becoming a baseline expectation in underwriting.

Smart AI Adoption Requires Quality Data

Running through nearly every session was a consistent undercurrent: the quality of your data determines the quality of your models. The most complex models in the world will still perform poorly if trained on incomplete or inaccurate inputs. Now, more than ever, access to high-quality, relevant data matters.

The “Decision Latency to Decision Velocity in AI Insurance” forum hosted by NayaOne put a fine point on this. The session focused on how organizations can use AI to accelerate decision-making, but the throughline was that speed without data integrity is a false promise, and many organizations aren’t yet comfortable putting new AI tools directly into production.

This is why teams are looking to leverage synthetic data when experimenting with AI tools, creating a safe sandbox to evaluate results before rolling AI tools out in production. However, the catch, as the discussion made clear, is that synthetic data only works if it closely mirrors real data. If your synthetic dataset doesn’t reflect the distributions and edge cases of the real world, you’ve just moved the problem upstream.

What This Means in Practice

Both themes from the conference point to the same underlying need: access to complete and verified data that carriers can trust.

Our Property Hazards, Perils & Crime API delivers a comprehensive geographic risk profile for any property — pulling together wildfire risk scores (including fuel load, drought frequency, and vegetation burn data), flood zone assessments grounded in FEMA map data, wind and hail event history, and more. Rather than requiring underwriters to manually research or aggregate risk signals from disparate sources, Fenris returns all of this in a single API call at the point of quote.

Additionally, Fenris offers a propensity-to-bind model for auto insurance, designed to remove the data quality barrier entirely. With over 100 million outcomes already incorporated, tens of millions of predictions are delivered every month to users looking to improve early pipeline data utilization.

The idea is simple: bring your leads, and Fenris handles the rest. With nothing more than a name and address, the model generates a bind prediction powered by Fenris’s proprietary feature set: no data preparation, feature engineering, or existing model infrastructure required. The model is also designed to adapt over time, shifting from a generalized base toward a profile that reflects patterns specific to your business. It’s a practical answer to one of the conference’s central tensions: you don’t have to solve the data quality problem before you can start benefiting from AI.

The Scout InsurTech Conference was a useful reminder that the industry is moving quickly — and that the organizations best positioned to keep up are the ones investing in the right data foundation today.