If you're an actuary, 2025 will likely be remembered as the year everything changed. Not in a dramatic, overnight way—but in a steady stream of 'wow, I can't believe I used to do that manually' moments that added up to something big.
Let me tell you about Sarah, a P&C actuary, and how her typical workday transformed this year.
What Changed This Year
GenAI Became Our Daily Companion
Remember when ChatGPT first came out and we all played with it but weren't sure how to use it at work? Well, 2025 was the year that changed for actuaries.
Sarah starts her Monday with a loss reserve analysis. She uploads her loss triangles and simply asks her AI assistant: "What's different from last quarter?" The AI spots that bodily injury claims for accident year 2023 are developing 8% higher than expected, drafts an explanation, and even suggests three possible reasons based on industry trends.
What used to take Sarah half a day now takes 20 minutes. But here's the important part—she still reviews everything carefully. The AI is great at drafting and spotting patterns, but it doesn't replace her judgment about why those patterns matter.
The key development: AI systems can now process thousands of pages at once (they call it "10 million token context windows"). This means Sarah can feed in years of reports and the AI can find connections she'd never have time to spot manually.
Traditional AI Got Really Good
Machine learning isn't new to insurance, but 2025 was when it became mainstream for actuaries. Sarah's company now uses ML models for pricing their commercial fleet business, and these models are analyzing telematics data from thousands of vehicles in ways that were impossible before.
The models find patterns like "drivers with this specific driving schedule have 40% fewer claims"—insights that translate directly into better pricing. And unlike a few years ago, Sarah can actually explain how the model reached these conclusions, which regulators really care about.
The Numbers That Tell the Story
Here's what happened across the industry in 2025:
- AI adoption jumped from 8% to 34% of insurance companies. What was experimental became operational.
- The Casualty Actuarial Society launched an AI Fast Track program because AI literacy became a core skill, not a nice-to-have.
- Companies using AI in underwriting reported 15-20% reductions in expense ratios.
- AI moved from back-office tasks to core actuarial work—pricing, reserving, and risk selection.
A Day in Sarah's Life (2025 Edition)
Morning: Sarah's working on homeowners claims after a coastal storm. She uploads claim photos, weather data, and loss information. The AI can now "see" the photos and connect visual damage patterns to claim costs. It identifies that specific roof damage types are settling 30% higher than normal, linked to recent building code changes.
Two years ago, this kind of analysis would have required days of manual review. Today, it's done before her second coffee.
Midday: Pricing a new product line. The cloud-based ML models run scenarios that used to take days, now completed in hours. Sarah experiments with different approaches, tests sensitivity to assumptions, and iterates quickly.
Afternoon: Sarah's testing something new—AI agents that can handle routine claim processing autonomously. When a new auto claim comes in, the AI validates coverage, checks for fraud indicators, estimates reserves, and routes the claim—all within minutes. Complex or ambiguous cases still come to human actuaries, but routine work flows through automatically.
What We're Learning (The Honest Part)
2025 hasn't been without challenges:
AI makes mistakes: Sarah caught her AI assistant making up statistics that sounded right but were completely fabricated. She learned to always verify, especially important claims.
Bias is real: Sarah's team discovered their AI was routing claims differently based on neighborhood data that correlated with race. They had to retrain the model with fairness constraints. This is serious business with regulatory implications.
The skills gap: Not every actuary is adapting at the same pace. Some colleagues feel overwhelmed by the changes. Sarah spends time mentoring others through the transition.
We still need judgment: The best AI in the world can't replace actuarial intuition about emerging risks, regulatory changes, or market dynamics. AI handles the "what" brilliantly—actuaries still own the "why" and "so what."
What This Means for Actuaries
Sarah's role hasn't disappeared—it's evolved. She's less of a number cruncher and more of a strategic advisor. AI handles the routine calculations, freeing her to focus on:
- Interpreting what the patterns mean for business strategy
- Validating that AI outputs make actuarial sense
- Handling complex cases that need human judgment
- Ensuring AI systems are fair and compliant
The boring parts of actuarial work? AI handles those now. The interesting parts—understanding risk, making strategic recommendations, dealing with genuine uncertainty—those are more important than ever.
Looking Back at 2025
As the year wraps up, here's what stands out:
This was the year AI went from experimental to essential in actuarial work. Not because it replaced actuaries, but because it made actuaries more effective.
The profession elevated: Actuaries who embraced AI tools found themselves doing more interesting work—exploring new products, analyzing risks at finer detail, and making bigger strategic impacts.
Challenges remain: Bias, transparency, regulatory uncertainty, and the need for careful governance aren't solved problems. They're ongoing responsibilities.
The opportunity is real: For actuaries willing to learn, AI isn't a threat—it's amplification. It's the difference between spending your day in spreadsheets versus spending your day solving meaningful problems.
The Bottom Line
2025 wasn't the year AI replaced actuaries. It was the year actuaries started working with AI in meaningful ways. The profession isn't shrinking—it's transforming into something more strategic, more impactful, and frankly, more interesting.
If you're an actuary reading this, the message is simple: embrace the learning curve. The AI tools that seemed intimidating at the start of 2025 are now essential to staying competitive. And the actuaries who figured that out early? They're the ones doing the most interesting work.
Here's to 2025—the year we stopped wondering if AI would change actuarial work and started figuring out how to make it work for us.
Enthusiast. Learner. Explorer.