Why Property Claims Teams Feel Friction with CCMs — and How AI Is Changing Everything

Enterprise Customer Communications Management (CCM) platforms are designed to consolidate communication systems and create a single, governed source for all customer-facing content.

For marketing or billing departments, this approach can make sense, sure.

But what about claims teams? Well, there’s about $100 billion in room for improvement through AI in claims handling, and claim correspondence is part of that.

CCM platforms are often expensive, with a steep learning curve that requires significant training. Their rigid, template-based architectures struggle with complexity. Every possible permutation may require a separate template. Handling state regulations, policy endorsements, and complex amendments often pushes these systems beyond their capabilities, forcing adjusters into time-consuming manual work.

Beware the Copy and Paste Monster – It’s Creating Indefensible Policyholder Communications

Some desk adjusters will dutifully commit to the hours-long process of using CCM templates. This requires looking up policy information, exclusions, and endorsements to correctly cite relevant policies in denials. Depending on the system, it also means laboriously altering formatting.

You know what does not take two hours? Copying and pasting that very similar letter from two weeks ago.

It’s an open secret. And in a dispute, those errors regularly make claim settlement letters indefensible. Consider a few common scenarios:

• A similar but different policy gets cited. Both would have resulted in a denial, but the inconsistency makes an immediate settlement the lower-cost option.
• Citing a common endorsement not held by a particular policyholder, a claim gets paid out when it should not have.
• Many letters with errors after a hurricane result in accusations of systemic issues, frustrated customers, and risks of increased regulation.

Why Are Claims Teams Comfortable with an AI Solution?

In the case of new AI solutions like Voltaire, the answer is simple: it’s faster than a CCM and more accurate than copying and pasting.

That’s why as much as 80% of adjusters at carriers using Voltaire use our AI voluntarily. That’s a high adoption rate for any software with adjusters, let alone AI.

But… isn’t the issue with LLMs that they’re… not always accurate? What would be the point, if adjusters just need to double check policy language?

As several carriers have attempted and found out, giving some documents and a prompt to an LLM will not produce accurate claim letters.

It requires many AI agents, working in tandem with other automations and software, to consistently outperform legacy methods in speed and accuracy. As one experienced adjuster put it “this saves me around two hours a day.”

How it works:

Templates remain, but far fewer are needed. Voltaire uses generative AI to create an entire, contextually accurate claims letter. Base policy language, endorsements, and exclusions appear accurately cited in a well-formatted Word document. Adjusters still adjudicate and must enter denial reasons, in claims requiring it. In fact, they have hours more every day to do so.

From Adjuster Efficiency to Financial Performance

Traditional CCM platforms build their business case around softer metrics like improved brand consistency or customer experience.

In contrast, AI in claims management is projected to reduce loss adjusting expenses by 20-25% and claims leakage by 30-50%. Even with a modest 5% reduction in claims leakage, along with savings in other areas, Voltaire’s business case shows 200%+ ROI in the first year.

Early Carrier Adopters Will Benefit the Most

What would your carrier do with 30% more margin for their mission, without IT lift or months- or years- long implementations? AI continues to produce a lot of hype, and it cannot always deliver on its promises.

However, Voltaire can easily prove its speed and accuracy in a brief demonstration. In our experience, the reputation of carriers and adjusters as AI-avoidant falls apart quickly when getting their hands on AI that just works.