DeepSeek reportedly allows insurers to maintain full data sovereignty, meaning proprietary data remains private and protected. (Free use image)

The barrier to entry for artificial intelligence is lowering.

While the technology has been the focus of all industries, cost has been a major factor in slow adoption. Most advanced AI models require massive cloud-based infrastructure, making them the domain of well-funded enterprises.

But with the emergence of DeepSeek, powerful AI no longer requires prohibitively expensive infrastructure. Organizations can deploy smaller, more efficient versions of large-scale language models directly within their data centers.

Many of the principles used in developing DeepSeek, a Chinese artificial intelligence startup, have changed the way developers think about AI learning. Most western AI labs have primarily relied on massive scale to achieve better performance, keeping the barriers of entry for competition high.

In contrast, the DeepSeek team invested in efficiency and obtained significant results in a cost-effective manner. Some analysts estimate DeepSeek’s models are 20 to 40 times cheaper to run than comparable models from OpenAI, enabling high-end AI capabilities on standard servers.

Addressing the elephant in the room

But the value of DeepSeek for insurers is not using cloud-based versions of the technology, but deploying it on-premise. Using the technology in the company’s private data center or cloud ensures that sensitive data never leaves the organization’s control. It is more secure than cloud-based APIs, which transmit confidential information externally. Insurers maintain full data sovereignty when they keep the AI model and all interactions within the company’s firewall, meaning proprietary data remains private and protected.

Additionally, DeepSeek’s permissive licensing means that insurers have everything necessary to deploy and tailor the AI in-house. This means that insurers can own and maintain the entire AI supply chain, which has compliance and governance benefits.

Using DeepSeek to improve insurance operations

Insurers can deploy DeepSeek to improve processes throughout the entire insurance life-cycle. Several areas are ripe for significant transformation:

  • Risk assessment: DeepSeek can enhance underwriting and reduce manual workloads by reviewing submission documents, summarizing details such as property information, highlighting key risk factors, suggesting preliminary assessment decisions or flagging anomalies based on underwriting guidelines.
  • Claims payments: The technology can extract essential details from claims forms and documents, determine coverage, and draft initial claim resolutions for simple cases. It can also cross-check claims against known fraud patterns and flag suspicious elements for investigation.
  • Customer experience: Insurers can use the AI to power customer service agents so they can understand and answer complex customer questions enabling insurers to provide round the clock customer service. The technology can also tailor recommendations and suggest coverage additions or adjustments based on the customer’s profile and exposure.
  • Compliance standards: DeepSeek can act as a virtual compliance assistant for internal operations by continuously monitoring regulatory updates and industry bulletins and summarizing any new requirements or changes in plain language for the legal/compliance team. It can scan company documents to ensure they meet required regulations and flag any deviations. DeepSeek can also automate routine tasks like drafting standard email responses, filling in forms by extracting data from attachments or compiling reports.

Best practices for adopting DeepSeek

When it comes to AI, it’s important not to overhaul the entire insurance operation at once, but identify one or two areas first. Insurers should ensure they have the right infrastructure and talent in place to fine-tune and maintain the model. A pilot program can help to test the technology, especially if the company focuses on one business line or one region, monitors the results, and makes adjustments.

Organizations can gather staff feedback on the technology’s performance. It’s important to have a governance framework for how AI will be used in decision making, including defining the areas that require human oversight. Compliance teams should be involved early to review outputs and make sure they meet regulatory standards. After a successful pilot and an established governance framework, insurers can then scale up usage.

DeepSeek levels the AI playing field lowering the cost of entry and enabling more companies to implement advanced large-scale language-learning models internally within their systems. Insurers can now implement AI solutions on-premise giving them complete control over their data and ensuring security and privacy. Deploying DeepSeek, insurance organizations can elevate underwriting, claims, customer service, and compliance while reducing costs, improving accuracy, and maintaining control over sensitive data growth.

Brandon Nuttall is chief digital and AI officer at Xceedance, a global provider of insurance-focused consulting, technology, operations and data solutions. He has roughly two decades of experience in the insurance industry, and a proven track record curating ecosystems that combine the best of industry professionals and digital solutions.

Any opinions expressed here are the author's own.

NOT FOR REPRINT

© Touchpoint Markets, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.