Skip to content


Marc Jones is Shift Technology’s CTO

Every few years a new technology emerges that seems to capture the zeitgeist perfectly. Some of these technologies never fully reach their potential and are remembered fondly as an “oh, what could have been.” Others, like the smartphone, are true disruptors that changed or created industries, and made their mark on the world forever. With the emergence of generative AI, it feels like we are in that place all over again.

One of the crucial elements of a “world changing” technology is the potential range of its impact. The Segway, for example, was going to change the world of personal transportation. 20 years in and it is increasingly rare to encounter a Segway rider out in the wild. Clearly, the Segway’s impact was quite limited. The smartphone… need I say more? And with this criteria in mind, we should clearly put generative AI much closer to the smartphone camp than the realm of the Segway. 

Yet, for a new technology to be considered truly revolutionary, it must have applications across multiple audiences, industries, and vertical markets. It must be able to be used safely, by consumers and businesses alike. And it must be truly useful over time. Generative AI fits this bill.

Required Characteristics of Transformative Technologies

When talking about a transformative technology it is important to understand that not all industries or individual users will adopt it in the same way. And even within a single industry, we may be talking about very different use cases depending on varied business requirements. However, there are some constants that must be considered when evaluating the potential of a transformative technology. These include:

  • Security - can the technology be deployed and used safely without exposing users to undue risk?
  • Effectiveness - does the technology represent a shift in the status quo that allows users to accomplish something more quickly and efficiently, or even new things that simply weren’t possible before?
  • Accessibility and scalability - is the technology able to be used by those who want to when they want to?

If a technology meets these basic requirements there is a good chance that we are talking about a true game changer. So, how does this apply to generative AI and the insurance industry and how can insurers make certain their use of the technology hits these three crucial points?

Making Generative AI Secure

For insurers thinking about using generative AI, security is one of the primary concerns. The solution must be safe and secure in order to protect both the business and its policyholders. One of the best ways to ensure this is to work with a trusted partner who has a proven track record of deploying enterprise solutions with security at the forefront. For example, Shift is working with Microsoft and its Azure OpenAI solution to add new generative AI capabilities to our insurance decisioning solutions, making it easier to bring generative AI to the insurance enterprise in a safe and secure way.

Another key aspect of making generative AI as secure as possible is related to the data these solutions ingest to do their job. It is critically important to know where that data is coming from and who has access to it. For example, will your generative AI solution have access to data outside the enterprise (maybe not such a good idea), or will it only be able to access data from within the enterprise that has been carefully cleaned, denoised and curated?

Making Generative AI Effective

Like all artificial intelligence, Generative AI is only as good as the data it is provided. All too often we hear the refrain that more data is better when it comes to AI. In the case of generative AI, that may not be the best way to think of things. We have already addressed the concept of data security and why cleansed, curated and denoised data may be more secure than collecting directly from the internet or other public sources. When we are talking about the effectiveness of generative AI, those same qualities are equally important. The goal of generative AI is to produce outputs that benefit the user. 

For individuals experimenting with technology such as ChatGPT, the desired output may be a first draft of a business email or a similar document. In such a situation having access to a lot of different data may be useful. However, insurance is an incredibly nuanced industry and the outputs of generative AI will have a very specific business purpose. Insurers may, for example, be looking to generative AI to augment fraud detection capabilities or support claims automation strategies. In that case, a more nuanced approach to making data available to the solution is required. By carefully curating what data is accessible to the generative AI solution in use you can train your models more efficiently and effectively and ensure that the responses generated will have the correct context and be truly effective for the business.

Making Generative AI Accessible and Scalable

Any technology solution is only as useful as it is accessible and scalable. For many insurers, that will mean making generative AI an integral part of the solutions their employees are using everyday. As opposed to deploying a standalone solution and encouraging employees to “figure it out” embedding the technology can help ensure seamless adoption and quickly received benefits. As part of an established technology solution already deployed in the enterprise (again, think fraud detection or claims automation solutions) employees don’t need to learn and perfect a new way of doing things. The new way is simply part of the technologies they use everyday, just providing more effective results.


Generative AI shows real potential to deliver positive impacts to not only the insurance industry, but many others as well. For insurers, it will be critical to understand where generative AI can deliver the most benefit and develop winning strategies for incorporating the technology into their current mix. Security, effectiveness and scalability will be the keys to success as insurers determine how to make the most of generative AI.

For more information about how Shift can help you adopt  generative AI to meet the unique challenges facing the insurance industry – contact us today