Takeaways:
- In 2023, The Standard built its own AI chatbot for internal use
- An AI team established in 2024 applied Microsoft Copilot to build on the chatbot
- Current AI focus covers productivity, solutions and development
Greg Chandler, executive vice president of information technology at The Standard, has worked for the insurance and financial services company for 17 years. The Standard markets its products to employers on a B2B basis. About three years ago, Chandler began leading the company's AI efforts, focusing on internal uses at the company, for about 3,000 employees. He began his current role in August. Digital Insurance spoke with Chandler about the progress The Standard is making with AI and its future plans for the technology.
This article is excerpted from an interview and a follow-up email, and edited for clarity.
How did The Standard introduce the use of AI in the company?
We began our AI efforts in 2023 with an effort I sponsored called AskStan – an AI-based chatbot internally built by a volunteer team. We held weekly demonstrations and iterations within IT until it was near MVP. The internal system was primarily leveraging LLMs obtained outside and re-hosted internally for privacy reasons and was used for asking questions around corporate policies.
This initial effort was as much a technology evolution as change management. The approach of testing AskStan on our employees with an ability to provide real time feedback to the tool got them acquainted with the technology in a work environment. Multiple learning forums and feedback sessions were held and through the effort, people became less intimidated by the technology and learned how to increase productivity as a result.
It's a unique environment here, where the value of AI is seen from the CEO on down. There's absolute commitment and actual understanding of the role AI can play in our industry, particularly at our company. I thought it would be a little bit tougher road, knowing we had to get our mind around how we leverage this technology here.
How did you further develop the company's use of AI?
Over the course of 2024, we built a dedicated Applied AI team, hiring impressive talent from competitors, MIT, and other industries. This team continued to evolve the original AskStan technology, but at the same time were developing new approaches to solve the much harder challenges our business faces.
In parallel, tools like Microsoft Copilot continued to evolve such that it made sense for us to transition the AskStan policy engine over to Copilot. We now have deployed 100% of Microsoft's Copilot technology to every employee in the company. Questions originally posed in our AskStan MVP are now handled by Copilot, as well as numerous other productivity-oriented uses. We have over 85% usage of Copilot on a daily basis across the company.
We got people to share what they were asking AskStan, and how they used it. That lowered the temperature in terms of fear of AI or threat of job elimination, because that is not our focus at all. Our focus is deep customer service empathy. Our employees are talking to customers at a crisis point – a loss of a loved one, a disability – something where empathy and that human-to-human connection is really, really important. We're trying to enable our employees to spend more time with our customers, by creating tools that allow them more time to do that, reducing the time it takes to research and get information for them.
What is The Standard's current strategy for AI?
Our AI strategy is three-pronged: Copilot for broad employee productivity; SaaS or fit-for-purpose solutions for commodity use (Contact Center AI, Salesforce assistants, etc.); and internal development to solve business-specific problems.
Initially, our customers will see AI from a Contact Center perspective and be indirectly benefitted by internally developed AI that speeds claims approvals, for example. Our internal use cases include claims, underwriting, commercial mortgage and investments. We believe there is more value in solving the business process-oriented problems that bring real money to the bottom line.
Over time, we expect externally facing chat interactions focused on certain customer use cases. Given we have three distinct types of customers – employers, brokers and claimants – those use cases are very different. For employers, it might be asking for certain insights into historical claims experience, or absence trends of employees. For brokers, it might be asking into products, or revenue generated from closed cases with us.
For claimants, it might be asking what types of coverage they have or the status of a claim, and would be multi-channel in nature. The AI will do all the heavy lifting to present a claims advisor all the information, as well as a recommendation to approve or deny. The human will always be the one to review and approve or deny the claim, which still saves hours of time per claim.
I'm working to reimagine the processes that underpin claims administration or claims approval, developing tool sets to realize value.
Are there other functions that AI could support at The Standard?
We envision numerous capabilities that AI can accommodate in our business. Some capabilities include eligibility determination, case severity determination, approval or denial recommendations, fraud detection, RFP prioritization, evaluation and product matching, underwriting, and RFP response creation, as well as many others.
If I have limited time and expertise from underwriters, I want to first focus on what makes the most sense, the most profitable cases that are coming in, that we align with. With a broker that has a better success rate, you can use that initial filter to create efficiency for underwriters.
As I continue to grow the capabilities, we have a higher probability of exponentially realizing value across the enterprise, versus one off solutions.
