HomeSocial Impact HeroesC-Suite Perspectives On AI: Heather H Wilson Of CLARA Analytics On Where...

C-Suite Perspectives On AI: Heather H Wilson Of CLARA Analytics On Where to Use AI and Where to…

C-Suite Perspectives On AI: Heather H Wilson Of CLARA Analytics On Where to Use AI and Where to Rely Only on Humans

An Interview With Kieran Powell

How reliable and consistent is your data? When you have a reliable data source, you can apply AI with great success, but patterns cannot be derived from your data — a human must take on that task. I became a data hunter early in my career, seeking companies with access to rich data sets like Kaiser Permanente and AIG that would allow me to develop analytics and reporting tools that are actionable.

As artificial intelligence (AI) continues to advance and integrate into various aspects of business, decision-makers at the highest levels face the complex task of determining where AI can be most effectively utilized and where the human touch remains irreplaceable. This series seeks to explore the nuanced decisions made by C-suite executives regarding the implementation of AI in their operations. As part of this series, we had the pleasure of interviewing Heather H. Wilson.

Heather H. Wilson, Chief Executive Officer of CLARA Analytics, has more than a decade of executive experience in data, analytics and artificial intelligence, including Global Head of Innovation and Advanced Technology at Kaiser Permanente and Chief Data Officer of AIG. While at AIG, she was named the Insurance Woman of the Year by the Insurance Technology Association for her data innovation work. Wilson has been a steady supporter of diversity. She launched the Kaiser Permanente Women in Technology group, focused on mentorship and retention for women in math, technology and science, and at AIG, she launched Global Women in Technology and served as Executive Sponsor of Girls Who Code.

Thank you so much for your time! I know that you are a very busy person. Our readers would love to get to know you a bit better. Can you tell us about your “backstory” and how you got started?

My professional journey evolved through three chapters: consulting, data leadership, and my current role as CEO of CLARA Analytics. I was very influenced by my college career in the 1990s. I graduated fluent in languages with a focus on international business, just as global trade and international relations were becoming central on the world stage. To prepare for this global landscape, I pursued multiple internships in various countries, which led me to the first chapter of my professional life as a consultant. I chose consulting to learn how to adapt to new client situations, understand complex issues, and develop solutions. This role honed my skills in client service, delivering engagements, and managing complicated stakeholders, all of which have been crucial in shaping me into the leader I am today. I sometimes call our team “claims therapists” because of the issues that they’re dealing with. Additionally, it gave me the ability to think about this world of data and analytics.

As a Chief Data Officer and Chief Data Scientist across various industries, I applied data and analytics to solve critical problems. Within healthcare, we focused on preventing the onset of chronic conditions, and in financial services, we detected fraudulent activities. Another example is helping a retail space understand which products should be in which stores and what’s trending for patterns. So all of it was kind of a build for me of the three chapters professionally in consulting, then being a chief data officer or chief data scientist in several industries, which lead me to where I am now. These experiences in leveraging data to drive business decisions across sectors have culminated in my current position, where I continue to apply this multifaceted expertise.

It has been said that our mistakes can be our greatest teachers. Can you share a story about the funniest mistake you made when you were first starting? Can you tell us what lesson you learned from that?

As I mentioned, the first chapter of my career was in consulting, and in consulting, you will do anything to make sure that you’re delivering for your clients. Of course, there’s engagement and a binding contract, but you really will go above and beyond to deliver exceptional results in hopes of building strong relationships — and to collaborate on future projects. So my transition from consulting into industry required a significant shift in approach. It became crucial to pause and ensure proper socialization of ideas; it’s not just for a consultant to push it through the brick wall. So while my intentions were always positive, I had to learn and understand that a socialization campaign helps to move ideas along to a place of execution.

At the start, I didn’t do that. When I first went into industry, some of my projects didn’t move forward as I thought they should have. And had I known then what I know now, years later, decades later, I would have spent more time in that first role, engaging more with my colleagues on some of the initiatives and understanding that it wasn’t about me — it was really about all of us moving together.

Are you working on any exciting new projects now? How do you think that will help people?

I took on the role of CEO of CLARA Analytics because I’m very passionate about the outcomes that we achieve on so many different levels. Yes, there are financial outcomes through a tremendous ROI of 5x to 20x. Yes, there are operational efficiencies that come about because our tools are a second set of eyes over documents and surface insights or items in the documents that could be missed. But I think the reason that I accepted the role is because of our strong focus on the claims handler and the claimant.

The claims handler is similar to a triage nurse in an emergency room. They have a plethora of workloads and know where to focus and where to prioritize so that they’re not bogged down every day deciding which file to look at and which decisions to make. A decision support tool like CLARA puts a microscope over the haystack and allows claims handlers to surface which files are most important to focus on and which decisions they need to consider. We refer to it as a GPS because you may know all the different routes of these cases based on your experience, or you may not know because you’re a new adjuster. We want you to drive the case like you would drive a car, and CLARA gives you the heads up on the traffic, the construction and the accidents. We want to be in front of you and bring all critical details to the surface because as an adjuster, you have only so many minutes in the day, and we want you to spend time with the injured party so that they can return to work faster. The tools that a claims adjuster uses are so important because the cases are only getting more heightened, whether it’s from a litigation perspective or the complexity of the cases.

Thank you for that. Let’s now shift to the central focus of our discussion. In your experience, what have been the most challenging aspects of integrating AI into your business operations, and how have you balanced these with the need to preserve human-centric roles?

At CLARA, we truly believe in human-centered design and having humans in the loop. I often use the GPS analogy because whether it’s Google Maps or Waze, you’re still driving the car. We need those adjusters to take the wheel in managing their cases, while CLARA, akin to your Waze app, is there to highlight the issues that require your attention and guide you in managing those cases.

It’s really important that you know we are non-prescriptive. We provide various routes based on insights from our models, but the choice remains with you, just as it does when you decide which path to take while driving. A claims adjuster’s decisions may be influenced by factors like cost or the duration of claims processing, and CLARA is designed thoughtfully to support these choices without dictating them. Additionally, we equip adjusters with objective data derived from their claims files, adding a layer of intelligent predictability. This allows them to stay ahead of cases, enabling them to deliver the level of service that claimants expect during challenging times.

At CLARA, we believe our AI platform enables the claims handler to approach each case with empathy because they have the data at their fingertips regarding what they need to address with the claimant. We believe that’s how you give empathy to a claims handler, who can then, in turn, effectively channel that empathy into meaningful engagement with the claimant.

Can you share a specific instance where AI initially seemed like the optimal solution but ultimately proved less effective than human intervention? What did this experience teach you about the limitations of AI in your field?

There is a growing focus of interplay between humans and AI models. As expectations rise, particularly regarding the accuracy of models when detecting information within documents and claims files, it’s important to acknowledge that there are times when models are not 100% accurate, when a model may have missed a data point. What’s good about machine learning is that in our AI platform, we have a feedback clue where if the adjuster is not taking action and provides us with reasoning on why, then the model picks up those notes. This process helps the model understand how to better handle specific files and align with the adjuster’s preferences.

Just as humans are lifelong learners, our models are designed to continuously learn and adapt. There are times when an AI model may struggle to interpret information from a document. For example, it can’t read something that appears on a document, so we make sure that we send back that document. And sometimes a human can’t read it either. It’s very important that we’re able to surface items that we’re unable to read. But there are times when a model will miss data points. With CLARA, we have a wonderful feedback loop that enables us to teach and retrain our models, ensuring that they improve over time and become increasingly accurate in their analyses.

How do you navigate the ethical implications of implementing AI in your company, especially concerning potential job displacement and ensuring ethical AI usage?

Insurance is a highly regulated industry. However, it’s not just about the regulations or security. One of the things that we make sure of for the efficacy of our models is that we test constantly for bias and that the data is not passed into our models, especially with the medical documents we receive. This allows us to not have bias when we are sending outcomes, insights and recommendations to the claims professional through CLARA’s platform. Also, we remain very objective. We do that by the data that we use on our AI platform. We use the data that we receive from our clients through their core claims, HRIS system, or documents they send us from their medical bill review or legal review entities. We choose not to use data that could be pulled from public sources which can be more subjective.

Maintaining objectivity is crucial for us, and we are dedicated to adhering to our ethical guidelines. This commitment ensures that we operate within the regulatory framework established for our clients, allowing us to provide reliable and trustworthy services..

Could you describe a successful instance in your company where AI and human skills were synergistically combined to achieve a result that neither could have accomplished alone?

A great example of a human and AI partnership is with our claims document intelligence solution, CLARA Claims Document Intelligence Pro. Claims DocIntel Pro analyzes, summarizes and provides insights on medical records and legal demands related to a claim that can sometimes be thousands of pages long. These complicated documents can take hours for a claims professional to review. The challenge for AI is that the documents include document types not publicly available, with legal and medical terminology that does not translate well in the context of a claim with out-of-the-box large language models (LLMs). We solved the first problem of document type identification with human readers that tagged and identified document types in order to train our AI. The partnership between human labor and AI solved a key issue, but summarization presented a set of new challenges.

Summarization of key medical terms and legal language in the context of a complicated bodily injury claim is a task that can only be quickly completed with both human and AI cooperation. Claims DocIntel Pro will summarize and analyze the claim for the adjuster who then evaluates the output and provides any needed corrections or context. This human-in-the-loop model continues to train Claims DocIntel Pro, making it even more powerful. Regardless of how well-trained the AI model is, high severity claims with bodily injury will always present unique challenges. Claims DocIntel Pro can quickly complete 90% of the hard work and augment the claims professional’s ability to make data-informed decisions. The professional must still use their judgement to refine the output and make nuanced decisions, all leading to better claim outcomes. Neither task could be done alone and be accomplished in a reasonable time with the level of accuracy required to get the best outcomes — a true example of AI and human partnership working together to accomplish time-consuming tasks in just minutes.

Based on your experience and success, what are the “5 Things to Keep in Mind When Deciding Where to Use AI and Where to Rely Only on Humans, and Why?” How have these five things impacted your work or your career?

  1. How reliable and consistent is your data? When you have a reliable data source, you can apply AI with great success, but patterns cannot be derived from your data — a human must take on that task. I became a data hunter early in my career, seeking companies with access to rich data sets like Kaiser Permanente and AIG that would allow me to develop analytics and reporting tools that are actionable.
  2. How much EQ is needed to successfully complete the task? Human engagement is needed for an empathetic experience. It certainly has been wonderful to develop AI solutions that drive massive efficiencies, but the real magic happens when you combine insights from AI with the EQ of a human to deliver on the best experience. Working in insurance claims, I have learned the real value of human engagement, especially when injured parties are involved.
  3. How much data is required to read and process for good results? AI is a perfect tool for the avalanche of data that is available to the modern worker. Predictive analytics and generative AI tools can process massive amounts of data and connections that a human could not process in a lifetime. Look for AI to scale your team with augmented intelligence for the more nuanced tasks.
  4. Can you deliver a good user experience? When making a decision to either automate with AI or use traditional processes, you need to take into consideration the cognitive load and context switching required by the user. If the AI-enabled process disrupts workflows or requires a user to switch from screen to screen, you should rethink the application. In my time at CLARA Analytics, I have focused on getting our AI platform integrated with core systems like Guidewire and Origami Risk to reduce the amount of context switching for the adjuster. We also have developed our platform to be a headless application, allowing for insights from outside sources to be seamlessly integrated to give the best user experience and a single view of outstanding risk.
  5. How severe and frequent are your cases? You should understand severity and frequency to help guide your AI strategy. For example, in the insurance world, low severity and high frequency claims can generally use AI to accelerate claim closure with minimal human intervention. On the opposite side of the spectrum, higher severity claims with lower frequency really need to have AI to provide context but have much more human intervention to help prevent escalation.

Looking toward the future, in which areas of your business do you foresee AI making the most significant impact, and conversely, in which areas do you believe a human touch will remain indispensable?

In casualty claims situations, it is essential that a human representative engages with the injured party to manage the case effectively. The best way to manage that case is to equip a human with an AI platform that offers valuable insights. This technology enables them to grasp the potential developments of the case, ensuring informed decision-making and better care for the injured party.

Our competitive edge lies in our extensive contributory database of millions of cases. With over a decade and a half of historical data, we can draw on past experiences when a new case arrives. This allows us to predict potential outcomes and identify key factors to consider, serving as early indicators and/or warnings of possible developments in the case. The two working together, I believe, give the best service to a claimant, with the soft skills and then the predictability and the GPS in front of the routes of your files. Both create a significant impact and foster strong engagement with a claimant.

You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. 🙂

I am a great believer that insurance is there when bad things happen personally or commercially. I also believe that insurance can play a bigger role in making the world safer — workplace safety, road safety, trucker safety, car safety, product safety, etc. The velocity and volume of claims data across the insurance industry can be highly predictable, helping identify patterns for prevention and safety for the greater good of people across so many environments. Trending millions and millions of aggregated, anonymized claims could make the world safer!

How can our readers further follow your work online?

Heather H. Wilson: www.linkedin.com/in/heather-h-wilson-40a61b8/

CLARA Analytics: https://claraanalytics.com/

CLARA Analytics on LinkedIn: www.linkedin.com/company/clara-analytics/

CLARA Analytics on X: @claraanalytics; https://x.com/claraanalytics

CLARA Analytics on Facebook: www.facebook.com/claraanalytics

This was very inspiring. Thank you so much for joining us!


C-Suite Perspectives On AI: Heather H Wilson Of CLARA Analytics On Where to Use AI and Where to… was originally published in Authority Magazine on Medium, where people are continuing the conversation by highlighting and responding to this story.