Prevent costly errors with AI-driven process mining, ensuring smoother workflows and smarter decisions across industries.
Have you encountered a bad situation that was made worse by something that is meant to help? Here’s a recent example of mine – I had to take my son to an emergency room while vacationing in Asia but the most frustrating part was dealing with insurance when we got home. The agent who initially processed my claim put me (and my money) in limbo – no external or internal follow-up communication, inaccessible and invisible in the client portal – because they didn’t follow the process for handling non-English documents. This poor customer service was entirely preventable and, though I’m not an insurance industry expert, I’m going to tell you how.
I started this article with my personal experience because all service providers need to consider customer impact when designing their AI adoption. Unfortunately for me, health insurance is a relatively inelastic service. The insurance company – let’s start to see ourselves in their position now – has many customers locked in for the year irrespective of individual satisfaction. It also means that customer acquisition is relatively fixed. Insurance companies are not alone in having profit margins that are won and lost in processes. They’re also not alone in having a customer base that includes stubborn engineers who will spend above-average time investigating problems to discover a root cause (hi, that’s me). Even though I can’t switch medical insurance, the original agent’s mistakes followed by my persistence led to an undesirably high touch time for the insurance company (getting personal again, I digress…)
Whether your organization manages insurance claims, manufactures automotive components, or facilitates the food and beverage supply chain, profitability is influenced by how well your people, processes and systems are harmonized. Fortunately, some of the up-and-coming solutions embedded with AI have started to measurably improve the balance with people, processes and, ultimately, profit. One of the solutions with a high yield potential from relatively low effort is called Process Mining. Gartner defines it as “a technique designed to discover, monitor and improve real processes (i.e., not assumed processes) by extracting readily available knowledge from the event logs of information systems”. What gives process mining the potential for high yield with low effort is that it leverages information that your business processes already generate but traditionally ignore outside of IT troubleshooting. Process mining users are provided with unprecedented visibility of process flows and deviations. Analysis of those deviations turns into data-driven continuous improvement with the possibility of incorporating process improvements that were already proven through execution even though they weren’t pre-planned.
Here’s how process mining could have saved present and future costs for the insurance company in my scenario.
- Digital process definition: Ahead of my problem, the business process flow for routing claims would have already been discovered and acceptable paths defined.
- Omnipresent monitoring: My claim being dumped in no-man’s land would have been flagged as a deviation by process mining not finding any status update after a defined duration.
- Active alerting: The deviation would have alerted a supervisor to investigate the claim and properly reroute it, saving a future agent from having to chase it down.
- Continuous improvement: From there, the supervisor could have assigned the specific training course on dispositioning non-English documents, preventing a future occurrence and/or laying the foundation for a performance improvement plan. Some process mining vendors provide enhancement features that leverage AI to recommend process improvements which, in this case, could have led to an organization-wide improvement based on the lessons learned from this single process deviation
Since process mining and process intelligence solutions use data available in information systems, companies can gather new insights without having to upgrade entire systems. This means that my insurance company could have a years-old claims management system (let’s be honest, they do) and still deploy process intelligence on it. It also means that the company could leverage the process intelligence tool on their IT ticket management for tracking technical support, their ERP for tracking a procure-to-pay or order-to-cash process, the list goes on.
A case has been made for insurance processes, and hopefully you’ve caught onto the potential this technology has for your organization. Now how do you start a pilot project? Process mining is a relatively new space with a few established front-runners and several interesting challengers. Here are five tips to find the right partner for your organization:
- Check for relevant use cases: Since most enterprise software generates the minerals that process mining seeks (log files), the applications are plentiful. Rather than starting a science experiment, make a short list of high-volume processes in your company and survey vendors’ case studies to make sure a few match.
- Value business expertise in your industry: This is a new solution space and most of you work in industries that have deeply embedded practices and terminology. Imagine you’re in the automotive supply chain and have to answer, “What’s this EDI step?”; or handle perishable foods and hear “Do we need to keep this cold chain process intact?”. Your projects will move much faster with a partner that is familiar with what you do.
- Don’t underestimate technical integration: Any good project manager knows that software setup is a critical path activity, and that time is money. Technical integration of the process mining solution to your ERP, CRM, etc. is required. A select few process intelligence vendors built their reputation on those core systems and offer process mining as part of their suite. If you have Microsoft, QAD, or SAP ERP and supply chain software you should inquire about pre-packaged process mining integrations and use cases.
- Evaluate the differentiators: Each vendor hopefully has a unique angle as to why (they believe) they provide above-average value. It’s fine to approach this with skepticism, but it could serve you to consider how that uniqueness could value you. One example is real-time alerting on specific units – this enables you to take faster and more precise action when process deviations inevitably occur.
- Know your users: Your organization can probably find areas to save with process mining whether or not you have a well-staffed continuous improvement / digital transformation team. If you can’t make it somebody’s regular job to be in a process mining solution, think hard about how the software can assist more sporadic users. Generative AI is making its way into enterprise systems, and a secure LLM could accelerate the occasional user’s interaction with process intelligence.
My ordeal with insurance claims highlights a critical lesson for companies with customers (that’s all of us): effective integration of AI solutions, like process mining, can significantly enhance customer experience and operational efficiency. The example underscores the importance of harmonizing people, processes, and systems to mitigate preventable mistakes and improve profitability. Process mining, by leveraging existing data to provide deep insights and continuous improvements, represents a promising approach to achieving these goals. As organizations contemplate implementing such technologies, it’s crucial to consider relevant use cases, industry expertise, technical integration, unique vendor differentiators, and user needs. By doing so, you’ll not only resolve current inefficiencies but also pave the way for a more responsive and streamlined future.
Explore AITechPark for top AI, IoT, Cybersecurity advancements, And amplify your reach through guest posts and link collaboration.
