March 7, 2023

AI Won’t Save You From Your COI Tracking Problems

Share on facebook
Share on google
Share on twitter
Share on linkedin

COI tracking software is an essential tool for businesses to manage and track certificates of insurance which are crucial documents that provide evidence of insurance coverage for a particular individual, entity, or property. These pieces of software make businesses more efficient all while keeping risk transferred off their plates

Some utilize AI and machine learning algorithms to analyze and categorize data, extract relevant information from certificates, and automatically update records. In those use cases, AI may be beneficial. The problem comes when AI is relied upon to make decisions. These programs rely on weak AI, called Artificial Narrow Intelligence (ANI), that can only act on a prescribed set of rules. Deviate from those rules, or have any amount of gray area, and you’re not going to get reliable output from that AI. That means your COIs mean next to nothing if they’re read incorrectly. Thanks, AI.

AI’s Use and Where it Fails

The use of AI and machine learning in COI tracking is only increasing, with many companies looking to automate their certificate tracking processes. However, it is clear that relying solely on these technologies to make decisions is a far cry from best practice. AI and machine learning algorithms, while powerful and efficient, cannot catch all of the nuances of contract language. Not by a long shot.

Insurance coverage contracts are super complex documents, often containing legal jargon and specific language that can be difficult to interpret without a deep understanding of the legal system. Nuance exists beyond what AI can understand, let alone detect. It only knows what we want it to know. While algorithms can be trained to recognize certain keywords and phrases, they lack the context and experience needed to truly understand the meaning behind the language. This leads to misreading COIs and making fallacious decisions based on bad assumptions. 

For example, imagine a COI tracking software that is being used to track certificates for a construction company. The software is programmed to recognize specific keywords and phrases related to construction insurance (among others). It may not be able to recognize more nuanced language that is specific to that industry. As a result, it may incorrectly analyze the COI leading to errors in record-keeping and potential liability issues for the business if/when it makes decisions based on that information.

The prevailing thought is to force AI to do what we want it to instead of it showing us what it can do. By taking this approach, we pigeonhole AI. Another issue is that depending on the language models the AI was trained on, it could have perceived or inherent biases built into how it makes decisions. 

Bias Illustrated 

We asked an artificial intelligence to explain how AIs can have a bias. Here’s what it said: “An AI can be biased if it is trained on biased data or if the algorithm or decision-making process contains implicit biases.”

We then asked if that specific AI has biases. It said, “…my responses can reflect the biases that may exist in the data I was trained on or the way the questions are asked. It’s important to keep in mind that while I strive to be as impartial as possible, my responses should always be considered with a critical and discerning mindset.” That last part is what sets us apart from some of the other COI solutions out there. Those discerning mindsets? They’re the insurance pros on our team who review your COIs with their human eyeballs and human minds. It makes a difference.

All AI has this issue. The AI built into your self-driving car or the one in your smartphone—they all have biases that are hard to iron out. The AI that created that text (ChatGPT) mirrors the same flaws used in some COI tracking software.

Insurance Language is Complicated and Ever-Changing

Another challenge in relying solely on AI and machine learning in COI tracking is that it cannot account for changes in contract language over time. Contract language constantly evolves, with new terms and clauses being added to reflect changes in the legal landscape and business practices. AI and machine learning algorithms need to be constantly updated and retrained to keep up with these changes which can be a time-consuming and resource-intensive process.

Moreover, contract language can vary widely depending on the industry, geography, and the specific nature of the transaction. As a result, relying solely on AI and machine learning can lead to errors and inaccuracies. In a world where errors affect bottom lines, these mistakes have financial implications.

Another issue with relying solely on AI and machine learning in certificate tracking software is that it can lead to a lack of accountability and transparency. If a mistake is made by the software, it may be difficult to determine who is responsible for the error leading to confusion and frustration, especially if the mistake leads to financial or legal consequences for the business.

Fun fact: OCR, or optical character recognition, isn’t AI. myCOI uses it to scan your COIs, but we don’t rely on it to make decisions beyond basic recognition of data in certain spots. 

You Need a Human Touch

Recognizing the limitations of AI and machine learning in certificate tracking, myCOI has always leaned in on people. We have specialized professionals on staff who have a deep understanding of contract language and legal frameworks to oversee this process. This can also involve incorporating manual checks and balances into the software to ensure that human expertise is brought in at key decision points. We pioneered this approach. It’s our secret weapon. It’s why we can warranty the decisions we help you make regarding your coverages and the legal mess it could imply. 

Incorporating human expertise into the certificate tracking process has several benefits. First, it improves the software’s accuracy and reliability by creating safeguards. No AI is bulletproof

Second, it increases transparency and accountability by ensuring there is a clear process for making decisions and addressing mistakes. 

Third, by incorporating people into the process, businesses gain insights into the nuances of contract language that may not have been captured by the software. This helps identify areas where the software can be improved and optimized over time, ensuring it remains up-to-date and effective. 

You Can’t (And Shouldn’t) Do It Alone

There is no substitute for real, human experts who know, definitively, what the best course of action regarding your coverages is—no machine can offer that. myCOI has built success on this. We help you ensure your compliance in a way that truly allows you to rest easy knowing something downstream won’t come swimming upstream and bite you in the pocketbook. Relying solely on AI to make these decisions opens up even more risk and liability for you.

Get this in your Inbox

The free myCOI Newsletter will send blog posts, industry news, and more straight to your inbox.


Is Your COI Process Healthy ?

How do you think your certificate of insurance tracking process stacks up?

We’ve got a quick, seven-question quiz you can take right now, free, that will help you gauge the health of your COI tracking system.