Happy New Year to everyone! As we step into 2024, we're excited to be back and are eagerly looking forward to a year filled with success and innovation. Wishing everyone a joyful and prosperous new year!
TLDR: At CourtCorrect, we have long been developing and deploying AI systems to improve and simplify the handling of consumer complaints in regulated industries. This week we have taken a huge step. We have a developed an end-to-end AI-based modelling pipeline that produces a detailed, action-driven root cause analysis ("RCA") report on the basis of a large number of individual complaints files. Think in: 3,000 complaint files, out: detailed RCA report. Below, we explore how we built this pipeline and why it matters (AI-generated RCA report linked at bottom of page).
Why is this tool useful for complaints teams, especially in regulated industries like financial services?
While there are many reasons to perform root cause analysis, we think there are three main motivations for complaints teams, especially in regulated industries, to consider using AI to decipher complaints data:
Data clustering unbiased by human preconceptions about pre-existing root causes, ie objective data analysis
Significant time and cost savings
Turbo-charging Consumer Duty compliance
Objective Data Analysis Using Unsupervised Data Clustering
One of the main challenges facing complaints teams that perform root cause analysis is that you usually end up finding what you're already looking for. After all, complaints teams spend most of their time reviewing customer complaints, so they usually have a good understanding of what kind of process or product failures cause these complaints to arise in the first place.
There is nothing wrong with this and it is a useful starting point. But complaints teams following this framework run the risk of simply sorting complaints into pre-conceived "buckets", rather than trying to objectively interrogate the data to identify both issues you may and issues you may not expect. Because it is by definition unsupervised (ie the model simply reviews the data without any preconceived notions of what it contains), AI-based RCA has the potential to identify both root causes that may already be on your radar as well as those that are not. This is a significant improvement upon the conventional approach.
Significant Time and Cost Savings
It may seem obvious, but it is worth stating still — for most organisations, root cause analysis is a massively time-consuming exercise. Often, it involves multiple teams pouring over troves of data for many hours to ultimately piece together an image of what is happening underneath. Our new end-to-end AI-driven pipeline produces similar or even better results in seconds. This frees up the time of complaints teams to focus on higher-value activities, for example to ensure that the insights derived from RCA are actually propagated throughout the business, ie that processes and products are actually fixed.
Consumer Duty Compliance
Finally, for many regulated industries like financial services, performing RCA is of course a strict regulatory requirement. It is not an optional "nice to have" but an regulatory "must", the absence of which can lead to enforcement proceedings. The Consumer Duty, introduced by the FCA in July 2023, of course requires firms to perform RCA, but arguably goes further by encouraging firms to engage in continuous improvements to avoid causing foreseeable harm to consumers. "Harm" surely becomes "foreseeable" once the firm is in or should reasonably be expected to be in possession of data indicating actual or potential harm. Root cause analysis is the remedial process by which firms can prove that they are indeed taking action to avoid such foreseeable harm and complying with the duty.
Given that compliance with the duty therefore requires an ongoing process of monitoring, identifying and remediating root causes of actual or potential consumer harm, the ability to generate reporting and management information (MI) repeatedly and effectively (ie by using AI) becomes all the more impactful.
Fortunately, the FCA seems to be quite open to the use of AI to satisfy these new and stringent requirements, as mentioned in their Annual Public Meeting last October:
“Firms need to invest in data collection that they can analyse in order to continually improve outcomes for consumers [..] we are pleased to see that firms are considering the adoption of AI and are thinking about the impacts on the firms and financial services industry. ” - FCA Annual Public Meeting 2023
Teaser: Root Cause Analysis with AI
At this point you must surely be waiting, mouth-wateringly, for an explanation of how root cause analysis can be effectively performed with AI to solve the various issues and deliver the many benefits described above.
You'll have to talk to us directly if you'd like us to spill the beans, but we will say this: this root cause analysis report linked here was generated 100% by AI on the basis of multiple thousands financial service complaints against a bank.
If you would like to learn more about what AI-driven root cause analysis can do for your organisation, please request a free trial of the CourtCorrect Platform here, fill out our contact form here or simply drop us a line at firstname.lastname@example.org. We'd be pleased to help you improve your RCA and complaints resolution model for the benefit of all.