Innovation Interviews: How Cardano embraces AI
Felix Mantz, director at Cardano Advisory, gives Pensions Expert a tour of how his company is using cutting-edge technology to enhance its services to trustees.
Many companies say they use artificial intelligence (AI), but quite often what they mean is advanced automation. This can improve efficiency and reduce errors but is often designed for specific and repetitive tasks.
In the case of Cardano, however, it seems different. The consultancy group’s various uses of natural language processing tools have made a tangible difference to its work on sponsor covenant – and it has even generated an AI-driven finance director to help train trustees.
Fresh from winning the Adviser Innovation Award at Pensions Expert’s Innovation Awards last month, Felix Mantz, director at Cardano Advisory, offered a sneak peek at how his company is embracing AI to benefit trustees, scheme sponsors and members.
Dear Sam
Trustees must go through a wide range of training to fully understand their role and responsibilities. Inevitably, this often means slideshow after slideshow and graph after graph. It’s no wonder that Cardano’s interactive negotiation training tool, dubbed “Dear Sam”, has proven so popular with its trustee clients.
When Mantz offers to introduce me to Sam, I jump at the chance to see it in action.
The tool is based on the same kind of technology that drives ChatGPT – an advanced natural language processing program that can interpret text inputs and give detailed, dynamic responses.
Mantz drafts a short opening negotiation letter, setting out a request to a sponsor for additional contributions to a scheme to help it prepare for buyout. Our hypothetical scheme is fully funded on a technical provisions basis, but needs an extra cash injection to reach a buyout-ready level.
Within a few seconds of inputting our first letter, we have a response in the form of a formal letter informing us – in firm but polite language – that the employer has no intention of providing additional capital. Sam is keen to work collaboratively with us and looks forward to hearing our proposals.
Given our time constraint, we quickly push for contributions to reach buyout-level funding, but Sam is having none of it. The responses each time are uncannily realistic, even when we suggest meeting over a coffee to discuss further.
At the end of the demonstration, the program presents a review of our negotiation approach and suggests improvements – giving more data, aiming for a compromise position – all of which are easily understandable and implementable.
“The trustees that we’ve used this with, they have all loved this,” Mantz says. “It goes down as the best training session they’ve had because it’s not just PowerPoint slides to flick through. You can get ‘hands on’ with this and test knowledge – you'll find the trustees that are strong on investments, and those that are strong on covenant or funding.”
Covenant computing
Cardano has also been enhancing its governance tool, Discover, using AI. One form of the technology processes reams of data from previous covenant assessments to make basic assessments more efficient. It can predict the most important likely drivers of scheme covenant risk using this huge database.
On top of this, another form of AI known as natural language processing is used to process relevant news stories and other written sources of information to highlight potential impacts on the sponsor such as industry developments and policy changes.
“From a trustee’s perspective, it means you get a notification when your covenant actually moves based on this information, rather than having to review every news story yourself,” Mantz says.
This combination of language processing and more “traditional” predictive AI provides Cardano and its clients with a powerful and practical tool that can support trustees in an efficient and reliable way.
What’s next?
Over the past year, Mantz explains, Cardano has been focusing on the “everyday efficiency” aspect of AI applications – automating repetitive tasks as much as possible.
He says: “As you get more efficient with something, and you automate more and more tasks that make up a job that you’re trying to do, all of a sudden you get to a point where that job is nearly fully automated. This makes it a game changer.
“We are looking for areas in which we can get a ‘win’ today, but that also lay a good foundation for us to build towards ‘game changers’ over time.”
Within the Discover portal, Mantz says Cardano is looking to expand its covenant tool to incorporate peer group data. Future developments could also include information relevant to the new Funding Code.
However the company plans to use it, AI looks set to play an increasingly important role in pension provision over the next few years.
Editor’s note: This interview was conducted before Mercer’s acquisition of Cardano was announced.