MindBridge CFO Matthias Steinberg has a lot of empathy for finance leaders under pressure to harness generative AI in order to improve their finance operations and companies. 

The main tension, as he sees it, is the gap between the creative, language-oriented strengths of the newer generative artificial intelligence technology and the numbers-oriented talents of the older machine-learning technology that is good at processing and analyzing massive amounts of data. 

The CFO and their finance teams face high expectations driven by the market hype around generative AI, Steinberg told CFO Dive in an interview. But, right “out of the box,” the large language model-type genAI is not ideal for working with numbers, he said.

“Currently, the work is being done to combine large language model systems with machine learning and other technologies to build tools that can really work with an enormous amount of data,” Steinberg said. “That creates a lot of friction and pressure for CFOs to deliver tangible results while the technology is still getting there.That’s where I see a lot of the pain.”  

Ottawa, Canada-based MindBridge’s business model sits squarely at the center of that pressure and it’s not alone. When Hewlett Packard Enterprise CFO Marie Myers doubled down on the use of its agentic AI tool “Alfred,” named after superhero Batman’s trusted butler, HPE worked with Deloitte and Nvidia to make sure it got “deterministic outcomes,” meaning it gets the same answer every time it asks the same question.  

That non-deterministic element of LLMs can be very valuable for marketing, but is a concern in finance. “It appears we do like our numbers to be 100% correct and it needs to be deterministic, and if I ask my tech the same question three times it needs to be, three times, the exact same answer,” Steinberg told CFO Dive, noting that the solution lies in combining the different technologies. 

MindBridge’s SaaS platform is used by auditors and other companies to find analogies and risk in financial data and systems. Steinberg describes it as a kind of monitoring tool that automates all the work of assurance with a company’s ledgers but which does not replace enterprise resource planning systems. 

“We’re not trying to replace that,” he said. “We ingest the data that's in the ERP systems and increasingly also other operational systems, like a booking or billing system. We ingest the data, run the analysis and play that back to the user in its simplest form, telling the user, here’s potential for high risk items that you need to follow up on.”  

Those users have included large audit firms such as Big Four firm KPMG as well as corporations like Chevron. While the pricing of the analytics systems vary, clients pay annual licensing fees starting in the low six-digit figures, he said. 

Founded in 2015, MindBridge was first evolved using unsupervised machine learning to drive its analysis, but it is now working on a new product with agentic AI that it will launch in several months. That new agentic wrapper will transform and simplify the user’s experience from one with multiple screens and analytics required, to a more streamlined interface.

“Up until today, an auditor or financial professional would make sure the data is loaded into our tool, the tool runs and there’s billions of combos that need to be done. Then, that all gets played back to a different dashboard and the user does all kinds of slicing and dicing,” he said. “In the future, you open your browser or tool in the morning, there’s a chatbox and all you do is say, ‘what do I need to do, what has happened in the last 24 hours?’” 

Steinberg, who has held the finance reins at MindBridge for nearly four years, has an MBA from INSEAD and a master’s degree in engineering from RWTH Aachen University, according to his LinkedIn profile. Earlier in his career, he got experience in private equity, working with The Boston Consulting Group and Summit Partners. As CFO he also helped take the Germany-based Ionos public, before moving to Canada to take the job with MindBridge.

When it comes to advice for other CFOs looking to tap AI, Steinberg believes it’s important to pick a relatively discreet project to inject AI into, such as accounts payable or investor relations, to find champions for change who oftentimes are younger staff, and to keep it contained to the finance team. 

“There’s no silver bullet and sometimes it’s overwhelming," he said. “But take one project and get going.”