Walled Garden LLMs Build Enterprise Trust in AI

It’s been nearly a year since generative artificial intelligence (AI) tools first hit the shelves.

In the time since, the technology has captured both the public imagination and lawmaker attention, but a clear game-changing impact that assuages concerns about data security and privacy, as large language models (LLMs) tend to hallucinate misinformation, has yet to emerge specifically for the enterprise.

After all, most future-fit back offices were already chock full of predictive AI, machine learning (ML), and robotic process automation (RPA) solutions meant to solve for legacy pain points and manual workflow bottlenecks.

This, as tapping generative AI for intelligent document processing and real-time information surfacing is increasingly looking like it will be, if not a grand slam, the solid base hit that sees generative AI capabilities start to be integrated across the broader business world.

In a sign of the changing times, Morgan Stanley announced Monday (Sep. 18) that it is launching a new generative AI tool for financial advisers and their support staff that is designed to better facilitate access to the bank’s massive library of research reports and documents.

The internally-focused software platform is a bespoke solution based on OpenAI’s GPT-4, and it isn’t the only one under development at the bank. Morgan Stanley is reportedly also currently piloting another gen AI tool called Debrief, which automatically summarizes the content of client meetings and generates follow-up emails.

Also on Monday, enterprise generative AI platform Writer announced a new $100 million funding round to help further commercialize its LLM for the enterprise product — a sign of investor confidence in gen AI’s business-centric utility.

Read more: Generative vs Predictive AI’s Role Across the Future of Payments

Generative AI Usage Within Enterprises Hitting an Inflection Point

By leveraging natural language processing (NLP) for document analysis, AI can unlock the power of organizational data in the blink of an eye, representing a capability phase shift for departments like accounting and finance, even legal.

Real-time techniques for automatically reading, understanding and analyzing business documents will not just help firms streamline their internal processes, but will have downstream effects on areas like accounts receivable and payable processes by eliminating tedious manual data entry requirements while ensuring accuracy and compliance.

“We are in that economic cycle where every cost you can beat out of the process is necessary right now … how to save money and how to eliminate those manual steps in the processes is top of mind,” Ingo Money CEO Drew Edwards told PYMNTS.

And some of the most promising efficiencies to find are in the finance department.

“Whether it is an accounts receivable report, accounts payable report, treasury report, cash flow forecasting, or something else, there is a time lag between having a report from a financial system and being able to take action,” Veena Gundavelli, the founder and CEO at Emagia, told PYMNTS. “Many finance executives are slowed down in their decision making because they have to spend time on analysis … but you can give AI any finance document like a remittance, invoice, bank statement or a lockbox image, and it understands and reads the data and produces the data.”

By removing the time spent by finance leaders on analysis, generative AI can help speed up analysis and accelerate the agility of organizational decision-making.

Read also: Can AI Fix Healthcare’s Zettabyte-Sized Data Problem?

The Power of Automated Document Data Extraction

“If you think about a typical organization, one that’s been in business for 10-plus years, and consider all the data that they have amassed, particularly unstructured data — this data was generally produced and then shelved,” Taylor Lowe, CEO and co-founder of LLM developer platform Metal, told PYMNTS. “[Gen AI tools make it] much easier to access and retrieve, and people can move much faster as a result.”

Image a finance employee being able to ask a spreadsheet using NLP the questions they want analyzed rather than having to model out their own analysis — this is the future that generative AI integrations can help build.

“The amount of paper that is still passed around in the B2B space continues to stun me, and it’s somewhat by choice, but more and more, I think businesses are looking for a better way,” Shawn Cunningham, managing vice president and head of Capital One Trade Credit, told PYMNTS.

PYMNTS Intelligence finds that automating manual tasks is an easy win for finance teams. Roughly 63% of CFOs in a PYMNTS study noted that the increased speed from AR automation reduced invoicing errors.

Improving payments allows these companies to focus their efforts on product innovations, mergers and acquisitions, joint ventures, alliances and partnerships to solidify and improve market position — the same way that integrating generative AI capabilities allows staff to free up their time to work on higher-level, strategic initiatives that give more back to the business.