The AI Feeding Frenzy Continues

A Top Technology Initiative Aritcle – Feb. 2024.

In the fall, accounting software makers including Wolters Kluwer, Thomson Reuters, and Intuit all had their conferences. Every event included AI. Many publishers of accounting software hold their events in the early part of the year, typically as sales kick-off meetings and to announce new features. We’ve been lucky to be invited to the AICPA AI in Accountancy Symposium, the Acumatica Summit in Las Vegas, the Oracle NetSuite Influencers’ Conference in Austin, ZohoDay 2024 in McAllen, Texas, the CPA Practice Advisor Top Thought Leaders Symposium, and Sage Transform 2024. We have recorded podcasts on the highlights of each event to provide you with a summary of the publicly available information from the companies including insight and observations from fellow technology observers, writers, and consultants.

It looks like AI integration and business strategies for 2024 and beyond are being integrated. Key learnings on AI developments include 1) no one AI model will solve all the business problems at hand, and 2) AI will become a part of our products. We won’t think about AI, just like you don’t think about SSL or TCP/IP. Many of the vendors don’t have a good AI strategy but know they must have an “AI story,” so finding useful and real AI technology will be a challenge.

My co-presenter for The Accounting Technology Lab, Brian Tankersley, and I noted a variety of AI concepts at each event. While AI integration and models are immature, the few benefits being delivered are real. Continue to look for new product features that include AI technology in 2024, but buyers should continue to be cautious.

So, What Are We Seeing in the Frenzy?

Product management teams have returned to their senses and are looking for practical applications of AI in their products. Most of the wise product managers know that ease of use, accuracy, and practicality of the AI enablement is critical. Their teams have typically developed dozens of prototypes to test their ideas and to prove effectiveness.

One specific AI feature we observed in various prototypes was narrative reporting. The financial reporting tools are looking at results, explain key findings, allow for further review, and make recommendations for action or remediation. We can see narrative reporting being useful for board meeting preparation, advisory work with management teams & owners, and for explaining concepts shown in the financial numbers in a way that non-accountants can understand them. While some vendors are quite far along with their narrative reporting, and have already released an early version, most expect to deliver later in the year.

Another key concept was that the size of the AI model matters. We were particularly impressed with Raju Vegesna of Zoho’s explanation of the company’s use of narrow, small (SLMs), medium (MLMs), and large (LLMs) language models. Raju showed how they were applying different size models effectively within their product. For example, their popular Zoho Expense product includes a receipt scanning and OCR capability.

They showed how a small language model was more accurate on the receipt than a narrow language model and how both were more accurate than pure OCR. We’ve seen this same concept and approach used by Wolters Kluwer in their TeamMate Document Linker product, using AI to extract data more accurately than OCR alone.

Another conceptual model advanced by Raju of Zoho was AI + BI + CI = DI. That is, Artificial Intelligence + Business Intelligence + Context Intelligence = Decision Intelligence. He demonstrated this with prototypes of searching for documents using context for data, relations, and permissions. For example, if you are looking for documents on a project modified in the past week, can you do one search to find them all? Additionally, I hope all dashboard and analytics tools run in the future, like Zia (Zoho Intelligent Assistant) does, for profit analysis and diagnosis.

Effective integration is another area where AI can be used. Integrating AI to summarize content and assist with responses makes sense for routine tasks like email. Now that Microsoft Copilot is available for $20-30 per user for most business-level subscriptions, we can get assistance directly from Copilot with Outlook email responses. Likewise, Google Gemini Advanced with Ultra 1.0 is now available for Gmail for $19.99/user/month.

Similarly, Zoho demonstrated how they are applying AI in their mail platform with greater levels of privacy than any other provider. Accounting software providers like Acumatica, NetSuite, and Sage are also integrating email into their accounting products, but in the short term, product managers have to navigate some unreasonable pricing expectations from Microsoft, OpenAI, and others. Sensible integrations should appear with greater frequency in 2024.

Another application area of AI is Financial Reporting, as we discussed in my October column on Your Firm and Your AI Financial Reporting. However, in less than 90 days, demonstrations of prototypes have raised my expectations more. In addition to traditional KPI and dashboard reporting, the modeling powered by AI in various platforms is notably enabling.

For example, Nick Meyer, Vice President of Extension Products, whom I knew from his SAP days, and Jason Wells demonstrated both NetSuite Planning and Budgeting (NSPB) & NetSuite Analytics Warehouse (NSAW), which had both been enabled with AI. NSPB may be the best advancement in budgeting and planning tools I’ve seen in ten years, and the use of AI for forecasting was impressive. NetSuite has a strategic initiative to get all customers on NSPB, which seems doable with the $17,000 for 50-user price point.

So, What Makes Practical Sense with These New AI-Enabled Products?

In this column, I try to point out imminent or released products. In other words, attempting to give you practical advice and guidance from the trenches is paramount. I understand the “Osborne effect” and recognize that not all prototypes make it to market. However, I have far greater confidence that our traditional vendor suppliers of software will release useful AI-based products during the next year.

I could continue to provide examples, and there are many more to discuss. However, I wanted you to understand that the AI frenzy will continue but become more practical and integrated this year. I’m excited to see what the next 90, 180, and 365 days bring in AI-enabled accounting technology.