A finance leaders' guide to assessing generative AI

Don't be the guinea pig.

Finance leaders are always searching for solution providers who understand their pain points and can offer ways to improve and grow the business. AI builders, especially those using generative AI, promise to solve these problems. But most of their finance tools and GPTs are still in their infancy. They can’t yet provide useful feedback for optimizing policies, answer vendor queries with real invoice data, or offer real, data-driven insights.

While there’s nothing wrong with exploring a new technology, the last thing your finance team needs is to be an AI automation test subject. So what's the best way to be sure you’re buying a mature, effective solution?

Have you noticed how AI writing tools use words and phrases like “Furthermore,” “In conclusion,” or “In the ever-evolving landscape of finance…”? It sounds like a 6th grader learning to write essays from a template: impressively vague. It can be prompted to sound more human, but it takes work to get there. And yet, many of the companies creating finance AI automation solutions write about their generative AI using these poorly designed tools. It makes you wonder how well they truly understand the technology.

Over the last year or so, legacy automation solutions that have handled expense audits and invoice processing in the past have jumped on the AI bandwagon. These companies pitch it like their products can now do it all. The truth is, it’s aspirational.

That’s because the same problems of scale, ease of use, and accuracy that these legacy systems suffer still persist. Adding AI functions onto existing, rule-based automation might make them slightly faster, but in the end, it’s still the same old solution you’ve been struggling with. It’s vanilla ice cream wrapped in a Ben & Jerry’s New York Super Fudge Chunk® carton.

To scratch the surface and see what's underneath, you need to challenge the person cold calling into your work day to ask you about upgrading to an AI solution. The only way to be sure you're getting the most mature, secure AI for finance is to ask questions. Hard questions.

Like, “Can your solution manage both mundane and inhumanly complex tasks so my team can get home at a decent hour? Will it catch all the non-compliant behavior we couldn’t find so we no longer need our current solution’s massive team of offshore reviewers? Can you prove your claims of speed, accuracy, and security?”

There are so many pain points to solve, the questions themselves feel overwhelming.

If you're evaluating AI solutions, here's a helpful list of important questions. If you’d like some practice, give us a call, first. We’ve been doing this for a decade and we’ll gladly set the bar for the providers you talk to next. (Forewarning: It’ll be set high.) Then, print out this worksheet and call the other finance AI providers to compare answers.

We’ll still be here when you’re done being unimpressed.

Ask the right questions.

Here are some important questions finance executives should ask when assessing AI solutions. For a more in-depth exploration, check out our guide to evaluating GenAI for Finance.

1. Technology maturity

   How long have you been in the AI business and what kinds of AI technology are in your solutions? How are you incorporating generative AI?

2. Accuracy

   What is the source of your AI's training data? How do you ensure your products’ data privacy, security, and regulatory compliance?

3. Proof

   Can you give me a live demo of your AI's autonomous capabilities? Can you demonstrate its generative AI functions, too?

4. Core competencies

   Which processes in the [AP or expense audit] workflow still require human intervention or oversight? How and where is generative AI used in the workflow? What else can your AI do?

5. Scalability and customization

   Does your AI learn quickly and adjust to new formats, policies, and workflows? Can it rapidly scale up or down as our business needs change? How much can we customize?

6. Change management and support

   How much will IT need to be involved in the integration process? How long will it take to deploy and train my team? Do you have ongoing training and customer support available?

7. Customer testimonials, references, and metrics

   Can you share some customer references and testimonials? Can you prove their ROI with hard numbers such as hours or dollars saved?
   
   

AppZen has been building AI for enterprise finance teams for over a decade. We're not 5 college kids based in an apartment, or a legacy provider “bolting on” a new technology we barely understand. AI is in our DNA. If you're a leader searching the “ever-evolving landscape of finance” for a mature AI automation solution, it helps to know that AppZen is a trusted partner to 1/3 of the Fortune 500, with the knowledge and experience to help you unlock the true potential of finance AI for your organization.

Give us a call to see our AI in action. We're ready when you are.