Advisory services platform Fieldguide announced the release of its new Fieldguide AI solution, which aims to automate the repetitive low-value parts of advisory work so the accountant can focus on more meaningful tasks.
Specifically, the new software can:
- Draft control tests based on control activities: Fieldguide's AI suggests the appropriate control tests to perform based on the control activities identified in an engagement.
- Draft requests based on control activities: Fieldguide's AI suggests the appropriate requests for documentation and evidence based on the control activity identified in an engagement.
- Assess a document's adequacy for addressing a control: Fieldguide's AI analyzes documents related to an engagement and suggests whether they meet requirements for a control activity.
- Suggest improvements to a document: Fieldguide's AI analyzes documents related to an engagement and suggests recommendations to better implement the control activity.
Fieldguide CEO Jin Chang, in an interview, emphasized the privacy aspects of its new AI solution. He said the program does not use user-submitted data to train or improve AI models, and overall the information is kept secure and private. This gets around the ethical issue of using certain AI programs, like ChatGPT, which store all query data in a company server.
"Theoretically, AI should be learning from the activities and feeding back to the large language model, but because we have considered all the enterprise security and privacy considerations, our AI does not feed back to the general large language model; it simply learns within the context of the engagement. We paid a lot of attention to the security considerations, because we know CPA firms are handling a lot of sensitive data, and we want to make sure that we're enhancing that, if anything," said Chang.
Chang said he envisions the Fieldguide AI not as a replacement for the human accountant but as a force multiplier, comparing it more to a co-pilot in an airliner who assists the human pilot. While the co-pilot may handle many mundane tasks, the human ultimately remains in the loop.
"The other thing I like to talk about is I think about AI almost as the copilot to the profession. In an airplane, the co-pilot is going to handle certain key parts of the cockpit command; similarly, AI is going to be like a complement to what practitioners do, and it really helps practitioners focus on what they're special in — advising clients, bringing their expertise to the table, and just spending a lot of face time with the client on really valuable activities," he said.
In this respect, it is not so much about replacing humans with machines but, rather, having single humans use machines to do the work of several other humans at once. Chang said this is with an eye toward addressing the CPA pipeline issue, where the number of people entering the profession has been tapering off.
"I'm so excited about the marketwide impact that technologies like Fieldguide and our AI capabilities can bring to the space. Everyone in the industry is talking about the talent pipeline, and how it's simply not going to keep up with the growth of the overall market. So what that means is that CPA firms need to find ways to produce more per practitioner, whether they are outsourced internationally or onshore. We believe that generative AI and current AI technologies can be that step-change in technology that can help the entire industry be more productive on a per-practitioner basis," he said.
