Accounting Today:Wiss and Co.等事务所重构客户AI咨询,应对成本上升
Tokenomics: Helping clients manage rising AI prices
Accounting Today采访多家会计师事务所发现,面对AI从订阅制转向按用量计费带来的成本上升,事务所正将自身成本控制经验引入客户咨询。Pigment调查显示,83%的财务负责人称相关支出超出预期,典型超支幅度为11–25%,17%称超支超过50%。

Елена Бутусова - stock.adobe.com
- Key Insight: Find out how CPA firms are restructuring their client advisory services around rising AI expenses.
- Expert Quote: "The trainings have gone from prompting and how to use these things to building governance" — Hrishikesh Pippadipally, CTO, Wiss and Co.
Accounting firms are not the only ones facing higher AI costs than before. Indeed, as more providers change from subscription-based to consumption-based enterprise pricing, many clients are experiencing the same kinds of challenges.
Processing Content
This has led to no small amount of consternation in the business world. For example, according to finance solutions provider Pigment, 83% of finance leaders said consumption-based AI costs have pushed spending above expectations, while just 13% have stayed within budget. The typical AI budget overrun was 11–25%. More than a third of respondents said they are more than 25% over and 17% said they are more than 50% over.
In response, CPA firms — having seen their own fair share of AI cost increases necessitating their own cost control measures — are taking what they've learned at their own organizations and bringing it to their clients. Leaders acknowledged that what they have been experiencing with AI costs has affected the advice they give to their clients.
Hrishikesh Pippadipally, chief technology officer for Top 100 Firm Wiss and Co., noted that his own firm has instituted user segregation, token monitoring, and efficiency training to control costs, and Wiss now trains their clients in implementing this governance model themselves, reflecting the growing maturity of the AI sector in general.
"The trainings have gone from prompting and how to use these things to building governance and cost monitoring from day one. Because you cannot retrofit that later. Everyone needs to build that guidance and cost monitoring … . That's the governance model you need to build," he said in an interview.
Charles Apigan, AI practice leader for Top 50 Firm LBMC, said his own firm's experience in tamping down AI costs adds to their credibility as trusted advisors to their clients. What he has noticed is that many clients are on the same page as they are when it comes to AI: more concerned about optimizing spend, they're starting to reel back some of their more extreme experimentation in favor of governed and controlled use for specific business functions.
"We become more valuable because we sell ourselves on being the trusted partner in AI, right? And so I have no choice if I say I'm going to be your trusted partner to give you the best advice possible. If using Claude to solve a problem is the right approach, I'll let you know. If it's to create your own workflow and that way we can pinpoint how to do every step of the process, and that'll be a lot cheaper, then I would rather that. If it's still better to have a human do it. I'll tell you that too," he said in an interview.
Cheryl Burke, the project management officer for Top 25 Firm PKF O'Connor Davies and a specialist in operations management, said that their AI advisory engagements always started with governance, noting that it's essential to gaining better visibility into the organization itself that will enable a more thoughtful strategic plan. Phil Comerford, the firm's CFO, added that this tends to involve questions about what is and is not allowed, how they set up teams to surface the best ideas, how they structure the compliance aspect, and more. Burke added that the degree to which cost control comes up in this conversation depends on the client, some don't think too much about it, and some think about it a lot.
"I think since a lot of AI is still evolving, when they ask about cost control, and we can talk about that a lot, each one approaches it differently. Some people are very strict, and they want to lock it down and give everyone a certain number of tokens, and then that's what it is, but the client needs to decide that," she said.
Travis Miskowitz, technology CAAS partner at Top 25 Firm Aprio, echoed the others in saying that clients have gotten over the initial novelty on AI and now want to talk about practical specifics. This sometimes involves price, but this is usually within the context of the larger business goals they want to accomplish with or without AI.
"The client conversation around AI is becoming much more practical. Our clients are more interested in where AI can save time and make a difference in the business. Our advice is to always start with the business problem. What are you trying to accomplish? Where are people spending time today? Where could technology help? AI can be incredibly useful; however, that doesn't mean it's the answer to every problem," he said in an email.
He added that, despite the changing environment, clients want what they have always wanted: someone who understands their business, who can connect the dots, and who can sit across the table from them and say, "Here's what the data actually means, and here's what I think we should be thinking about next."
"And ultimately, that's where the value is. Technology can help us work differently and give us better access to information. The advisor still brings the experience and judgment to understand what that information means for a particular client and what they should do with it," he said.
Technology Editor
Chris Gaetano is the technology editor for Accounting Today. He brings with him more than a decade of experience covering the accounting profession... Read full bio
来源:Accounting Today(美国会计行业媒体) · accountingtoday.com