Microsoft’s New Sales Pitch for AI: Spend Less Money on Humans
Microsoft has a new pitch to persuade businesses to spend money on artificial intelligence–powered services: If you use them, you won’t need as many employees.
The potential of AI to replace human workers is an old idea, but one most companies have avoided bringing up explicitly, for fear of suffering reputational harm and political attacks. But as tech companies try to overcome customer uncertainty about the value of AI, they’re becoming more direct about that possible benefit.
The Takeaway
• CEOs are asking for assurances that buying AI will save their companies money
• Until now, AI sellers have shied away from promoting software as a labor replacement
• Google changed a customer quote to remove reference to savings on labor cost
CEOs and chief financial officers need assurances that adopting AI will lead to labor cost savings or reduce payments to external contractors, said Jared Spataro, chief marketing officer for Microsoft’s 365 Copilot software.
“What CFOs are rightly asking for is, ‘Show me what you took out of our budget by using AI,’” Spataro said. “We’re hearing that loud and clear.”
Microsoft is sharpening its pitch as its customers consider moving from experimenting with conversational AI that automates writing, coding, analytics and other tasks to actually using it at scale. Until now, many large companies have resisted expanding budgets for Microsoft’s Copilot AI until they have clear evidence of savings in other areas.
Other businesses are paying for such AI only if they can reduce spending on other enterprise or cloud software. Such expectations have prompted Microsoft to give a more candid sales pitch, Spataro said.
One promising group of customers for Microsoft is corporate marketers. Finastra, a British fintech company that generates $2 billion in revenue annually, recently used Microsoft’s 365 Copilot to handle marketing work it typically would hire external agencies to do, CMO Joerg Klueckmann said.
For a recent ad campaign, Finastra marketing staff interviewed 50 financial sector executives regarding what they found useful about its product and used Copilot to transcribe the interviews. Copilot then parsed 4,000 pages of transcribed interviews to generate ad copy featuring the executives’ testimony.
Klueckmann said that in the past, Finastra hired outside creative agencies and market research firms to create such campaigns, which would typically cost more than $60,000 and take several months. Instead, Finastra is paying for a handful of marketing employees to use Copilot to do the job, for a fee of $30 per employee per month. (It’s also paying for hundreds of other employees to have access to the Copilot tools for other purposes.)
Klueckmann says he doesn’t anticipate this type of technology would prompt him to reduce his more than 60-person marketing team, however.
“I think we’re going to produce more stuff with the people we have,” he said.
Office Automation
Despite the limitations of today’s conversational AI, which still makes mistakes, Microsoft can point to a variety of companies that are saving money from the technology. Some companies, for instance, have started using generative AI to replace roles like customer support or junior salespeople.
Spataro said corporate legal departments are using Microsoft’s 365 Copilot and other AI applications to avoid hiring outside lawyers, though he didn’t cite specific examples.
Broadcom, the semiconductor giant, has been scrutinizing the economics of its AI purchases as it automates IT help desk support for employees, said Stanley Toh, who oversees the department. By using software from AI startup Moveworks, which uses large language models from OpenAI and other providers, Broadcom has kept its IT support staff below 20 people over the past five years, even as the company’s head count grew from 10,000 to more than 50,000.
Toh’s team spends less than $2 million annually on AI—far less than the roughly $8 million annually it would be spending if its IT staff had grown in proportion to its overall staff.
“When I need to do contract renewal, this is the [return on investment] that I show to our executives to justify why we’re spending on this AI software,” he said. “The pure motivation is efficiency and cost avoidance.”
Until now, software executives have gone out of their way to allay concerns that AI would replace humans. In January, CEO Satya Nadella downplayed the idea, saying AI would change “the shape of these jobs” while creating more jobs by making tasks like software development easier for untrained workers.
Microsoft rivals such as Google and Amazon have similarly steered clear of saying customers can use their AI to reduce or avoid growing their workforces—even when one of their own customers says that’s what it’s doing.
Erasing ‘Labor Cost’
For instance, in August, Mirko Bibic, CEO of telecom firm Bell Canada, said during a conference call with analysts that Google AI tools had helped automate Bell’s customer service call center by using chatbots to handle some customer requests, as well as generating written responses that Bell’s human customer service agents could then send to customers.
As a result, Bell took “costs out of our business, which contributed to $20 million in labor cost savings,” Bibic said.
Notably, a Google blog post touting the same example amended Merko’s quote to remove the phrase “labor cost” so it just says “$20 million in savings.”
A Google spokesperson declined to comment. A Bell Canada spokesperson said some of the $20 million in labor cost savings came from the Google AI software, while some came from changes to how the company delivers fiber and changes to its billing and ordering system, but that person didn’t provide a breakdown.
Sellers of AI have long shied away from publicly acknowledging their technology could lead to fewer jobs because of the political heat that could bring.
Typically, “these companies won’t say there’s job replacement because it looks bad for their reputation and image, and they don’t want to spark an outcry,” said Lisette Espín Noboa, a researcher at the Complexity Science Hub, a Vienna-based tech policy think tank.
In its new pitch to customers, though, Microsoft is also touting its own savings on labor costs thanks to AI. The company says that using technology from OpenAI, it built internal tools that help salespeople automatically generate customer lists and perform cold outreach, and it separately built a tool to help customer support agents send automatically generated responses to customer questions, allowing them to handle more cases.
Microsoft executives say these tools are a reason the company has maintained its sales growth rate despite laying off 10,000 people last year, and the executives have been encouraging salespeople to use that example to entice customers, according to three current sales employees.
“We’ve been able to improve our throughput per [customer service] agent by 12% using Copilot. That’s real money. That means we don’t have to hire as many people,” Microsoft’s Spataro said.
That example might be overstating AI’s impact, however, because there’s also been a broader software spending recovery over the past year.
