Microsoft EVP: Embrace AI Agents to Rewire Business Processes Now

In September, Rajesh Jha the executive VP overseeing Microsoft’s AI-powered assistant Copilot appeared on Microsoft’s “Worklab” podcast, to discuss the history and future of technology. Loking back over the last 30 years, the “compression of innovation” in this wave of AI “is like nothing that we’ve seen before,” Jha said. What does he wish business leaders understood today about AI agents? “It’s not some distant future. It’s happening now…” Jha shared stories about how he’s using AI in his own executive-level job at Microsoft. He offered tips from other real-world implementations at emerging AI-native companies. And he even remembered the important management lessons he’s learned from watching top executives at Microsoft. How should other businesses be approaching the rise of AI today? Armed with a steady supply of real-world stories, Jha offered up his own insights. The kind of insights you can only get from running a $100-billion business with 30,000 employees… Embracing AI Agents: The Present Reality “It’s possible today to take full advantage of agents,” Jha emphasized. “The security model exists, the identity model exists, the user interface exists. The hard work here is to actually go pick the processes that give you the most bang for the buck — and then be rigorous about measuring that…” Fortunately, there’s a tool for that. “This is why we invested in something called the Copilot impact dashboard — so customers can take their core key performance indicators, and they can measure how Copilot is moving those key performance indicators. But it’s not just a theoretical benefit. Jha’s product management team researched 30,000 people across different organizations — in 30 different countries — as well as emerging “AI-native” companies. So where do they see AI making a difference? After getting workers a productivity boost, there’s a second place they customers getting a real return on investment, Jha says. It happens by reconfiguring high-cost (or high-value) business processes with agents “that can automate a bunch of those processes to be either more effective or more efficient. “That I think is changing the way work happens… I see customers actually taking a business process, and they are rewiring that for a world where people an AI can work together to automate that, to make it more effective and more efficient… “It’s not trying to solve 100 business processes, but taking a few and going really deep, and measuring the ROI and tweaking that. Because then the payoff is right there.” Leadership Lessons From Microsoft’s Cloud Journey And there’s a reason why executives have to lead this transformation, Jha says. “If a support organization, a marketing organization, an engineering organization is wired to work the old way, they are not going to automatically re-wire themselves for a world where AI can do a bunch of tasks, and people’s tasks change.” It requires leaders saying “how do I lead the way where I empower and I get to a world where AI is assisting, to agents and people working together.” Jha illustrates his point by remembering a moment 15 years ago, when Steve Ballmer asked him to bring Office to the cloud. At the time Microsoft was incredibly profitable, while “the cloud was a question mark for many at Microsoft.” Was the technology mature enough? Would the lower margins be worse than their on-premises business model? As a long-time server company, could Microsoft transform fast enough? “There was a lot of, ‘Let’s hold back — let’s see if this trend is really real or not.'” Steve Ballmer “showed incredible courage — by going all in.” Jha remembers it as “a very big moment for us,” describing Ballmer as “unambiguous” about cloud computing being a software-delivery route of the future with lots of benefits. “We are just going to go all in.” And specifically, “What Steve did was he gave license to people to go and learn, even if we were not perfect on Day One. The lesson? Be wary of hesitation — whether it’s cultural, or based on your skillsets or business models. “It’s very hard for the teams to rally behind something where the leaders themselves are half-hearted. The hard parts are business-model change, culture change, skill change– “They’re all incredibly hard. And that’s why there has to be a commitment from the top, that we’re going to see this through.” Jha agrees when podcast host Molly Wood adds that it also takes “persistence and belief… so you stay on that road, even when it gets hard.” But adds that it has to be tangible. “It’s one thing to say it. It’s another thing to allocate resources to that belief. We have a quote which is, if you really want to see the strategy of an organization, you’ve got see where they’re allocating resources.” Microsoft’s Pivotal Moment With GPT-4 Jha hammers home the point with another story: when Microsoft CEO Satya Nadella invited leaders of Microsoft’s three biggest products to see GPT-4, and specifically because “I’ve got to get you guys to be believers.” Watching a big-screen demo — at the home of Bill Gates — Jha was “completely blown away” by its ability to perform not just on multiple choice questions, but also when asked for longer, written answers. It even answered Bill Gates question about what would it say to the parent of a sick child. Jha called the experience “profound,” a realization that after 35 years, this heralded a change in “the fundamental interaction pattern between people and their devices” — that it’s now possible to express a user’s intent at the high level of a human-to-human interaction. (Where before the user was left to perform what’s essentially low-level clerical work.) And this moment required a new kind of response. “I have never in my 30 years… ever gone to my team and said ‘Drop all your plans. Let there be controlled chaos. We’re going to move 100 miles per hour.'” So what was that moment like? Jha says after the AI demo, “None of the existing plans mattered anymore. I huddled all my senior leaders at the top 50, 60 leaders, and I said, ‘Folks, I want you all to run 100 miles an hour. It’s going to be very uncomfortable, because we are going to unleash some amount of chaos. But let’s make sure we harden our processes, that this chaos does not make its way to customers.” AI Agents in Action: Modernizing Microsoft Workflows Jha points out that his typical emails have been composed after consulting his staff and leadership team, or reviewing a customer’s history. But “Now Copilot is doing that for me — it’s taking a look at my past emails, it reaches out to the customer service database, it tells me the latest status on this. It creates me a draft that I then go write, and I send it out.” And the Researcher agent built into Microsoft 365 Copilot works with your enterprise permissions and your enterprise data. Jha calls it “a real breakthrough” with “the makings of a digital employee,” like someone who joined a company and accessed its institutional knowledge to “be a part of producing output for the company. Jha remembered when his father was running a large steel plant in India, and visiting his office 20 years ago while his father’s secretary reviewed his morning messages. While he’d teased his father for being old-fashioned — since Jha carried his email on his phone — “Now I understand he was a smart guy — and I’m a digital clerk. I do all the clerical work myself — you know, I’m sorting messages, I’m replying to stuff… “That is what AI is now going to do. It’s going to take the clerical part from all of us, and will automate a lot of clerical parts to let the human ingenuity and the creativity, and really let us focus on the intent and the meaning of our work.” Jha recently tried Microsoft’s “Researcher” tool when crafting a six-month plan — feeding it a wide variety of data on the competitive landscape, customer feedback, and the team’s accumulated ideas. “It was incredible. It was able to get through my email and documents that I hadn’t fully read, but my team was iterating on. It looked at last year’s plans, it took a look at the competitive landscape and gave me a great eight-page document that I can now go and tweak and make it my own, overlay my perspective, and use as a starting point.” Jha’s team also built a “Know Your Customer” agent, which ingests support tickets and past communications and data on their current adoption levels. He says he’s “very excited” to use that, and often ends up sharing its output with intrigued customers… And he also uses AI a lot in his personal life — for planning trips or researching products. “I use it for a lot of that too.” So when asked for one actionable insight about AI, Jha says, “Go embrace agents. Pick out your most important processes, reimagine them — how agents and digital labor can re-wire that.” The post Microsoft EVP: Embrace AI Agents to Rewire Business Processes Now appeared first on The New Stack.