Move past disruptive innovation's hype 

In the case of generative AI, personalisation means cultivating solutions that serve real and quantifiable needs in the market. “Generative AI is currently good for three things: understanding intent, synthesising knowledge, and producing natural-language responses,” Mathur says. 

“That means we first need to recognise both the breadth and limits of those capabilities: Generative AI is not some kind of panacea for every pain point. Then, we can ask ourselves, where might those capabilities be best suited to solve existing problems? And whose problems exactly are we solving?”

Mathur and his team at the India Technology Center are addressing people issues via the same user base they already use to guide all their work: system admins, end users, service agents, and so on. IT personas might sound about as staid as disruptive innovation sounds glamorous. But relying on existing personas, and the real-world customer input that they represent, ensures technology investments deliver meaningful business returns.

“The problems we want to solve with generative AI are already well validated by people who fit those personas,” Mathur says of the India Technology Center’s current explorations. “You’re not aiming to come up with one universal solution, but instead finding ways to lower the bar for different groups to do more strategic and creative work.”

As an example, Mathur points to service agents who face a constant tug of war between productivity and personalised service to their customers. “You cannot humanly parse huge unstructured case histories while also constantly shortening time to resolution,” Mathur says. 

“A generative AI solution, however, can synthesise knowledge and produce natural-language responses based on that synthesis. For your typical service agent, that could look like turning those huge case histories into a single paragraph of common-sense text with just a click. And that gets you to the best-of-both-worlds state that these agents have sought for years.”

If this kind of persona-based approach to generative AI can deliver such obvious value, why aren’t more companies adopting it? “Persona-based development is both easy and hard,” Mathur admits. “It’s easy because the business case makes obvious sense: You are using generative AI, or whatever new technology you deal with, to tap into existing demand amongst existing customers. 

“But it’s hard because building solutions for each specific persona, integrating them into those persona’s existing workflows, just involves a whole lot of engineering time and labour compared to a one-size-fits-all approach. Generating this kind of value takes heavy work, and there are no shortcuts.”

That prompts another question: If building best-in-class AI comes at such a cost, why not buy it instead?

The high failure rate of acquisitions—as high as 70% to 90%, according to some—has yet to dissuade the technology industry, with more than 45% of dealmakers expecting tech M&A activity to rise this year. Caveat emptor, Mathur cautions: A wrongly purposed acquisition can cost technology providers more than it’s worth.

“You have to ask: Are you acquiring the company for its technology or simply its accretive revenue?” Mathur says. “For us [at ServiceNow], the purpose of any acquisition must be that the technology can help us better serve our customers. But for those acquisitions to deliver real customer value, [it] takes substantial amounts of commitment and sacrifice.”

As one pertinent example, Mathur points to ServiceNow’s acquisition of G2K and its AI platform, which draws real-time data from real-world sources like video cameras and internet of things devices. “G2K’s technology is entirely embedded in the physical world, which gives us the chance to extend the ServiceNow platform out of the cloud and into on-premises environments like retail outlets or warehouses,” Mathur says. “At the same time, dealing with video and internet of things data is something new to our platform. 

“Our job now is to incorporate this new technology into our platform, in a way that proves frictionless and fruitful for customers. That takes a lot of careful planning, testing and, most of all, time.”

Mathur uses the term “replatforming” to describe the process: an end-to-end remediation of the acquired party’s technology to fit naturally into the acquirer’s technology stack. It’s an approach that can take months of intensive redevelopment of the new technology and existing platform alike—during which the acquired software and applications aren’t being sold. “There is definitely a significant revenue cost to doing this right,” Mathur says. “But we’ve seen all sorts of historical examples where a company simply bolts on the new technology in a hurry to sell more and it ends up degrading performance and harming the overall customer experience. 

“The sales hiatus ensures we only present our customers with something they will find useful and valuable, rather than causing them to justifiably question why we made the acquisition in the first place.”

Mathur’s ethos appears to be increasingly shared by the broader market. Recent studies found that 29% of firms now start planning long-term operating models for their acquisitions early on during deal screening, compared to just 1% that did so in 2019. Those operating models extend well beyond simply technology.

“Any sensible acquisition for us brings both technology and staff, the latter of whom often hail from very different cultures and organisational structures to your own,” Mathur observes. “Even as you’re working to get the technology seamlessly integrated, you’re also having to learn about and understand a whole new world of people.” Replatforming culture, I suggest, and Mathur looks pleased at the idea.

“Replatforming software or culture, aligning disruptive technology to what customers tell you they need—none of this is easy,” he concludes. 

“It takes conviction and sacrifice that rarely make the headlines. But doing the job well is worth it—for our businesses, our customers, and ourselves.” In other words, stick true to fundamentals of creating value and serving others, even—or especially—when it stings. The hype cycle itself is just part of business as usual, after all.”