Ghosh, who spoke exclusively to TOI, his first media interaction since he was appointed to the management committee in July this year, said change was becoming critical in today's high-velocity, software-driven world. "We are telling clients, if you want to be a disrupter in the market, you have to change to move fast and create value," he said.
Three changes are seen to be vital. The first is in the way software code is written. Traditionally, monolithic applications have often been built from the ground up. These are slow to implement — sometimes taking years — and slow to change.
"What's needed today is a more liquid way to write applications, write them in a modular structure, with reusable components — sourced internally and externally. This allows for the rapid assembly of applications in support of dynamic business needs," said Ghosh, who was with Infosys before joining Accenture in 2003. This, he said, also requires an open architecture since different components have to be quickly linked together.
Accenture notes that when firms like OpenTable, the online restaurant reservation system, and Uber want to add new application functionality such as messaging customers when their table or car is ready, they don't build it themselves. Instead, they tap into cloud providers like Twilio, which offer the functionality as part of a prebuilt platform. In turn, both OpenTable and Uber share their application components with other developers from firms as varied as Starbucks and TripAdvisor.
Ghosh said companies building liquid applications should do so with a cloud-first, mobile-first mindset. Applications need to be engineered to operate and scale in the cloud, since that may become inevitable even if the initial deployment is on premise. They must also be engineered for mobile devices since that's how customers and employees increasingly interact with software.
Ghosh's second key recommendation is to infuse intelligence in all applications. This is made possible by advances in data science such as natural language processing, machine learning and cognitive computing. So now, applications can automate routine tasks, can do integrated analytics (using data across the organization and outside it), and also self-learn and, with that knowledge, self-heal when a problem occurs.
The third recommendation is to allow business partner and customer ecosystems, as well as the internet-of-things environment, to easily connect to your applications. This would allow the community to quickly build value-adding solutions on top of your platform.
John Deere, manufacturer of tractors and other industrial equipment, introduced an open, online platform called MyJohnDeere in 2012 to help agricultural producers manage all the data related to their equipment and operations. DuPont Pioneer, for instance, delivers near real-time, field-level data via this platform, and this data helps farmers and dealers make important revenue-generating decisions about seed, fertilizer and purchasing. The platform has thus enabled John Deere to expand from being a provider of agricultural equipment to a provider of data-based agricultural services.