AI agents, not chatbots, will define enterprise tech’s next decade: Salesforce India VP

The conversation around enterprise artificial intelligence is rapidly moving beyond chatbots to autonomous, “agentic systems” capable of executing complex tasks with minimal human intervention, a shift that is already beginning to reshape hiring, workflows and decision-making inside companies.

That transition is visible within Salesforce itself. CEO Marc Benioff recently pointed to how AI-powered coding tools are boosting productivity to the point of reducing hiring needs.

“I’m not hiring more engineers in FY26 because I was using coding agents and I was allowing the productivity from the coding agent to give me the extra capacity that I needed for the year,” Benioff said in a recent interview.

His comments underscore a broader shift underway across enterprises: AI is no longer just assisting workers; it is increasingly doing the work.

For Salesforce, this evolution marks the beginning of what it calls Enterprise General Intelligence (EGI). In an exclusive interaction with Business Today, Deepak Pargaonkar, Vice President of Solution Engineering at Salesforce India, described EGI not as a single model, but as a system-level capability designed to operationalise AI across business workflows.

“It requires an operating system,” Pargaonkar said. “One that gives agents governed data and context, encodes the business logic that already runs the enterprise, and provides the observability to understand what agents are doing and why.”

Salesforce’s approach is built around what it calls an “agentic enterprise” stack, structured across four layers: context, work, agency and engagement. The goal, Pargaonkar explained, is to “convert raw intelligence into real work, and to move beyond individual agency to enterprise agency, with thousands of agents and humans orchestrating complex tasks across teams, functions, and companies.”

From pilots to production

The shift from experimentation to scaled deployment is already underway. According to Salesforce’s CIO study, AI implementation surged 282% in 2025, signalling that enterprises are increasingly embedding AI into core operations rather than treating it as a side project.

India is at the forefront of this transition. Salesforce’s State of Sales data shows that 91% of Indian sales professionals view AI agents as critical to business success.

“These aren’t experiments anymore, they’re augmenting human judgment in real workflows,” Pargaonkar said.

However, companies that remain stuck in pilot mode are often grappling with foundational challenges, particularly around data readiness and governance.

“AI agents are only as intelligent as the unified, real-time information they draw from, and leaders who understand that are moving faster,” he added.

Why are some sectors ahead

Adoption is not uniform. Financial services, healthcare and manufacturing are leading the shift toward agentic systems, largely because of how their operations are structured.

“Financial services firms have mature data infrastructure, regulatory requirements to document decisions, and workflows — compliance checks, loan approvals, claims processing — that map naturally to agent-driven automation,” Pargaonkar said.

“When your processes are already governed and auditable, deploying agents is an extension of existing discipline, not a leap of faith.”

In healthcare, the push is driven by the sheer scale of administrative workload. Tasks such as documentation and coordination are “exactly the kind of high-volume, rule-bound work agents handle well,” he noted.

The deployment challenge

Despite growing momentum, deploying enterprise AI at scale remains complex. Customisation, in particular, continues to be a major bottleneck.

Salesforce is attempting to address this through its Agentforce ecosystem, which includes prebuilt agents, templates and industry-specific solutions. Its AgentExchange marketplace already hosts nearly 800 reusable agent assets from over 160 partners, aimed at reducing deployment time.

Still, Pargaonkar is clear that customisation cannot be eliminated.

“Customisation is unavoidable and necessary,” he said. “Where customers invest time is in the areas that drive the most value: equipping agents with the right business knowledge, configuring them for specific workflows, and ensuring data readiness.”

Ultimately, he argues, data quality remains the defining factor in performance.

“A well-deployed agent on clean, governed data will outperform a sophisticated one on fragmented data almost every time.”

Rethinking the talent pyramid

As enterprises scale agentic systems, the implications for jobs are becoming harder to ignore. Benioff’s remarks on hiring reflect a wider recalibration underway across the tech industry.

Pargaonkar, however, sees this as a transformation rather than a reduction.

“The talent pyramid is transforming,” he said, adding that the shift “creates more opportunity than it displaces.”

The real challenge, he argues, lies in how organisations redesign work. Employees will need to be reskilled to manage and collaborate with AI agents, while human effort shifts toward higher-value functions.

“India has an extraordinary talent base. The opportunity is to redirect that talent toward the skills that agents can't replicate — judgment, creativity, and the ability to orchestrate human-AI collaboration at scale,” he said.

As enterprise AI moves from chat interfaces to autonomous execution, the companies that succeed may not be those with the most powerful models, but those that can integrate agents seamlessly into their workflows.