AI agents in sales: what Indian organisations should expect in 2026

Higher targets. Less time. More demanding buyers. That is the sales reality many Indian teams are walking into in 2026. Salesforce’s latest State of Sales research positions AI agents as the lever teams are reaching for. In India, 91% of sales professionals surveyed say AI agents are essential to business success, and AI is now being used widely for day-to-day work such as prospecting, forecasting, lead scoring, and drafting emails.

The report also points to the real bottleneck: not effort, but administrative friction and messy, disconnected data. Even as sellers spend close to a day a week prospecting, many say they still lack the time to do adequate outreach. The findings are based on a double-anonymous survey of 4,050 sales professionals globally, including 250 in India, conducted in August–September 2025. In this interview with Dataquest, Sangeeta Giri, SVP and Chief Operating Officer, Salesforce South Asia, explains what agent-led selling looks like in practice, and what foundations sales leaders must fix first.

The State of Sales research puts AI adoption front and centre this year. What was the core intent behind the study, and what did you set out to understand?

We wanted to understand how sales teams are adopting AI, what use cases are showing up in a meaningful way, what benefits teams are actually seeing, and how all of this can improve the life of a salesperson. We ran the research across 20-plus countries and spoke to a little over 4,000 people globally.

We also ensured the study reflected the reality of selling. Sales is not only the seller in the field. It is a system made up of multiple roles. So we took inputs across sales managers, field sellers, partner sales teams, SDR and BDR teams who qualify and build pipeline, and operations and enablement teams that support execution. We also looked at a slice across performance levels, because real-world teams have high performers, mid performers, and underperformers, and that mix tells you where AI is actually helping and where teams are still getting slowed down.

The report suggests that administrative friction is a bigger drag than effort or skill. When you looked at sales work in India, what did you find about how sellers actually spend their week?

When we studied how time is spent, we realised we were not getting enough customer-facing time. We did an internal exercise in India as well. We brought 100-plus managers together and asked: how are you and your teams breaking up time?

We looked at calendars and broke it down across customer meetings, prospecting, training, and internal work that goes into getting ready for a customer meeting. What we realised was that more than three to three-and-a-half days could go into internal work and preparation versus actually sitting face to face with customers. That became the problem statement: how do we flip non-selling time and selling time so we improve customer-facing time meaningfully, without simply adding more people?

Where did AI agents start making the first real difference once you began introducing them into your sales workflows?

The first gains came from reducing the work around planning, entering data, creating notes, prospecting preparation, and a lot of the internal work that sits around selling. As we started using AI more consistently, another point became clear: data accuracy becomes very important because sellers can only move faster if the information they see is reliable.

As we introduced AI, we realised that pulling information intelligently from different sources and presenting it in a usable way improves both internal planning meetings and customer conversations. The time that earlier went into searching, switching between tools, and building context manually starts coming down.

You described a pre-meeting workflow using a Slackbot. What does that agent-assisted planning look like in practice for a seller?

The idea is to give the seller a ready-to-use briefing before the meeting. A seller can go into Slack and ask: I’m meeting this customer, can you give me everything I need?

The agent can pull information from the sources available and share back a background on the customer: are they a prospect, have they shown interest earlier, do they already use Salesforce, if they do then how do they use it, and are there any past interactions or follow-ups that were missed.

It also gives broader customer context: what industry they are in, how big they are, what they are trying to achieve, how they position themselves, and what they are competing against.

Then it connects that to what the seller needs to sell. It can provide information on the customer’s technology stack, based on publicly available information and what has been collected in earlier interactions, and it can point to gaps or areas where there is a propensity to sell. It can also suggest industry use cases and show comparable customer examples or win stories that can help the seller frame a credible conversation.

The point is not to overload the seller with information. The seller can ask for three actionable insights and get that quickly. Something that could have taken one or two hours earlier can now happen in minutes, including while walking from the reception to the customer’s door.

The India data point is striking: 91% of Indian sales professionals say AI agents are essential to driving business success. When teams say “agents” today, what are they actually using, given maturity varies?

It varies widely, and that is worth acknowledging. For some, it starts with quick assistance to save time: helping prepare for a customer meeting, suggesting questions to ask, summarising information, or drafting communication in a contextual way.

For others, it is deeper workflow automation that impacts pipeline. For example, when interest spikes around a topic, the volume of inbound leads can become very high. But teams have a finite set of people, and some leads get dropped for lack of qualification or lack of time. Agents can re-engage those leads at scale with contextual messages, and they can run multiple rounds of engagement. As the prospect moves from curiosity to real interest, the agent can then set up a meeting based on calendar availability so that humans step in at the right time.

This is where agents become useful. They are not replacing the seller’s judgement. They are removing the time cost of searching, chasing, and formatting, and making follow-through more consistent.

The report also says 90% of Indian sales organisations are already using AI for tasks like prospecting, forecasting, lead scoring, or drafting emails. In your view, what separates “AI usage” from “AI-led execution”?

Using AI occasionally is different from embedding it into daily workflow. AI-led execution shows up when it changes how sellers operate: how they plan accounts, how they prioritise pipeline, how they engage prospects, and how quickly they move from insight to action.

If sellers still have to jump across dashboards and tools to build context, then the system has not changed enough. When agents are able to do that research and bring back a ready docket, the seller can focus on the conversation and the outcome. You start seeing better quality outreach, better prioritisation, and a higher quality of customer engagement because the seller is not spending hours piecing together information.

From what you are hearing from customers, how are they thinking about using agents for their own sales teams, especially around sales enablement?

Sales enablement and sales execution are converging. Sellers still need to learn products, industry context, and use cases. That learning cannot be skipped. But the way enablement scales can change.

Internally, we do a lot of ‘stand and deliver’, where sellers learn and then pitch back so we can grade them. In the past, that needed managers and multiple people in a room. At scale, that becomes hard to do consistently. With agents, parts of that pitch review can be done faster. Agents can recognise key points we have defined as important, grade the pitch against those guardrails, and provide feedback on what was missed. That makes learning more bite-sized, faster, and more frequent.

From a customer perspective, the same productivity logic applies. The market opportunity is larger than what any team can cover. They can’t keep adding more sellers, so they need to make sellers more productive by taking away mundane work and letting humans spend more time where judgement, relationship-building, and negotiation matter.

You also talk about prospecting pressure: cold calling is widely disliked, yet pipeline growth needs more outreach than teams can deliver. How do agents change prospecting without turning it into automation spree?

Prospecting needs both volume and relevance. The risk with any automation is that it becomes generic. What changes with agents is that the outreach can be more contextual because the agent can pull relevant background, past interactions, and specific signals, and then draft communication that reflects that context.

Agents can also help carry follow-ups consistently across multiple touches, which is hard to do manually at scale. But the handoff matters. When the prospect shows genuine interest, that is when humans need to step in. That is the balance: agents handle the repetitive work and continuity, humans handle the conversation where nuance and judgement matter.

There is also anxiety in teams about agents replacing roles. How have you addressed that internally, and what do you advise leaders to do to reduce that fear?

Initially, it was fear of the unknown. When people don’t know what agents can do, they worry about what it might mean.

What helped us was making adoption visible and peer-led. We asked our teams to share how they are using AI, and we received 100-plus short videos. As more examples came out, the variety of use cases became clear. People learned from each other, and it shifted from a top-down message to a peer-driven adoption.

There is also a very practical point. A lot of work sellers have to do is not glamorous: data cleaning, removing duplicates, correcting errors, filling omissions, standardising formats. When AI takes that burden off the seller and makes data hygiene easier, the value becomes immediate. The mindset changes from fear to leverage, because sellers can see how it makes them more productive.

A final point. The report highlights that disconnected systems and messy data can slow AI initiatives. What foundations should Indian sales leaders prioritise if they want agents to work reliably?

Data quality and connected systems become foundational. If systems are disconnected and data is inconsistent, agents will not deliver consistently. Hygiene becomes critical, and it is one of the areas teams must prioritise if they want ROI from AI.

Leaders also need to simplify where possible. Tool sprawl creates trapped data and slows down execution. The focus should be on trusted, connected data, clear guardrails for how agents are used, and a workflow design that brings insights into the seller’s flow of work rather than forcing sellers to chase information across multiple systems.

That is what makes agent-led execution reliable: clean data, connected context, and a human-and-agent partnership designed around outcomes, not novelty.