I asked Beerud Sheth a question that sounds simple until you apply it to your own campaigns: what changes when a marketer stops thinking about messages and starts thinking about conversations?
Beerud is the Co-Founder and CEO of Gupshup, a global conversational AI platform for marketing, commerce, and customer support. An IIT Bombay and MIT alumnus, he previously co-founded Elance, which later became Upwork. He has built Gupshup from a messaging platform into a global conversational AI company.
But what made our conversation on Dilse Omni Talks valuable was not just his understanding of technology. It was how clearly he connected changing consumer behavior to the everyday decisions marketers make.
Table of Contents
1. Every Message Has to Earn Its Place
2. A Reply Changes the Nature of the Interaction
3. The Campaign Metric We Too Often Ignore
4. The Neighborhood Shopkeeper Is the Best Explanation
5. Marketing Cannot Generate Interest and Walk Away
6. AI Agents Could Make “Offline” an Incomplete Description
The Shift I Would Make Before the Next Campaign
1. Every Message Has to Earn Its Place
Beerud began with the economics. He pointed to rising WhatsApp messaging costs and anticipated RCS rate increases in some markets, including India. His larger point was straightforward: when messages become more expensive, indiscriminate volume becomes harder to justify.
A brand cannot keep sending more and assume that enough people will eventually respond.
There is also a consumer cost. Irrelevant messages lead people to block brands, opt out, or simply stop paying attention. Beerud connected those reactions to a broader pressure on messaging: higher prices, weaker delivery, or declining attention can all make the old approach less effective.
What stayed with me was the responsibility this puts back on the marketer. Before sending, we need to think carefully about who should receive the message, what it should say, when it should arrive, and how we will respond.
Beerud’s Five Filters for Better Messaging
Beerud described five qualities that messaging now needs. I would turn them into a practical campaign checkpoint:
Deliberate: Be clear about why this message needs to exist.
Targeted: Decide who genuinely needs it, rather than defaulting to the whole database.
Precise: Make the communication relevant to that customer’s situation.
Thoughtful: Consider timing and the questions the message might create.
Analyzed: Examine the response instead of treating delivery as completion.
These are not five ways to polish a broadcast. Together, they push us toward a more responsive relationship with the customer.
2. A Reply Changes the Nature of the Interaction
The distinction between messaging and conversation is not whether a brand uses WhatsApp, RCS, or another interface. It is whether communication remains one-way.
In a broadcast, the brand decides what to say and pushes it out. In a conversation, the customer contributes intent. They ask for something, clarify a requirement, or raise a concern.
Beerud explained that a customer’s question creates an invitation to respond. The brand is no longer merely trying to interrupt someone. It has an opportunity to be useful.
He also pointed out that these exchanges can be both more effective and lower-cost. I took that as a reason to rethink the interaction, not as a universal pricing promise across channels and message categories.
“The era of broadcast marketing is over.”
Beerud Sheth
That became the starting point for a masterclass in conversational marketing. Not better campaign copy alone. Not another channel. A different way of thinking about customer engagement, sales, and the experience that connects them.
My immediate reaction was that a person messaging back is often showing stronger intent. They are giving us something a click alone may not reveal: what they need to know before moving forward.
This distinction sits at the heart of how conversational commerce changes customer engagement. The response is not an interruption to the campaign. It is where the next useful interaction begins.
3. The Campaign Metric We Too Often Ignore
As Beerud spoke, I started thinking about our measurement habits. We discuss open rates and click rates constantly. How often do we give reply rates the same attention?
Especially in mid-ticket and high-ticket purchases, customers have questions. Are there more colors? Can you show me other designs? What is the quality like?
Those questions are part of selling. Yet our campaign planning often ends before we prepare for them.
I admitted that I have made this mistake myself. Most of my energy has gone into the copy, the creative, and getting the campaign out. Much less has gone into anticipating what the customer might ask next.
The uncomfortable question is: have we designed a campaign to generate a click, or a conversation that can lead somewhere?
My “Next Five Questions” Framework
This is a working framework I took from our discussion, rather than a named model Beerud presented:
Relevant message → Customer reply → Clarifying questions → Useful recommendation → Appropriate next step
Start with relevance: Why would this customer care about the message?
Anticipate five questions: What might they ask about price, options, design, or quality?
Prepare useful responses: Help the customer evaluate, rather than repeat the promotion.
Plan the handoff: Decide when an AI agent, voice agent, person, or store should take over.
Review progress: Look beyond opens and clicks to replies and movement toward a sale.
The number five is simply a useful planning discipline. It forces us to imagine the customer’s side of the exchange before pressing send.
4. The Neighborhood Shopkeeper Is the Best Explanation
Beerud then used an analogy that made the whole idea feel familiar.
Think about a large store such as Walmart or Costco. There are aisles full of products. You walk around, compare options, and work out what suits you. Much of ecommerce follows that structure: rows and columns of products, with the customer doing the searching.
Now think about the neighborhood shop.
The shopkeeper asks what you need, shows three or four options, listens to your reaction, and returns with a more suitable selection. The experience evolves through the exchange.
That is conversational commerce. Not just access to inventory, but help choosing.
Beerud’s point was that this interaction is curated, personal, and responsive. The customer does not have to navigate everything alone.
It also explains why premium service feels premium. Someone knows your name, understands why you are there, and responds to your requirements. The value comes from being understood, not simply being shown more information.
My takeaway was that conversational AI should be considered through this service lens. Can it make the experience more helpful, rather than merely automate the delivery of more messages?
5. Marketing Cannot Generate Interest and Walk Away
Beerud asked me how the old marketing mindset differs from what comes next. I thought about the traditional separation between marketing and sales.
A newspaper or television advertisement generated interest. Sales then took over. That thinking carried into messaging: send the Rakhi campaign, the Diwali offer, or the Easter promotion, generate footfall, and move to the next campaign.
But a conversation narrows that distance.
If a brand announces a new iPhone and the customer immediately asks about price, the next response is already part of a sales conversation.
“You cannot just generate interest and leave it.”
Saurabh Agrawal
Beerud agreed and took the point further. Brands divide the journey into awareness, interest, decision, action, sales, and support. The customer experiences one brand.
The neighborhood shopkeeper does not stop being responsible after creating interest. The same person recommends, sells, cross-sells, and helps when something goes wrong.
That is the continuity businesses need to recreate. Marketing can still have specialist teams and responsibilities. But the customer should not experience those boundaries as a dead end.
As marketers become more accountable for sales, revenue, attribution, and ROI, this shift matters even more. The work cannot end with proving that somebody clicked.
6. AI Agents Could Make “Offline” an Incomplete Description
The final part of our conversation widened the discussion beyond messaging campaigns.
Beerud challenged the assumption that online marketing and offline footfall must remain separate experiences. Customers carry connected devices into stores. Their physical presence does not stop their purchase decision from being digitally assisted.
He described a future in which shoppers could ask an always-available AI agent about products, brands, deals, and offers rather than queue for a small number of assistants.
Importantly, that agent may belong to the brand, or it may be the customer’s own personal agent. Beerud expects agents to become increasingly available through phones, glasses, and other interfaces.
That was his forward-looking argument: commerce becomes agent-mediated, whether the customer is physically in a store or shopping remotely.
The question for brands is whether they participate in and support that interaction. Our broader discussion of AI in omnichannel marketing explores the same direction: marketing, selling, and supporting need to feel connected.
The Shift I Would Make Before the Next Campaign
I came away with a simple test. Before launching another campaign, ask: if a customer replies, are we ready to help?
That question changes the brief. It makes us plan the response, anticipate uncertainty, connect the handoff, and measure more than attention. Technology matters, but the starting point is a mindset that stays with the customer beyond the first message.
Beerud’s lesson was not that brands should stop communicating. It was that communication must become more deliberate, useful, and responsive. The opportunity is not simply to send better messages. It is to build better conversations.
I am Saurabh Agrawal and we come with a new episode on Dilse omni talks every fortnight and cover different aspect of omnichannel with amazing speakers.


















