ARTICLE AD BOX

Summary
We may be hurtling into an era of autonomous agents, but there are important reasons not to sideline the chat window.
Like many frequent users of chat assistants, I’ve found my equation with them. We’ve got to know each other. They seem to have figured out the way I work, and though they can be annoying, they try to avoid stepping on my toes. But just as I’ve become accustomed to conversation with AI, the industry wants to move on.
Apparently, the era of talking to chatbots is already over. Did companies spend vast quantities of money and effort fine-tuning conversation, only to hand over the reins to agents and tell us we don’t need to say a thing?
Chatbots don’t sound human by accident. Lots of work has gone into shaping that tone. Systems are trained to detect mood from text, infer context in a conversation and even respond with something resembling empathy or understanding. The goal was simple: Make interaction feel natural enough that users would trust it and return to it.
Personas were introduced to widen that appeal. Grok, for instance, experimented with a bouquet of voice personalities, including the notorious “Unhinged” mode that does nothing but spew abuse if that happens to be your preference. Gemini can take standing instructions on how to respond, whether warm and witty or factual and formal, and has recently been going all out to wrap all responses up in personalisation.
Despite facing lingering irritations such as sycophancy and hallucinations, most users eventually settle into a conversational mode that suits them.
Call my agent
But developments in AI happen at breakneck speed, and the fact that we are moving into a world of AI agents is something you hear on an everyday basis.
Agents are a big change from the chatbots we’ve grown used to. While AI chatbots read or hear your prompts and give you a response, agents take that next step. They act. When given a goal, they can even divide the work among other agents to speed up the process. With permissions, agents can access different apps and spaces and retrieve the data they need to perform their tasks. There’s no back-and-forth conversation needed, and so the emphasis is no longer on what was a beginner mode for the AI and user relationship.
The appeal of agents is easy to understand. Conversation was the breakthrough that made AI accessible, but execution is what makes it powerful. That shift reduces friction and moves AI closer to transactions, workflows and measurable outcomes. For companies, that’s progress. It signals maturity. It opens clearer paths to revenue. Compared with that, conversation can begin to look like the prelude rather than the main act.
Money, money, money
There’s another reason companies are so eager to move on from plain chat. Agents open up possibilities for revenue. An AI agent that compares prices, books tickets, fills forms or places an order is not just demonstrating capability; it is participating in economic activity. That creates clearer business models. It allows AI companies to position themselves as intermediaries in transactions.
Obviously, this is easier to monetize than talk. In fact, companies have considered bringing advertising into the chat window. That shift is already visible in how the most popular chat products are being re‑engineered. OpenAI has begun testing ads inside ChatGPT for free and low‑cost tiers in some markets, with clearly marked ‘sponsored’ boxes appearing under responses when users ask about things like trips, gadgets or recipes. So, the chat window itself is being pulled closer to the checkout counter. The same space where you ask questions is now also where brands bid to meet you at the moment of intent.
That’s all very well for companies. But for users, it will blur a boundary that made chatbots feel novel in the first place. When every answer carries a quiet invitation to click to buy, dialogue stops being just a way to think aloud with an AI and starts to look more like the pre‑sales funnel for an invisible marketplace.
Keep talking
Despite the headlines, the chat window isn’t disappearing. Everything we do isn’t directly related to productivity and specific tasks. When I argue with Gemini over a French phrase or discuss a song with ChatGPT and get right into the song’s background, the lyrics, and the instruments used, I don't need an agent.
Conversation remains essential because it is how humans naturally think, decide, and build trust, and those needs don’t disappear just because software can now act on our behalf. Studies on decision-support systems show that users tend to trust conversational interfaces more than equivalent graphical ones, even when accuracy is held constant. In any case, we often want reflection, not action.
Agentic AI shines when the task is clear—refund the order, rebook the ticket, update the CRM. But many real-life queries are fuzzy and involve wanting to explore our feelings and opinions. A purely ‘one‑shot, no‑chat’ interface makes it harder to spot or question something that doesn’t seem right. When conversation is the glue between agents, how can it not be between humans and AI?
It’s more accurate to say the chatbot era has become the baseline, not that it’s dead. Industry pieces arguing that the chatbot era is over usually mean that plain Q&A in a text box is no longer the exciting differentiator, and what matters now is conversation plus action. So no, ChatGPT, this isn’t goodbye.
The New Normal: The world is at an inflexion point. Artificial intelligence (AI) is set to be as massive a revolution as the Internet has been. The option to just stay away from AI will not be available to most people, as all the tech we use takes the AI route. This column series introduces AI to the non-techie in an easy and relatable way, aiming to demystify and help a user to actually put the technology to good use in everyday life.
Mala Bhargava is most often described as a ‘veteran’ writer who has contributed to several publications in India since 1995. Her domain is personal tech, and she writes to simplify and demystify technology for a non-techie audience.

1 hour ago
1




English (US) ·