An AI agent for sales: less busywork, more deals and quota

According to the research, almost 80% of salespeople missed quota last year. And here is the interesting part. Those who have truly learned to work with AI hit their quota 3.7 times more often than everyone else.

Not because AI sells for them, but because it takes the busywork off their plate and gives back time for talking to customers. After all, selling itself takes up less than a third of a salesperson’s working time. Everything else is eaten by researching accounts, updating the CRM, prepping for meetings, and the constant fiddling with data that is forever out of order. No wonder things turn out the way they do.

Plenty of people can use AI, yet only a few get that kind of result out of it. In this piece we will look at where the profession is heading according to major research, what an AI agent actually is and how it differs from an ordinary chat, which specific salesperson’s tasks it already handles, and how it helps bring in more deals.

Where the market is heading

In the recent Salesforce State of Sales report for 2026, which surveyed more than four thousand salespeople, the picture is this. More than half already use AI tools, and almost nine in ten plan to by 2027. You would think the train has left and everyone is on it. But the most interesting thing is that using AI and getting a real result out of it are two very different things today.

And the difference here is huge. There are two formats. The first is conversational. You open a chat, paste in an email, a contract, or a chunk of correspondence, ask it to rewrite it more punchily, check the clauses of an agreement, or suggest an argument, get an answer, and move it back to yourself by hand. The second is agentic. You set a task, and the system itself goes into your CRM, email, calendar, and open sources, gathers everything needed about the customer, prepares a call brief or a draft email, and updates the deal record, and all that is left for you is to check and send. The first is a handy prompter within reach, it barely changes your actual speed. The second takes on the very busywork that leaves no time for selling.

That same 3.7x gap is a Gartner estimate, and it is about the fact that the point is not a couple of handy prompts but a different level of work. And a large study by top American universities together with OpenAI, “The Shift to Agentic AI: Evidence from Codex” (June 2026), shows where this gap comes from. Agents have already moved beyond development into ordinary work functions and are taking on ever longer tasks that a person used to do alone in an hour or more. But for now this is the lot of the few. Most of those who say “I use AI” are still stuck messaging with a chat and get almost no gain, while the real payoff is captured by the few who have moved to agents.

The market is already pricing this in. Teams where salespeople work with AI grow their revenue noticeably faster than those without it, and it is precisely the strongest salespeople who use these tools most often. The ability to work with AI is increasingly expected right at the door, and for salespeople themselves it gives back what they are actually paid for, time for the customer and for the deal. And yet, truly, at the level of an agent, only a few have learned this so far, and it is in that gap that the opportunity lies.

It comes down to a simple thing. The market already needs salespeople who can do more than just ask a chat, who can hand the busywork to an agent and tune it to themselves. For now there are few of them. Whoever learns this stops spending most of the week doing anything but selling and returns that time to deals and to quota. So the whole question is who moves into that camp first. And the good news is that getting there is far easier than it seems. Exactly how, we will cover next.

What an agent actually is and how it differs from ChatGPT

If you strip away the fancy words, an agent is a program that does not answer a question but performs a task. The difference is roughly like that between a consultant and a doer. A consultant tells you what needs to be done, but you still do it yourself. A doer takes it and does it, and you check.

An ordinary chat waits for you to bring it all the context yourself. You copy the correspondence, explain who the customer is and what stage the deal is at, and then move the answer back to yourself by hand. An agent is given access to the necessary tools once, and after that it goes into the CRM and email itself, pulls up the customer’s history, prepares a brief or a draft email, and updates the deal record. All that is left for you is to check and confirm.

It is arranged quite simply. The agent works in a loop. It received a task, figured out what to do as the first step, say pull up the deal in the CRM or read the latest call transcript, did it, looked at the result, moved on to the next step, and so on until the task is done. There is no need to get into this, just as there is no need to understand how a mail server works in order to send an email.

There is one important point, and it is simpler than it seems. For the agent to work not at random but by the rules of your particular sales, you do not need to train it by hand and spell everything out in text. Setup is more of a joint effort. You hand it your process, your customer profile, your deal stages, your winning emails and positioning, examples of past calls, and it works through all of it itself and assembles a set of rules it will follow from then on. You check this and correct it where needed. After that, these rules are reused in every task, and any request is handled just as evenly, whether you launched it yourself or a colleague covering for you did.

Which tasks it already handles

If you look at what such agents actually do for salespeople today, a few typical scenarios take shape. Below are the most common ones, just as examples, so it is clear what we are talking about. In reality you can automate almost any repetitive chunk of work this way; these are taken for illustration.

The most common story is the morning research before a call. Usually, before a meeting, fifteen or even thirty minutes go into googling the company’s news, glancing at the CRM, rereading the past correspondence, and remembering where you left off. An agent that has access to your CRM, email, and calendar assembles such a brief itself. It pulls up the customer’s history, can read the notes of all the meetings in a minute, checks fresh news and the deal status, and hands over a short summary you can read in a couple of minutes before the call. And it can work proactively. For example, in the morning it went through all the meetings in the calendar for the day itself and sent a brief on each in advance, before you even remembered them. One salesperson set up exactly such a task, which every morning checks the calendar a few hours ahead, pulls the context on each deal and the latest call transcript, and assembles a two-minute-read brief.

The second familiar scenario is updating the CRM, which you never quite get around to. Agreements from meetings and emails often never make it to the deal fields, and it surfaces only right before the pipeline review with your manager. Here the agent takes a lot off your plate. After a call it takes your notes or transcribes the meeting recording itself, understands it all, and updates the deal record itself, sets the next step, amounts, stage, and everything you agreed to record. If needed, it can create tasks for you in a task tracker itself. You just confirm it, and the CRM always stays up to date without the evening cleanup.

The third scenario is lead research and outreach personalization. A proper personal email takes ten, twenty minutes per person, and at scale there simply is no time for it. The agent takes this chunk on itself. One solo salesperson described how it works for him. He manually selects a list to fit his customer profile, and then the agent works through it itself, filters out the irrelevant ones, finds the right people and their contacts among the rest, gathers a short report on each company, creates a record in the CRM, and prepares a draft email right in the inbox. All that is left for him is to check, tweak the draft, and hit send. In his own words, the only thing he does not trust the agent with is sending outreach unsupervised, and that is right. Pre-screening that used to take days now takes minutes, and checking a single lead takes about a minute instead of an hour.

And the fourth scenario, the most telling, is when the agent assembles you an assistant across the whole pipeline. It keeps all your deals in mind and prepares itself the things there is usually no time for. For the Monday review it sends a digest of the deals with a note on where the risks are and what is slipping. Before a sync it assembles a list of meetings for the past and current week with types and statuses. Once a week it hands over the pipeline status and forecast on a single page. One GTM leader set up such a morning task, which pulls deal status from the CRM and spend figures from the data warehouse and drops a ready brief into the chat before the first meeting. This way you get a junior on the team who takes on not one task but a whole recurring layer of prep, while you plug in on the decisions and on selling itself.

And the most valuable thing is that these scenarios only get better over time, as the agent accumulates information. You set up a scenario once, and from then on it is essentially a recorded playbook of your work, one that gets reused every time and that you can gradually improve. Your way of running a customer, prepping for a call, and running a deal stops living in one person’s head. The same scenario can be handed to the whole team, and a new salesperson from day one preps for meetings and updates deals exactly the way your best one does, rather than figuring it out on the fly for months. You work out once how you sell, and from then on it works for you continuously.

What the limitations are

Here two honest caveats are needed. The first is that AI sometimes gets it wrong. If access to the data dropped for a second, it may substitute a plausible but made-up figure just to fulfill the request. That is why the agent is a first-pass doer, not a final authority. It takes the busywork and the prep on itself, but the CRM write goes through your confirmation, and outreach does not go out to customers without your supervision. The key decisions and final wording remain, of course, with the salesperson. During setup the agent is explicitly given the rule to rely only on real data, to flag where it is unsure, and to make nothing up.

The second caveat is confidentiality. Customer and deal data must not leak into someone else’s cloud, so it is wiser to keep files and access with you. That way you get the agent’s speed while controlling where the data sits yourself.

Keep these two rules in mind, a human checks and confirms, and the data stays with you, and the agent will not become a risk. It is simply a faster way to do the same work with the same confidence in the result.

How to start using this

You can assemble such an assistant yourself. But the point is not only spending a couple of evenings on the setup. You also have to figure out how it all works, what to connect where, how to describe the scenarios, which basic rules to lay down, try it, make a mistake somewhere, and redo it. That takes time, and not every salesperson has the desire to dive into it, especially with quota burning and plenty of their own work to do.

That is why there is an easier path, taking a ready-made, preconfigured solution. There are solutions like kvelo.dev that already work as agents and assemble a personal workspace for you, where everything needed, including the typical salesperson’s scenarios, is already built in and configured automatically, without long manual fiddling. It also brings together that very zoo of eight tools you usually jump between. After that you just work. You do the complex tasks at the computer, and the simple questions you solve right from a messenger on your way to a customer. You write in Telegram or WhatsApp “put together a brief for the meeting with the customer in an hour,” and the agent went into the CRM, email, and news, pulled up the deal history, and sent a ready summary. In essence you talk to it like a colleague.

By and large, the market has already decided that working with AI in sales is the new normal. The difference between those who stay messaging with a chat and those who learn to delegate to agents will soon show in the quota, in the time spent on the customer, and in who on the team grows and who drowns in reports and the CRM. That is why it is worth starting to get into this now. There is still time, and the chance to be among the first is still large. And to figure it out as simply as possible and immediately build it into your work, you can start with a ready solution like Kvelo.dev and just give it a try.