Between 2018 and 2019, I ran my first SDR team. We had calling booths, a whiteboard leaderboard and a bell for every meeting booked. On a good call-a-thon day, six SDRs would book around 30 meetings between them. I was figuring it out as I went, and volume was the whole game.

After ten years in go-to-market, I thought I knew outbound. In the last six months I've had to relearn a lot of it. Search became answer engines, LinkedIn changed, and the buyer on the other end of the phone changed most of all.

The easy story is that AI broke sales. I don't think it did. AI didn't break go-to-market. It exposed it. Every team that never had a clear customer profile, real insight into its buyers or a system that worked is now finding out, faster than before.

Most of what follows comes from my conversation on the GTM & AI Operators Show with Rahul Wadhwa, founder of School of SDR, School of Sales and School of Interviews, who was on the founding team of Nevara, an AI assistant for account executives, and a discussion with Akshay Doshi, a fractional GTM leader for seed to Series B startups, who spent the last six years as SVP of Sales at SpotDraft, building and scaling the sales motion from the ground up.

You can watch my conversation with Rahul Wadhwa on AI and sales here.

Your buyer has done the homework

Think about what you do when an appliance at home stops working. You take a photo and ask ChatGPT. You watch a few videos. Only when you're out of ideas do you call the service centre. Rahul's point is that B2B buyers now do the same thing. By the time they speak to you, they may have spent a weekend trying to solve the problem themselves. They've attended the webinars and heard every pitch. What they haven't heard is your point of view.

The data backs this up. In Gartner's survey of 645 B2B buyers, run in late 2025, 67% said they'd prefer to buy without a sales rep, and 45% had used generative AI on a recent purchase. Yet 69% still wanted a rep to check what the AI had told them.

"The first call used to be where the buyer learned about you. Now it's where they check their homework."

Akshay Doshi, Fractional Revenue Leader

Akshay sees buyers arriving with a shortlist, a view on pricing and questions they've already half-answered. They don't need the category explained. They need to know whether what they've learned holds true for their team, their tools and their timeline, and why they shouldn't just build it themselves. Even a self-serve product he sold for under $500 a year had buyers asking to talk to someone, not because it was complicated but because they needed help justifying the purchase internally.

That changes the first call in three ways:

1. It's a validation call, not a discovery call. Confirm or correct what they think they know.

2. Bring a point of view. AI can do the pre-call research, so show you've done your homework on their business.

3. Qualify on fit and urgency from minute one. The buyer arrives informed. The goal is a faster yes or a clean no.

The worst thing you can do is hide basic answers behind a demo. When an SDR tells a buyer "come to the demo and my AE will answer that", the buyer won't wait. Keep the product walkthrough for the demo. Give the basic answers now.

The first call is now a check on the buyer's homework, not a lesson.

Everyone has Claude. Your edge is the other 3 out of 10

AI learns from what's on the internet, which means it learns the average. The best article it can find on writing an outbound email was written to work for most readers. That's a 7 out of 10. The other 3 is you: what you've done, what you've seen, the notes in your own notebook that no model has read.

AI gets everyone to 7 out of 10. The last 3 is what you bring.

The trouble is that most people stop at the 7. Rahul knows fewer than five people who've built something with AI that's worth selling or using. Most salespeople pick up someone else's template and try to make it fit. And experience on its own isn't the same as expertise. A VP who hasn't made a cold call in five years will lose, every day, to someone who still does the work and knows what's happening on the ground.

Akshay found the same thing when his team broke down what its top SDRs did differently. The answer wasn't "used AI more". The best one used AI for deeper research on fewer accounts. The others in the top five got their pipeline from things AI can't do on its own: reviving lost deals with real context, bringing account executives in early, and blocking hours every day for calling. He sees four things that separate the best sellers now:

• Precision over volume. Use AI to find the few hundred accounts that really fit, not to send ten times more email to the ones that don't.

• Context the model doesn't have. Why the last deal was lost, who knows whom, what the buyer said on the last call.

• The live conversation. AI made writing cheap, which made a good conversation worth more. When his team put a parallel dialler in place, meetings booked went from one or two a month to 30 or 40.

• Commercial rigour. The best account executives look more like solution engineers: a sharp demo, a quantified problem, and a business case the champion can defend.

His warning for SDRs is blunt. The email-heavy version of the role is in jeopardy. The ones who thrive will call, read signals and handle objections alongside the automation.

None of this comes from a template. Rahul compares a market playbook to half-cooked chicken: you can turn it into any dish you like, but only if you know your ingredients. What worked six months ago may not work now.

Aim before you fire

AI is a spray gun. It can make anything you point it at more powerful, which is exactly why the aim matters more than ever.

Start with a simple test. Put your team in a room, hand everyone a blank sheet of paper and ask them to write down your ideal customer profile. If the answers differ, that's your number one problem. Everyone is hitting in different directions, and every conversation feels like a game of Chinese whispers. Until the ICP lives in writing, nothing downstream works.

Here's what a sharp one looked like at Nevara. The team started with just New York and San Francisco, where the founder could show up in person. They wanted account executives with two to three years of experience, at least six months in their current company, and a calendar already full of back-to-back demos. That last signal was the hard one. Talking to customers, they learned the demos came from marketing teams running ads, and that those teams had website traffic of 27,000 or more. That became the proxy.

A human found that signal. No AI could have. But once they had it, an agent could check every new company against the criteria. Their top tier, A+, met every criterion, had an SDR team of three or four, and was hiring. Because their buyers lived on LinkedIn, a targeted message campaign to them got a reply rate above 30%.

Then keep the list fresh. A list picked on 1 January is stale by February. Akshay ties the cadence to the segment: monthly for small businesses, quarterly for mid-market, every six months for enterprise. Two principles matter more than the cadence:

• Do the capacity maths first. His team once worked out that a rep with 177 accounts, converting 5% with half of those becoming qualified, would produce about five qualified opportunities a quarter. The target was 18 to 23. That wasn't ambitious. It was impossible.

• Stack signals, and use them to prioritise. One intent signal is a commodity; your competitors are buying the same one. Combine fit with timing triggers such as a new leader, a system migration, a funding round or a hiring spree. A website visit tells you when to reach out. It doesn't belong in the message.

If the book can't produce the number, the target isn't ambitious. It's impossible.

Do the things that don't scale

When messages went unanswered, the Nevara team went to the office. They printed coffee bags with their logo. The founder cycled to five companies a day, tagging each one in a post in advance to say he was coming. If the person wasn't in, he left the coffee with someone else, took a photo, and sent it to them. Some installed the product on the spot.

I've done my own version. At one company we made personalised photo frames and handed them to people at conferences. Enterprise deals took six to eight months and many touches, and the frames were one of the touches that worked.

None of this scales, which is the point. Paul Graham made the case years ago in "Do Things That Don't Scale": early on, the founder's job is to get close to customers in ways a process never could. The thing that stops most founders isn't the idea. It's discomfort. Rahul has worked at a company whose founder was too shy to talk to customers. Get as close to the buyer as you can, and try ten things. Some won't work. Some will work for a while. You only find out by doing them.

The 30-minute rule

Salespeople have never spent much of their week selling. In Salesforce's State of Sales research, based on a 2022 survey of 7,775 sales professionals, reps spent just 28% of their week actually selling. The rest went to deal admin, data entry and everything around the sale.

AI should change that, and Rahul's rule makes it concrete. No SDR should spend more than 30 minutes working out what to write, what to send or what to say on the phone. With the tools available now, anything more means something is wrong. If he were an SDR today, he'd spend about six hours on the phone.

Account executives have the same problem in a different form. Someone with back-to-back demos for five or six hours is five or six hours behind on follow-up by the end of the day. A call recorder and a simple AI skill that writes the summary fixes that. If AI can do it, a seller shouldn't be doing it.

The catch is that AI only frees up time if the basics are in place. In four years of quarterly reviews, Akshay saw the same line under "what could have been better" every single time: CRM hygiene. The team kept adding tools and the problem stayed. Tools amplify whatever discipline you already have. With messy data, AI just produces confident nonsense faster.

His fix is structural. Manual work like copying phone numbers, splitting lists and logging LinkedIn activity is plumbing, and it shouldn't sit with a rep. It belongs with a GTM engineer who owns the data, the signals and the automation. His working view is that one strong GTM engineer and a small, high-quality team can now produce the pipeline that used to take a much larger team. I've written separately about what GTM engineering is and what it replaces.

The best outbound team is the one buyers already know

Here's the paradox at the heart of Rahul's business: he teaches outbound, and all of his business is inbound.

It started with a lost promotion. Years ago, he and a close colleague were up for the same trainer role. He was hitting his numbers; she was better at being seen. She got it. He decided he wouldn't lose another opportunity for lack of a brand, and a colleague nudged him to start posting on LinkedIn.

Two habits made it work. The first is taking public stands. He calls out companies for unfair hiring, and six or seven of them have become clients as a result, including one whose CEO reached out to ask how to fix it. The second is research before content. He used Google Trends to find that far more people search for interview help than for how to get into sales. One video on interviews passed close to a million views, brought about a thousand followers overnight, and now adds about five thousand a month, a wider funnel that feeds his sales school.

Buyers reward this. In the 2025 Edelman and LinkedIn B2B Thought Leadership Impact Report, 95% of buyers said good thought leadership shows a vendor's value better than traditional marketing, and 79% said they're more likely to champion a vendor in a purchase if it publishes consistently.

Akshay calls a personal brand the most underrated pipeline asset an early-stage founder has, and the one most people quit before it pays off. When every inbox is full of AI-written outreach, trust becomes the scarce thing. His rules:

• The founder goes first. Founder content beats the company page. If writing is the bottleneck, do a weekly brain dump and have someone turn it into drafts the founder edits.

• Connect with your buyers, then post. Content that never reaches decision-makers is a diary, not a channel.

• Give it six months before you judge it. It compounds; don't measure it like outbound.

• Pair it with proof: customer logos, case studies and real numbers, or the inbound won't convert.

For reps, it helps but doesn't replace the work. A personal brand makes outbound land better. It doesn't make outbound optional. I've written about the order that makes LinkedIn work separately.

Sell the outcome, not the hours

AI has made a lot of work cheap, so buyers are asking what they're really paying for. Rahul used to pay an editor around $300 to cut each podcast episode. A tool now does most of it for about $50 a month. If editing an episode yourself takes three hours, you're effectively deciding whether your hour is worth more or less than $100. Or you pay someone for the outcome: a finished episode.

He's applied that to his own business. For company training, he now charges three times as much, guarantees a measurable improvement, and keeps working for free until the team reaches it. If a client just wants a training session, he's happy to walk away, because there's no accountability in that. He recently signed three large training deals on those terms. He'd like software to move the same way: credits that don't expire, instead of a monthly fee he keeps paying in months he doesn't use the tool.

Akshay, who runs his own fractional practice, has landed somewhere slightly different. His best model is a fixed monthly retainer with clear, outcome-based deliverables built in. The reason is practical: timelines shift with what the client prioritises, and he can't control that. Outcome pricing works, in his view, when the outcome is narrow, measurable and mostly in the provider's control, such as meetings against a proven customer profile, or a conversion lift on a funnel that already works, over a long enough engagement to measure it.

So his outcome lives in what he commits to, not in how he invoices. Every sprint has defined deliverables and a before-and-after measured together: win rate, sales cycle, pipeline coverage and ramp time. I think both are right. Guarantee the outcome when you control it. Commit to it when you don't.

My SDR team in 2018 won by making more calls than anyone else. That advantage is gone; anyone can send more now. The teams that win next will make fewer, sharper moves, aimed at buyers who already know who they are.

FAQs

How is AI changing B2B sales?

Buyers now research vendors on their own, often with AI tools, before talking to sales, and expect the first call to validate what they've learned. Sellers all have the same AI tools, so the advantage moves to judgment, context about the account, live conversations and a clear ideal customer profile.

Do B2B buyers still want to talk to sales reps?

Many prefer to research and buy on their own, but most still want a rep to confirm what they've learned. In a Gartner survey of 645 B2B buyers, 67% preferred a rep-free experience, yet 69% wanted to validate AI-generated insights with a sales rep.

What makes a good SDR in the age of AI?

Precision over volume, context the AI doesn't have, strong live conversations on the phone, and the ability to read signals and handle objections. Email-only SDR work is the easiest part to automate, so the role is moving towards calling and judgment.

How often should sales teams change their account lists?

It depends on the segment. A common approach is monthly for small businesses, quarterly for mid-market and every six months for enterprise, re-ranked using fresh signals such as funding, leadership changes, hiring or system migrations.

How much time should SDRs spend researching and writing?

As little as possible. With AI tools, deciding what to write and say should take minutes, not hours, leaving most of the day for live conversations. One rule of thumb is no more than 30 minutes, with around six hours on the phone.

Does a personal brand on LinkedIn help with B2B sales?

Yes, especially for founders. When prospects recognise the sender from their feed, outreach lands warmer. It compounds slowly, so give it about six months, connect with your target buyers before you post, and back it with proof such as case studies.

What is outcome-based pricing in sales services?

Charging for a result, such as more meetings or a measurable improvement, instead of time or effort. It works best when the outcome is narrow, measurable and largely in the provider's control. Otherwise, a retainer with clear outcome-based deliverables is often fairer to both sides.