I had a long conversation this week with a sales leader who was trying to work out where to put his bets. His company had the same problem most B2B businesses have right now: they know AI is reshaping how sales gets done, they can see other people doing interesting things with it, and they have no clear plan for how to actually adopt it without making things worse.
At the end of the call he asked me what I thought sales would look like in five years. This piece is the long version of that answer.
The short version: the way most B2B sales teams are organised today is a response to scale and specialisation problems that AI is about to solve. The next five years will see those teams reorganise around the assumption that one person, properly augmented, can do the work that used to require three. That has implications for how you hire, how you structure compensation, how you think about pipeline coverage, and what you should be building right now to be ready for it.
The four shifts
There are four changes I think are nearly certain over the next five years. They’re connected: each one makes the next one more likely.
1. The return of the fullstack salesperson. SDR, AE, and AM functions collapse back into a single role. Not because the work disappears, but because AI handles the mechanical layer that originally justified splitting them up.
2. A 5–10x productivity step-change for the reps who adopt well. Not 5–10% better. 5–10x more output at the same or higher quality. Reps who don’t adopt won’t keep up; reps who do will look like a different category of professional.
3. Smaller, denser sales teams. The teams that win in the next cycle won’t be bigger. They’ll be smaller, more senior, more technical, and will hit numbers that today require headcount that won’t exist in five years.
4. A different kind of salesperson. Less time spent on process and admin, more time spent on solutioning, judgment calls, and the parts of the job that actually require being human in the room.
Let me take each of these in turn.
Why the SDR / AE / AM split is about to collapse
The current B2B sales structure (SDR books meetings, AE closes them, AM keeps the customer) wasn’t a designed-from-first-principles model. It was a response to a specific scale problem. Closing reps were too expensive to spend their days prospecting, so prospecting got pushed down to a cheaper role. Customer success got separated out because account expansion required a different motion than new business.
That logic held when the mechanical work was actually the bottleneck. Researching an account properly took an hour. Building a personalised sequence took half a day. Pulling activity data out of the CRM to plan an outreach week took most of a morning. If you wanted to scale, you had to either hire more bodies or specialise the role so each body could go faster.
AI changes the unit economics of that mechanical layer. Account research that took an hour takes minutes. Signal-led prospecting (researching trigger events, then drafting a tailored angle for each one) is now something one person can do at scale. Activity analysis that took a morning is a five-minute Claude project query. The work that originally justified splitting roles into prospecting and closing functions is now the part that’s been most thoroughly automated.
What’s left is the work that always required judgment: reading the room in a discovery call, navigating a complex buying group, knowing when to push and when to wait, building the kind of relationship that survives a customer’s procurement process. That work doesn’t split well across roles. It’s better done by one person who owns the account from first contact to renewal.
So the structure unwinds. Not back to the 1990s “rep does everything manually” model. Forward, to a model where the rep does everything, but with AI doing the mechanical layer underneath them.
What 5–10x actually looks like
When I say a properly augmented rep will do 5–10x the work, I don’t mean they’ll send 5–10x more emails. The volume game is mostly already over and the deliverability environment punishes anyone trying to play it.
What I mean is this. A traditional AE, in a typical week, might run 8–12 first meetings, manage 25–35 active opportunities, draft a couple of proposals, do their CRM hygiene, and spend whatever time is left on prospecting. The rate-limiting step is not effort. It’s the mechanical work that wraps around each of those activities. Pre-call research. Post-call notes. Proposal drafting. Stakeholder mapping. Follow-up sequencing. Forecasting prep. The actual selling is a smaller fraction of the week than most people realise.
Pull most of that mechanical layer out and what you have left is a rep who can run more meetings because each one takes less wraparound time, manage more opportunities because tracking and follow-through is automated, turn proposals around in hours instead of days, and prospect properly because account research doesn’t cost a half-day per account anymore.
That’s not theoretical. Some of it I do already in my own work. Some of it is what I’m building for clients now. The 5–10x figure isn’t aspirational; it’s roughly what happens when a competent rep stops being the bottleneck on their own output.
The catch (and there’s always a catch) is that this only works if the underlying business is in good shape. AI gives a rep with a clear ICP, a good product, and an honest pipeline a massive multiplier. It does the same for a rep with a confused proposition and a CRM full of garbage, and the result is more confused outreach faster. If the foundations aren’t there, that’s the problem to fix first. AI Won’t Fix a Broken Revenue Function covers why in detail.
Smaller, denser teams
The natural consequence of a 5–10x productivity step-change is that you don’t need as many reps to hit the number. That’s uncomfortable to say out loud, but it’s the obvious arithmetic.
The teams that win in the next cycle will look different from today’s:
- Fewer people, more senior on average. The bar to be useful goes up. A junior rep who can’t operate the AI layer competently is a liability, not a cost-effective starter.
- More technical literacy as a baseline expectation. Not “can write Python” technical, but “can build and maintain their own Claude project, can structure a Clay table, can think clearly about what data and signals they need” technical. This is a real shift in the kind of person sales hires.
- Compensation models that look different. Higher OTEs for fewer roles. Less leverage on the SDR-to-AE pyramid because there are fewer SDR roles to leverage from. Comp committees and finance teams will need to update their assumptions.
- Different management structures. A manager managing 8 fullstack reps is doing a different job than a manager managing 3 AEs and 6 SDRs. Span of control changes. Coaching changes. The role of a sales manager becomes more about strategy and judgment, less about activity policing.
This is one of the bigger second-order effects of AI in sales and it’s the one most leadership teams haven’t started thinking about yet. The right org chart for 2028 is not the right org chart for 2025.
What it actually means to be a salesperson in five years
The fourth change is harder to quantify but I think it’s the most important one. It’s about what salespeople actually spend their time and attention on.
Most B2B salespeople today spend the majority of their working hours on process and admin. CRM updates. Sequence building. Follow-up logistics. Pre-call research that gets done in a hurry. Proposal drafting from scratch every time. Internal coordination. The actual moments of selling (real discovery, real solutioning, real negotiation) are a small fraction of the week.
When you take the mechanical layer away, what’s left is the cognitive work. Understanding the customer’s situation deeply. Diagnosing what they actually need. Solutioning around their constraints. Reading the dynamics in a buying group. Knowing when a deal is real and when it isn’t. Having difficult conversations about price, scope, timing. The parts of the job that require being a person, in the room, with judgment.
That’s a more interesting job than most B2B sales is today. It’s also a harder one. Reps who currently lean on volume and activity to compensate for shallow thinking will struggle. Reps who can think clearly about a customer’s business will thrive. The bar moves up and the work gets more like consulting than like the call-centre-with-a-CRM model that a lot of modern SaaS sales has drifted toward.
This is good news for the business, generally. It’s also why the recruiting and management challenges are bigger than people realise. You can’t take the average B2B SDR and turn them into the kind of fullstack rep this model needs. The hire profile changes. The development path changes. The career structure changes.
What to do about it now
If most of this is right, the implication isn’t to wait and see. Five years sounds like a long time but it isn’t, especially given how long it takes most companies to change how they hire, how they comp, and how they organise.
Three things worth doing in the next 90 days, regardless of where you are now:
Get the foundations honest. Clean CRM data, real pipeline stage definitions, an ICP you can actually defend, a sales process documented well enough that it could be handed to someone new. None of the AI work matters if these aren’t in place. If you don’t know whether yours are, What to Audit Before Hiring a RevOps Consultant is a self-check worth doing.
Build one Claude project that does something real. Not a demo, not a proof-of-concept that gets parked. One actual workflow your team uses every week. Account research, proposal drafting, post-call summaries, pipeline analysis. Pick the most painful mechanical task and build something around it that proves the model works in your specific business. The rest gets easier once you’ve got one working example. Building a Business Development Claude Project walks through how I do this for clients.
Compress the meeting-to-proposal gap. This is where the productivity gains show up first and where customers feel them. Reps that can turn a discovery call into a credible proposal in hours rather than days will close more business. The mechanism is a brand-aware AI workflow built on top of your sales collateral. From Meeting to Proposal in Hours covers how it works.
The bigger structural changes (smaller teams, fullstack reps, different comp models) take longer and don’t need to be solved immediately. But they’re the direction of travel and the businesses that start adjusting now will be in a much better position in three years than the ones still treating AI as an experiment.
A note on what I’m not predicting
I’m not predicting that AI replaces salespeople. The work of selling, the actual cognitive and human parts, doesn’t compress that way. Anyone selling you “AI SDRs” as a replacement for human reps in B2B is selling you something that doesn’t work outside of very narrow product categories. Buyers in serious B2B deals don’t want to be sold to by a bot, and the deals that matter most are the ones where the human work is the work.
I’m also not predicting that this happens evenly. Some segments (high-velocity SaaS, low-ACV transactional sales) move faster. Others (enterprise, complex services, regulated industries) move more slowly because the buying motion is fundamentally different. The fullstack model lands in different parts of the market at different speeds.
What I am predicting is that within five years, the gap between teams that have adopted this model well and teams that haven’t will be enormous, and that gap will be commercial, not technological. The technology is here now or will be soon. The hard part is building the operating model around it. That’s the work most sales leaders haven’t started yet.
For the operating model in detail: Building a Business Development Claude Project, From Meeting to Proposal in Hours, and The AI-Augmented Sales Manager cover the three workflows I’d build first. For the broader principles on AI in revenue functions, the AI & RevOps page covers what’s worth doing and what isn’t. And if you want to talk about what this looks like for your specific business, get in touch.
Free resources
Four phases for removing yourself from day-to-day sales without revenue dipping.
Free download22-point pre-raise checklist and a 90-day post-raise build plan.
Free downloadScore your revenue function across 5 areas. Find the gaps before you hire.
Free download