Here’s something I’ve been thinking about a lot lately.
Everyone’s worried AI is coming for sales jobs. And yes, it is. But maybe not for the reasons you think.
The uncomfortable truth? AI isn’t going to replace great sales reps. It’s going to replace the ones who were never really doing the job in the first place.
What a Great Sales Rep Actually Does
Let me break down what a truly great AE does for a prospect. Not what they say they do. What they actually do:
- They own the scheduling. Not “let me know what works for you” — they send the calendar link, they coordinate the five stakeholders, they handle the timezone math, they reschedule when the CFO gets pulled into a board meeting. They make it easy to buy.
- They run real demos. Not the generic product tour. They learn your business, they customize the flow, they show you your use case with your data. They make you see yourself already using it.
- They know the product, industry and competition cold. So they can tell you exactly why, when and where they win.
- They solve the actual problem. When you hit a technical snag or a weird integration question, they don’t just loop in SE and disappear. They quarterback the whole thing. They own it until it’s fixed.
- They sell the room you’re not in. This is the big one. They give your champion the deck, the ROI calculator, the one-pager, the talk track. They arm your internal buyer to sell when you’re not on the Zoom.
- They follow up relentlessly. Not annoying “just checking in” emails. Real follow-up. “Here’s the case study you asked about.” “Here’s how Company X solved that same security concern.” “Here’s the procurement language that worked for your competitor.”
- They simplify deployment. Before the deal even closes, they’re thinking about implementation. They’re looping in CS. They’re making sure this thing actually gets used.
Now Here’s the Problem
How many reps on your team actually do all of this?
Be honest.
In my experience, maybe 20% of reps truly do the work. The rest are waiting. Waiting for the prospect to schedule. Waiting for the prospect to follow up. Waiting for the prospect to sell internally. Waiting for the deal to close itself.
And look — AI can already do a lot of this:
- Scheduling? Solved.
- Follow-up sequences? Automated.
- ROI calculators and custom content? Getting there fast.
- Answering technical questions at 2am? Easy.
- Never forgetting a single detail from a call? Already better than humans.
AI doesn’t get tired. AI doesn’t get lazy on a Friday afternoon. AI doesn’t “forget” to send the case study.
The Real Competition
So here’s the question every sales leader needs to ask:
How much of the work are your reps actually doing — versus how much are they leaving on the table for the prospect to figure out?
Because AI is coming for the gap.
Not the great reps. The great ones will use AI to become superhuman — to do even more of the work, even faster, for even more prospects.
But the reps who were coasting? The ones who send three emails and mark a deal “closed lost” when they don’t hear back? The ones who make buying hard instead of easy?
That’s who AI replaces.
The Real Survival Test: Deep Product Expertise
But doing the work isn’t enough anymore. There’s something even more fundamental.
If your sales rep doesn’t know your product better than AI does, why would your prospect talk to the human?
Think about that. Really think about it.
Your prospect can now upload your product documentation to Claude and get accurate, instant answers to complex technical questions. 24/7. No scheduling. No “let me get back to you on that.” No fumbling through a demo.
I’ve tested this myself. Upload a company’s prospectus, pitch deck, or product docs to an AI, then ask it hard questions — the kind sophisticated prospects ask. It will beat most sales reps right out of the box.
And here’s what makes it worse: AI with memory gets better with every interaction. It learns. It compounds knowledge. It never forgets the Q3 release notes or confuses them with the Q4 roadmap.
Your average rep? They plateau at month six and start looking for their next job at month twelve.
AI Makes Buyers Smarter Than Reps. That’s A Real Issue.
I had a demo recently for an enterprise AI product. Should have been a slam dunk — great product, real traction, I came in ready to buy.
Then I asked about their MCP support. The rep had no idea what I was talking about. When I mentioned Claude, they said “my guys will look into it.” When I pressed on technical details, they responded: “Your secret to being good at AI was really sick prompts.”
Then they threw up a pricing slide.
I didn’t buy.
Here’s another one. One of my hottest portfolio investments was meeting with their largest prospect ever — potential $1M deal. The first meeting included their well-regarded CRO, someone with a strong background in “technical selling.”
When the prospect started asking technical questions, the CRO looked confused and asked: “What’s an API call?”
For the follow-up meeting, the founders made a brutal decision: they left the CRO sitting in the lobby and went into the meeting themselves.
They closed the $1M deal without him.
This isn’t about founders thinking they’re better at sales. It’s about recognizing that when you’re talking to sophisticated buyers about complex products, credibility is everything. And if you don’t know the product cold, you have zero credibility.
The New Math on Sales Roles
So what does this mean for sales teams?
Here’s my prediction: you’ll still need humans in “sales.” But many of them won’t be traditional reps.
They’ll be Forward Deployed Engineers.
Look at what’s already happening. Palantir invented this model — FDEs who work directly with customers, understand their specific use cases, build end-to-end workflows, and iterate until the AI actually works. Scale AI does it. OpenAI does it. Decagon has a team of “Agent Product Managers” who stand up AI agents for each customer.
What do FDEs actually do?
- Work directly with customers to understand their processes
- Build and customize workflows until they work
- Handle model training and edge cases
- Own implementation from Day 1 to live deployment
- Solve real-world problems daily
They’re part engineer, part consultant, part AI trainer. And they’re eating traditional sales roles alive.
The Deployment Gap Is the New Competitive Moat
We now have 20+ AI agents running SaaStr. Here’s what criteria we used to pick vendors:
We picked the ones that helped us the most.
Not the ones with sales folks who wanted to “get on a call” before showing us anything. Not the ones bashing competitors. Not the ones sending 100 LinkedIn InMails.
The ones that rolled up their sleeves and helped us get our AI agents running. Including the FDE team at Salesforce.
The greatest AI agent in the world — without top-tier help deploying it — is not that valuable for complex workflows. You need both: a great product AND a team to help you get it working.
The CEO of an AI agents company we really admire asked why we didn’t pick them. Honest answer? Their VP of Sales argued with us, wouldn’t do any work up front, wouldn’t give us access to features we needed, wouldn’t help train our agent.
So we went with someone else.
The Economics Are Brutal But Clear
The FDE model works beautifully at high ACV ($50K+). You can afford dedicated engineers for each customer.
But what about SMB? What about $5K ACV? Who’s doing 30 days of training for each customer?
This is where most AI companies are stuck. The ones that figure out how to systematize the FDE training process will win the SMB market. Capture expert knowledge once, deploy it at scale.
That’s exactly what we did with SaaStr AI. I did manual training for 60+ days. Now that knowledge works for thousands of users.
Most companies haven’t solved this yet.
The One That Does The Most Work Often Wins The Deal. That Won’t Always Be a Human.
The best sales reps have always known the secret: the one who does the most work for the prospect usually wins the deal.
But now there’s an addendum: and if you don’t know the product better than AI does, you’re already obsolete.
AI is about to expose every rep who forgot both lessons.
The question is: which kind of team are you building? One full of people who do the work AND know the product cold? Or one that’s about to get replaced?
The reps who survive will be the ones who become true product experts — or who evolve into FDEs, AI trainers, and deployment specialists.
Everyone else is just waiting for the email.
