By Kim Talarczyk, founder of Demand Recruiting
If inbound is already waiting and the problem is how fast you respond, an inbound AI agent can help you keep up. If you still need someone to work out the outbound motion, the ICP, and what a good meeting looks like, hire a human SDR, BDR, or hybrid first and give them AI for research and prioritization. Do not buy an AI SDR to create demand that does not exist yet.
Almost every founder I talk to right now is getting pitched an AI SDR. The promise is easy to like. Skip the first sales development hire, turn on a tool, and let the meetings show up. So the question I keep hearing is whether they still need a person at all.
The real decision is whether you need to keep up with demand you already have, or create demand you don't have yet. Those are different problems, and they usually call for different first moves.
Most of the companies I work with are B2B SaaS under $10M ARR. Read on for how I would help a company this size work through these questions. If you are still working out where sales development fits among your early hires, my post on the first 5 GTM hires for B2B SaaS founders covers the wider order.
Before you look at any tool or any candidate, name the thing that is actually stuck. This is where I would start.
| What is happening | What usually comes first | What to read next |
|---|---|---|
| Inbound leads wait too long for a reply | An inbound AI agent can help you respond faster, with a person owning the qualified conversations | When an inbound AI agent helps, below |
| There is no outbound motion yet and the founder is the only one who can book a meeting | A human SDR, BDR, or hybrid | How to hire BDRs and SDRs |
| Meetings are happening and deals are not closing | A seller who can close | Hire your first AE |
| The founder cannot yet describe who buys and why | Keep selling yourself and write down what works | Why first sales hires fail |
An inbound AI agent is for people who already raised their hand. This includes demo requests, form fills, and site visitors who want a fast answer. It can cover after hours, route each lead to the right person, and book the meeting while the buyer is still interested.
This works when two things are true. 1. There is real inbound volume, and 2. You have clear rules for what counts as qualified. If either one is missing, the agent has nothing to keep up with or no way to tell a good lead from a bad one.
Once a lead is real, a person owns the conversation, but the agent gets the buyer to the table faster. It does not replace the person who runs the call and follows through.
If outbound has not been built yet, I would start with a person. That usually means the ICP is still moving, no sequence has been tested, and nobody has agreed on what a good meeting looks like.
At that stage someone has to learn from every reply and every call and bring it back to the founder. Things like which titles answer, which messages get ignored, which meetings turn into real conversations. That learning is the job, and it takes a person.
Which role to hire, SDR, BDR, or hybrid, and what to screen for is in my post on how to hire BDRs and SDRs for SaaS. Whichever you choose, your first rep still needs a playbook from you. I wrote about why in most founders fail their first sales hire because the process wasn't ready.
If you need help building the sales motion itself, I point founders to Hannah at SALT for sales leadership and execution. You can find her on my partners page.
This is where I think AI earns its place. Behind a human rep, it is good at research, account prioritization, and finding a reason to reach out now. It can draft a first line hook that the rep then rewrites. It can take notes and set follow-up reminders so nothing slips.
What it should not own is the message or the relationship. The context that makes outreach land lives with the rep. A deal you nearly won last quarter. A conversation at an event. The tool does not always carry those things, and sometimes we still have to rely on human memory and context. Of course, building out content into your AI tools should be continuous so that AI improves.
If AI is writing and sending every email with the same stale context, the emails start to sound the same as everyone else's.
This is the order I would usually follow.
First, the founder sells and writes down what works. Then a person tests the outbound motion by hand until it gets real replies and good meetings. Once that is happening, AI helps that person and that motion do more.
AI scales a motion that already works. It does not invent one. If you skip the first two steps, you are automating a guess. Once a person has the motion working and manual work is what holds it back, that is when I would hire a GTM engineer to build the system around it.
Buying an AI SDR to create demand. If buyers do not know you yet and there is no motion that works, an AI SDR has nothing to multiply. You send more messages nobody asked for, and you learn very little from them.
Hiring an SDR when the real problem is closing. If meetings are on the calendar and deals stall after them, more meetings will not fix it. That is a seller problem, and it usually means it is time to hire your first AE.
Hiring any SDR before you can say what a good meeting is. Human or AI, the first sales development hire needs a target. If you cannot describe a good meeting, neither can they, and you will end up debating every one they book.
Turning on volume before one sequence works. Get one sequence to earn real replies first. Adding volume on top of a sequence that does not work only gets you more silence.
If you work through this and the answer is a person, that is the part I help with. You can see how a search works or get in touch.
Helping B2B SaaS founders and GTM leaders hire the sales, marketing, and RevOps people who move growth forward.
