Sales development representative is a job with an actual definition. Build a list, work out who on it is worth contacting, make contact, sort what comes back, get the good ones onto a calendar. That definition is why "AI SDR" works so well as a product name and so poorly as a description of anything. On one pricing page it means a sending tool with a language model writing the first line. On the next it means a research agent filling columns in a spreadsheet. On a third it is a chat window sitting on a contact database. All of them are sold under a title that implies somebody clocked in, and the founder pays for a role and receives a feature.

Start with what the human version of the job contains

A sales development rep at a working company does four things. They build a list of accounts and people. They decide which of those people are worth contacting this month and which are not. They make contact, by email or by phone or by whatever gets an answer. Then they handle what comes back, which mostly means sorting replies into interested, not now, wrong person, and stop emailing me, and moving the interested ones onto a calendar.

There is a fifth thing nobody writes into the job description. A decent rep tells you the list is wrong. After a couple of months of calls with the wrong title at the wrong size of company, they come back and say so, and that is often worth more to a young company than the meetings were.

Measure any software wearing the title against those five, one at a time, and the picture gets clear fast.

The parts the software genuinely does

List building is real. You can query a contact database by industry, headcount, location, and title and see how many people match before you commit to anything. Enrichment is real: software finds and verifies a work email or a direct dial faster than a person clicking through company sites at eleven at night. Drafting is real too, and founders undersell it, because the honest comparison is not some beautifully written human email. The honest comparison is the same template pasted a few hundred times with a first name swapped in.

Reply sorting works. Deciding whether an inbound message is interest, a handoff to a colleague, a hard no, or an autoresponder is the kind of task language models handle well, and it takes back the most demoralizing hour of anyone's week.

Then there is endurance, which never makes a landing page because it sounds boring. The four hundredth email goes out written as carefully as the fourth. The day-nine follow-up gets sent after you have forgotten the person exists. That consistency is most of why this category makes money.

The title starts writing checks around the word "rep"

"Rep" implies judgment. It implies someone picked these thirty accounts for a reason they could explain if you asked. What nearly every product in this category actually does is apply the filters you typed. If your idea of a good customer is wrong, the software has no way of noticing. It will run the wrong definition at speed, and you will decide that outbound does not work for your business when the truth is that you aimed it badly.

"Rep" also implies someone you can coach. You cannot take this software aside after a bad week. Correcting it means rewriting instructions, changing filters, rebuilding a sequence. Fine work, and worth doing, but it is closer to configuration than to management, and the pitch tends not to mention that.

The third overclaim is quieter. Plenty of products in this category treat their job as finished the second a reply is detected and classified. What they do next is send you a notification.

That salary comparison on the pricing page is doing sales work

You know the one. A monthly figure set against the fully loaded cost of a junior rep, salary plus tax plus tooling plus ramp. It is a frame designed to make a subscription look like a bargain, and it quietly invites you to buy the software as a replacement for a person and then judge it as one.

The smaller version of the claim is more useful. This software takes the mechanical hours out of a job you were doing yourself at midnight. If you are a founder at pre-seed who was never going to hire a rep in the first place, that is the whole benefit, and it is a good trade. What you should actually study on the page is the second and third row of the pricing table. Watch what the number does as sending volume, seats, or contacts climb. That curve tells you what you are buying far better than the comparison at the top does.

Most products with this title cover one drawer of the four

Here is the setup a founder usually ends up assembling. Apollo or Lusha for the contact data. Clay when the enrichment has to be smarter than a single lookup. Instantly for the inboxes and the sending. Calendly for the booking link. A CRM underneath so the records live somewhere. Zapier across the gaps, moving a record from one to the next.

Every boundary between those products is a place where the work comes back to you. Ten minutes here. A field mapping there. A Tuesday morning finding out why replies stopped syncing into the CRM three weeks ago. Add it up over a quarter and you are the integration layer for your own go-to-market, which costs more than any of the subscriptions do.

The worst seam is the one nobody bills for. A reply lands at four on a Thursday afternoon saying yes, send some times. Nothing happens until a human opens a tab. By Monday that person has stopped caring.

Sales development has always been part telephone

This is the biggest hole in the category and it barely gets argued about. Outbound generates inbound calls. Someone reads your email, does not reply, and two days later dials the number in the signature. If that rings out you paid for the lead and then threw it away. The fix gets sold to you as another purchase: answering services like Ruby and Smith.ai will cover the phone, billed by the minute or by the call, on top of everything you already pay for the sending tool.

What we ship under this heading, stated plainly

Rocketship does not sell a mascot with a first name and a stock headshot. It sells an AI worker with a defined shift, sitting in a shed next to the CRM, the app builder, quoting, scheduling, and payments, so that a reply and a phone call both land somewhere useful.

Launch is $24.99 a month. It finds buyers, writes to them, reads and answers what comes back, and your number is answered around the clock. One worker, one inbound line. Searching the market, counting how many people match your filters, and previewing 25 of them by name are free, so you can learn whether your market is 400 people or 40,000 before spending anything. Adding someone to your list costs one credit. A verified phone and email cost five, charged only when they are found. Launch includes 100 credits a month, Frontier 250, Mission Control 500.

Around that sit the unglamorous parts that decide whether outbound works at all: reply classification, bounce suppression, sending-domain checks before you burn your address, call recordings and transcripts, transfer to a human mid-call, after-hours handling. Frontier at $79 adds the scheduler, so an interested reply turns into a meeting on your calendar without you touching it, and a second worker.

Five questions to ask before paying for anything wearing this title

  • Does it find the people, or do I bring the list?
  • When a reply says yes, what happens next without me?
  • What happens when they call instead of writing back?
  • Where does the record end up, and what did I have to wire together to get it there?
  • What does this cost at three times my current volume?

Answer those and you can stop worrying about what the category is called. You will know which parts of the job you bought and which parts are still yours at midnight.