I took ten founder calls last week.
Elevator service contracts in NY. AI automation agencies. Self-storage SaaS. Pickleball ad sales. SEO for plumbers. Voice-of-customer tech. AI marketing for salon owners.
Same story almost every call.
The reason they were on the phone with me was almost never "we need more leads." It was: "the last sales hire did not work and I need to figure out what to try next."
So I dug.
Four founders, four costumes, one mistake
One founder runs SEO for home services. He has built a complete sales stack. Phone Burner dialer. AI call recording. Transcripts. Coaching. 30-second contract generation. He found one closer who, in his exact words, "closed 1 deal in 1 hour, then stopped working." Commission-only. No base. The rep won on a single dial and never came back.
Another runs elevator service contracts in NY, NJ, and CT. A couple of 1099 commission-only reps on the books. Target of 25 to 50 sales over six months, which is not a crazy bar. Reps not generating consistent effort. Same comp structure as the SEO founder. Same outcome.
Third one runs an AI automation agency. He came in expecting his appointment setter to cost $12 an hour plus $50 per booked meeting. Do the math on that. At that rate the setter has to book 28 meetings a month before they clear $2,000 in total comp on a 40 hour week. If they do not book, they do not eat. So they do not dial. So they do not book. So they do not eat.
Fourth ran an AI voice-agent shop. Used a Pakistan-based agency for seven months. His exact words on why he cut them: "lack of proactiveness, ownership, accountability."
Four founders. Four different setups. One thread.
The mistake is the comp plan, not the headcount
None of those reps "did not work." The structure they were placed in did not work. Commission-only. 1099. $12-an-hour-plus-bonus. Offshore-and-forget. The same lie in four costumes.
Sales is not a coin flip. A rep on commission only is paid like sales is a coin flip. If they win, they get paid. If they lose, the company loses nothing that month. The founder thinks they have offloaded the risk. They have not. They have just moved the cost to month four, when there is no pipeline and the rep is gone.
This is not a hot take. It is what Jeb Blount has been writing for a decade in Fanatical Prospecting. The reason most outbound dies is not that the script is bad. It is that no one is actually pulling the trigger every Monday at 7am. And the fastest way to guarantee no one pulls that trigger is to pay them only if they hit it.
The boring playbook
The fix is not exciting.
- Pay a base. Pay it on time.
- Set a daily quota that is a number of dials, not a number of deals.
- Sit on the calls.
- Coach the calls. Replay them the next morning.
- Replace the rep in week two if they cannot hit the dial quota.
- Keep them and pay a bonus on results if they can.
That is the entire first 90 days. None of it is a tool. None of it is a sequencer. None of it is an AI SDR. It is a base salary, a daily activity floor, and a manager who actually listens to the recordings.
It is also why the same founders keep showing up on my calendar with the same problem. They have tried every flavor of "do not pay them unless they perform" and every one of those flavors guarantees they do not perform.
One piece worth your time this week
Jason Lemkin published a piece on SaaStr on May 17 titled "Schmoozing Is Dead, Agents Are Hitting 120% of Humans, and Growth Is the Only Thing That Matters."
His take after wrapping SaaStr AI Annual: schmoozing on Zoom is dead, the SDR-as-receptionist role should go extinct tomorrow, and AI agents in some slivers of GTM are already at 120% of humans. He says Qualified set 682 meetings for SaaStr in the run-up to the event, with zero quota-driven sandbagging. SaaStr now runs 21 agents and 3 humans in their GTM org and has crossed $1M in AI-driven closed revenue.
His framing on the SDR who cannot answer the question is exactly right. The kid named Jason who graduated from San Mateo Community College and is there to qualify you for a meeting with Bob next week. That role is actively destroying pipelines. Kill it Monday.
Where I push back: agents are not at 120% of humans for outbound to anyone you actually want to sell to. The ten founders who called me last week tried the cheap automated low-accountability version of outbound first. It is the reason they were on a call with me. Lemkin's bigger point in the piece, that you need a 10x feature to bother, is the real signal. If your product is not differentiated, no rep, human or agent, fixes that. If it is differentiated, a trained human picking up the phone with full context will outwork any AI SDR on a deal worth more than $5k a year. Both can be true at the same time.
That is the read for this week.
If anything in here landed, forward it to a founder who is about to hire their first BDR. They will need the boring playbook before they buy the shiny one.
Justin