I had fourteen founder calls last week.
At least six of them showed me the same broken plan.
Skip the SDR.
Try anything else. AI on the phone. A commission-only AE. An offshore agency on retainer. Yourself, between sales calls and demos. Just do not hire and manage humans who dial all day.
Every variation breaks the same way.
Five real examples from the week
One founder runs a lead-gen agency for high-ticket home services. He built an AI caller in-house. The show rate is good. The book rate is good. The answer rate is destroyed by carrier-level spam filtering. His words, almost exactly: "calls are going to go through much better just based off spam and all these different things." That is why he was on a call with me. The AI he built cannot get the phone to ring.
Another founder runs a marketing agency for wedding venues and golf courses. He and his partner sit in the chair for three, four, five hours straight and dial. Then they do not dial again for a week. He told me, "Cold calling is always the thing that gets pushed off. Because it is not urgent. But if you do not, how are you going to book your appointment?" Two implosions in their last two hires. He is not undisciplined. He is running every other part of the business.
A digital marketing founder told me he hired someone who was, in his words, really good at sales. The rep closed one deal in his first hour. Then never touched the system again. Commission-only. No call activity. No follow-up. No nothing. The founder is back to doing it himself.
A SaaS founder ran a Pakistan-based agency for seven months. He could not get them to take ownership of anything. "They were not able to bring enough calls to my plate. But they were more depending on me to explain the services. My intention was to have a hands-off experience. The closing thing was done mostly from my side rather than them taking ownership."
A sales coaching founder had one of his AEs doing the entire cold-calling function. The wrong person on the wrong activity. Three or four meetings booked across two months as their entire output. No connect rate data. No dialer beyond Apollo. No follow-up sequence. The math was invisible to them because nobody was running it.
Same pattern every time
The founder reads about AI replacing SDRs. Or watches a podcast where someone claims their agent is closing six-figure deals at eleven on a Saturday night. Or reads a LinkedIn post about how cold calling is dead. They try to skip the structured, boring, expensive part. Hire one person on commission. Build an AI. Find an offshore agency that promises three meetings a week for eight hundred dollars a month.
None of it survives contact with the actual dialer.
The job that needs doing in 2026 still looks the way it looked in 2016. A small team. Real phones. A list. Time on dials. A manager watching connect rate, talk time, and meetings booked. A coach listening to recordings. A discovery script that does not embarrass the brand. Follow-up that goes out the same day, not when someone gets to it.
It is not the most exciting investment a founder can make. It is the one that produces pipeline.
One piece worth your time this week
Jason Lemkin published a piece on SaaStr this week called "A Great Year With Our 20+ AI Agents. But a Rough Week."
He is honest about what happened. One of his outbound AI agents decided, on its own, to run an A/B test. The "B" variant gave away free tickets to SaaStr Annual that nobody authorized. Cost him two thousand dollars before guardrails caught it. Another agent promoted SaaStr London 2025, an event that had already happened. A vendor pushed a hot fix and broke a workflow he had built over months. A workspace container would not load and the support advice was, in writing, to "let it sit overnight and see if it sorts itself out."
Read it. It is the most honest thing I have seen written this year about what running AI agents in production actually feels like.
Where I agree with Jason: the math at SaaStr is real. He went from twelve humans to 2.5 humans plus 20 agents and the work is still getting done. The 24/7 follow-up is real. The deliverability lift on Artisan is real. Outbound that nobody was getting around to is finally getting touched.
Where I push back: SaaStr is a fifteen-year-old brand with twenty million words of training material, a category-defining founder, every transcript and email and sponsor conversation already in their data warehouse, and Amelia spending fifteen to twenty hours a week running the agents. Almost none of the founders I spoke with this week have any of those ingredients. Jason himself says it explicitly in his own playbook: AI SDRs scale what is already working. They cannot fix what is broken. Most founders are trying to use AI as the fix. That is the broken plan I keep seeing on calls.
Build the human dial team first. Get the messaging working with humans. Get the data clean with humans. Get the connect rate, the objection map, and the discovery flow proven with humans. Then layer AI on top to scale what already works. In that order. The order is doing all the work.
If anything here landed, forward it to a VP of Sales who would push back on it. Reply if your week looked anything like the ones I described above.
Justin