A sales email with the wrong year shows why companies are turning to AI agents

A sales rep at a leading public software company sent SaaStr's team a pitch asking about its "end of 2024 goals." The year was 2026. The recipient wrote back to ask if the rep was serious. The rep replied "Hahaha fair enough, you got me there" and made no real attempt to sell.

What went wrong

The first email promised "insights" on buyer intent and AI efficiency but shared none. It listed three product lines and offered to "benchmark" the software. The second asked whether SaaStr was looking at automation tools "for the remainder of 2024." It also offered a menu of six categories, a sign that nobody had looked at the account.

SaaStr's founder, Jason Lemkin, blames the system more than the rep. A step written in 2024 with the year typed in apparently survived two planning cycles. A rep with hundreds of accounts and a quota cannot read every email, so the sequence ends up doing the selling and nobody checks it.

The agent approach

SaaStr says it makes about $20M in revenue this year with 3 humans and 20 or more AI agents. Its outbound agent reads each account record before every send, even the 400th email of the day. A human team at that volume cannot, and a standard sequence tool never checks.

The post admits a limit. An AI sales agent given only a name and a title writes email just as generic as the 2024 one, and sends 50 times more of it. SaaStr says it built its own account tool on its own data for that reason, rather than relying only on vendor data.

Buyers are using agents too

The post says AI agents are now a top-5 source of SaaStr's web traffic. An agent called 10K, which SaaStr calls its AI VP of Revenue, evaluates vendors for the company. In one case it shortlisted a vendor, ran the numbers and dropped it within 12 hours. It cited premium pricing with no conversion data, multi-month minimums and scarcity framing.

Lemkin expects that an email with the wrong year will soon be archived by an agent sorting an inbox. The rep would only see another non-response in his report.

What to do

These figures and claims come from SaaStr itself, a company that promotes AI agents, and were not independently checked. Still, the lesson holds for any business that sends mass email.

• Reread old email sequences for dates and offers that are out of date.
• Make sure a person or a tool checks each message against the account before it sends.
• Name something specific about the recipient, such as their own product or costs.
• Expect that some buyers will use AI to screen out generic pitches.