The floor keeps rising

I found a note from 2018 this week. It said: “If you’re using automation, you’re now average.”

I had left it as a post idea and never written it. Seven years on, it feels more relevant than when I wrote it - and more complicated.

By 2018, Outreach, HubSpot and countless other sales and marketing automation tools had gone mainstream. Having nurture sequences and lead scoring stopped being a differentiator. It became the floor. The early movers had extracted their edge, and everyone else had caught up.

What followed was not the promised future. Personalisation at scale - the right message for every buyer at every moment - mostly did not materialise. What we got was volume. More emails, more nurture tracks, more triggered workflows. Inboxes got noisier. Engagement fell. The tools that were supposed to make marketing feel more relevant made it feel less human.

The companies that came out ahead were not the ones who automated most aggressively. They used automation to do fewer, better things. The tool became table stakes. The judgment behind it became the edge.

We are still early in AI adoption. It feels like everyone is using it, but most are barely scratching the surface. The place where AI has embedded fastest - sales email automation - is telling. It happened quickly because it was easy to bolt onto existing processes. But if my inbox is anything to go by, it is surprisingly lacking in judgment. The volume is up. The relevance is not.

That will change. As platforms combine signals, get better context, and redesign their processes with AI at the core, the quality will improve. And then a familiar dynamic kicks in: AI will increasingly screen out the emails it helped write. The arms race that automation started is about to get a lot more automated.

The parallel with 2018 is there, but it breaks down in one important place.

Automation replaced processes. The ceiling was always the process itself. AI can help you think, not just do. Scale is solvable with the right technology. Judgment increasingly is too. What cannot be automated is deciding what is worth doing in the first place.

In 2018, I was right that automation made everyone average. What I missed was what the next level would look like - and who would get there.

Instantly analysed billions of cold email interactions for their 2026 benchmark report. The platform-wide average reply rate is 3.43%. Roughly 95% of cold emails get no response.

Instantly Cold Email Benchmark Report 2026

That is what the automation era produced in B2B outreach - and AI is accelerating the same dynamic. The tools solved for volume. Nobody solved for relevance.

Ask this question about each thing you are using AI for: is this helping us do what we already do faster, or is it helping us decide what is actually worth doing?

Most AI usage right now sits in the first bucket. That has real value - efficiency compounds. But the companies that got the most from automation were not the ones who ran the most workflows. They were the ones who used the time and data those workflows created to make sharper decisions.

The equivalent question for AI is not “are we using it?” It is “what are we learning from it, and what decisions is it changing?” If the answer is none yet, that is the gap worth closing.