Open your inbox and count the cold emails from the last week. Now count how many you read past the first line. If you buy anything for a living, the second number is probably close to zero, and you are not unusual.
Outbound used to work on a simple, unwritten contract. The seller did the homework, and in return the buyer gave them a few minutes of attention. The seller earned the right to interrupt by showing up with something relevant. That contract is broken, and the sellers broke it.
How we got here
AI outbound tools made it close to free to send a thousand "personalised" emails a day. Every company in the market bought the same tools, plugged in the same data providers, and pointed them at the same list of VPs. The emails mention your latest funding round, a podcast you were on, or your job change. Buyers have seen every variation, and they can spot the template within a sentence.
The result is noise. Reply rates fall, so teams send more. Domains get flagged, so they buy more domains, while mailbox providers tighten the rules: Google's sender guidelines ask senders to keep reported spam below 0.1 percent. The cost of sending went to zero, but the cost of being ignored keeps climbing. Each bad email also makes it harder for the next seller, including the good ones, to get a hearing.
If everyone can send at volume, volume is no longer an advantage. The advantage goes to whoever can create signal: outreach that proves a human thought about this specific buyer.
What signal looks like
Signal is anything a buyer can tell was expensive for you to produce. Expensive in time and thought, not money. Here are the approaches I have seen work in enterprise deals.
1. Insight they can use without buying from you. Send a short, specific observation about their business. Compare their checkout flow to two competitors. Point out a gap in their hiring plan against their stated product roadmap. Share a benchmark from companies of their size. If the buyer would forward it to a colleague, it is signal. If it only makes sense as a lead-in to a demo, it is noise.
2. Handwritten notes. It sounds old-fashioned, and that is the point. A handwritten card to a senior buyer, with two lines that reference something real, almost always gets opened and often gets a reply. Ten of these a week, to accounts you have actually researched, will outperform a thousand automated emails.
3. Small, curated events. A dinner for eight people in the same role, in the same city, with a useful topic and no pitch. Buyers want peers more than vendors. Host the room well, make the introductions, and the relationship starts on your terms. Follow up with something from the conversation, not a deck.
4. Warm paths, done properly. Ask customers, investors and former colleagues for a specific introduction, with a reason the buyer should care. "Would you introduce me to your CMO because we helped a similar team cut their content launch time in half" is a request people can act on. "Can you connect me with anyone relevant" is not.
5. Fewer accounts, deeper work. Pick 25 to 50 accounts that genuinely fit, and learn them properly: who the buying committee is, what changed recently, what they have tried before. Give each account the attention you used to give a whole territory.
A practical weekly rhythm
For a seller or founder running outbound, a week built around signal might look like this:
- Monday: research five accounts in depth and write one specific insight for each.
- Tuesday: send those five insights by email or LinkedIn, from a real person, with no attachment and no calendar link.
- Wednesday: write and post ten handwritten notes to senior buyers you have already researched.
- Thursday: ask two customers or investors for one named introduction each.
- Friday: review what got a response, and why. Keep what worked and drop what did not.
That is roughly 20 to 30 touches a week. It will feel slow next to a tool that sends 1,000. But measure meetings with the right people, not activity, and the comparison changes quickly.
Where tools still help
I am not arguing against technology, though I am sceptical of AI SDRs in enterprise sales. Use it to research faster, to find who changed jobs, to keep account notes in order, and to remind you to follow up. Just stop using it to fake the part that has to be human. The research can be automated. The judgement about what this buyer needs to hear cannot.
Rewriting the contract
Buyers have not stopped buying. They are doing more of it on their own: in Gartner's 2026 survey, 67 percent of B2B buyers said they prefer a rep-free buying experience, and 45 percent had used AI during a recent purchase. They have stopped listening to people who did not do the work. The sellers who win the next few years will be the ones who go back to the original deal: earn attention by bringing something worth the buyer's time.
If your pipeline depends on volume, now is a good time to test the alternative. Pick ten accounts this month and try it.