2010: Cold Visits, No Tools, No Data
In 2010, I walked into logistics offices in Parrys and Mannadi in Chennai with a friend named Rajkumar. No appointment. No introduction. No LinkedIn. Just the two of us showing up cold at companies that had no reason to hear us out.
Most offices had five to ten people sitting in a common hall. If we pitched, we pitched in front of all of them. You could feel the resistance the moment we walked in — that quiet, collective stare of people wondering who let these two in.
We did this for roughly twenty companies. Most said no. Two or three showed real interest. But I learned something that no job had taught me: rejection only feels unbearable until you walk toward it repeatedly. After that, it just feels like weather.
That was outbound in its purest, most unscalable form. No tool. No sequence. No data. Just two people, a pitch, and a willingness to feel uncomfortable in front of strangers. I have been doing outbound in some form for a decade since then — and across that decade, I have watched the entire discipline transform so fundamentally that some of what I spent years perfecting is now either irrelevant or actively harmful.
The World I Came From: Pre-AI Outbound
Timing Was Luck, Not Precision
There is a theory that has circulated in B2B sales for years: at any given time, roughly 3% of your target market is actively looking to buy what you offer. The other 97% are not ready.
I believed this theory. I could not find a better one. And I could not identify that 3% in advance — there was no tool I could afford. So I did what every small outbound operation did: I reached out to enough companies that I would statistically hit the 3% by volume.
When a prospect said "you emailed me at the right time," I used to smile. Not because I was clever. Because I got lucky.
Personalization Was Exhausting and Finite
You could write a genuinely tailored email to ten people in a day if you worked hard. Twenty if you were disciplined. But at a hundred, something broke. The emails started blurring. You were technically personalising but the humanity had already drained out of it.
Most teams chose volume. The sequences got longer. The follow-ups got more aggressive. Response rates crept down. The market became noisier and everyone complained that cold email was dying — when what was actually dying was lazy personalisation at scale pretending to be something it was not.
Warm-Up Was Automated and Getting Easier to Detect
In the early days, automated warm-up tools worked reasonably well. That window closed gradually and then all at once. Modern inbox providers got dramatically better at detecting behavioral signatures — the timing intervals, the precision of click events, the mechanical consistency no human actually produces.
At Tacticalism, we moved to manual warm-up. Real people, real emails, real engagement. The irony: in the era of AI, one of the most important things we do manually is the warm-up that used to be automated.
You Needed Engineering for Everything
Attribution, enrichment pipelines, multi-step automations, data cleaning — all of it sat behind a technical barrier. If you were bootstrapped and non-technical, your ceiling was whatever a marketer could accomplish with spreadsheets and manual work.
What the AI Era Actually Changed
Timing Became Precision
Tools like Clay changed the buying signal question fundamentally. A job posting is a buying signal. A technology install or uninstall is a buying signal. A LinkedIn post where a CEO describes exactly the challenge you solve is a buying signal.
When a prospect now says "you emailed me at the right time," I do not have to smile with relief. I can smile because I knew they would.
Personalization at Scale Became Real — With a Catch
The tools exist now to personalise at genuine scale. Clay can produce a tailored one-liner for every recipient. AI can conduct company research and draft a version of your pitch incorporating what a specific founder has said publicly about their challenges.
But something is still missing. Even technically perfect personalisation can feel inhuman. You can know what company someone works at, what their job title is, what their company recently announced — and still produce a message that reads like it was assembled from parts rather than written by a person who actually thought about them.
The answer is what I call a Personal Narrative Repository — a structured collection of real experiences that AI draws from to add a human layer on top of technical personalisation.
I ran a controlled test: personalised messages to around 200 people, humanised messages built with the repository to another 200. The humanised version outperformed on deliverability, open rates, and reply rate. The human layer is not decoration. It is the signal that tells the recipient this was written for a person, not assembled for a category.
The TAM Question Got a Sharper Answer
The actual countable number of companies that fit your ICP and have the problem you solve — that number determines everything about how you run outbound.
- Under 200 accounts: ABM-style. Every account matters too much to burn. 5 per week with deep research, multi-touch, relationship-building sequences.
- Thousands of accounts: Tiered. Deep work on highest-value, structured personalisation for the core market, intent signals triggering promotion between tiers.
AI made the intent data accessible. But the strategic decision about how to use it based on your actual market size — that remains human work, and most teams skip it.
Workflows Belong to Marketers Now
n8n, Clay, Instantly — the stack that would have required engineering resources three years ago is now accessible to someone who is not a developer, willing to learn, and comfortable with logic rather than code. The competitive advantage large teams had simply by virtue of having engineering resources has narrowed significantly.
The One Thing That Has Not Changed
If your product has not yet proven that real people will pay for it, you should not be running outbound at scale, AI-powered or otherwise.
Automation scales what already works. It does not discover what works. That discovery phase is irreducibly human — the equivalent of walking into those logistics offices in Parrys with Rajkumar, feeling the resistance, learning to read it, and adjusting your pitch in real time because no tool can do that for you.
Validate manually first. Then scale intelligently. The era changes. The principle does not.
What This Means for You
- Know your actual TAM in account count before deciding on a strategy — not revenue TAM, the countable number of companies
- Use intent signals to find the 3% rather than hoping to hit them by volume
- Invest in humanising your copy at a level the tools alone cannot produce
- Do not skip the validation phase — automated outbound cannot tell you whether anyone wants what you offer
- If you're a marketing lead: ask whether your outbound strategy was designed for the market you actually have or the market you assumed you had
The luck is optional now. The discipline was never.
- Pre-AI: outbound ran on luck and volume — 3% theory meant spraying and hoping
- AI changed: intent signals, personalisation at scale, marketer-owned workflows
- What didn't change: validate manually first, then scale intelligently
- Personal Narrative Repository is the missing layer between technical personalisation and genuine human resonance
- TAM in account count determines your entire outbound strategy — most teams skip this calculation
- Manual warm-up now outperforms automated warm-up — the irony of the AI era
Frequently Asked Questions
