The Tools Got Better. The Outreach Got Worse

There's a version of this story that gets told as progress. Better tools. Faster execution. AI-generated personalization at scale. Sequences running on autopilot while the team focuses on closing. On paper, it looks like efficiency. In practice, most of the people on the receiving end of it can tell exactly what happened.

The tools got better. The outreach got worse.

I've been in and around sales and marketing for more than two decades, long enough to remember when relationships were built differently. When understanding a potential client meant actual research, actual conversation, actual investment of time before expecting anything in return. That approach wasn't just nostalgic courtesy. It produced better results because it created real context before the ask.

What changed isn't that relationship-building stopped being valuable. What changed is that the pressure for speed created a shortcut culture, and AI handed that culture a megaphone.

Account-based marketing (ABM) is a good example of how this plays out. In a boardroom under growth pressure, it sounds compelling. Focused. Targeted. A strategic play that signals discipline rather than spray-and-pray. And it can be all of those things.

But many companies approach it with a quantity-over-quality mindset, which is precisely what ABM isn't supposed to be. They get excited about an automation tool, upload a contact list from a target account, pull together just enough surface-level information to call it targeted, then use AI to generate outreach at scale. The sequence runs. The metrics track. Nothing meaningful happens.

I saw this directly with a sales rep on our team. Not a criticism of him. He hadn't done anything wrong by his own understanding of the task. But the emails he'd built were awful. Inserting a first name isn't personalization. Mentioning that someone uses a competitor isn't relevance. A sequence of unsolicited emails that shows no real understanding of the recipient isn't targeted outreach. It's automated noise with a veneer of effort over it.

What he'd been given was a workflow, not a strategy. Build the target account list. Use ChatGPT to write the emails. Load it into HubSpot and set it to run. Follow the prospects on LinkedIn and like their posts in parallel.

That's activity. It has the shape of a strategy without any of the substance.

The problem isn't that AI was involved. The problem is that no one had defined what the outreach was actually supposed to accomplish, what a good message looked like, what would make a recipient feel like the sender understood their world, or what the guardrails were for protecting the relationship before it had even started.

The deeper risk with AI-assisted outreach isn't that it produces bad emails. It's that it scales bad judgment efficiently.

A salesperson can be excellent at demos, relationship development, and closing while still having no real understanding of what happens when automated outreach starts eroding trust before they ever get a conversation. Those are different skills. Most teams don't develop the second one before deploying the tools that require it.

The result is what most people experience daily. Generic emails written to sound personalized. LinkedIn messages that claim to be tailored and clearly aren't. Outreach that's slightly polished, slightly targeted, and completely forgettable, arriving in volume because the tools made volume cheap.

The answer isn't to reject the tools. It's to use them for the part of the process where they actually add value, which is research, not messaging.

The shift worth making is using AI to understand the person and the company more deeply before the message gets written, not to write the message instead.

There's more accessible signal than most teams use. What causes does the person engage with publicly? What content are they sharing or commenting on? What events, communities, or conversations are they active around? What has the company announced recently? Where does the business appear to be heading? What language does the person actually use when they talk about their work?

None of this requires anything proprietary. It requires effort. And that effort: organizing, synthesizing, identifying what's actually relevant, is exactly the kind of work AI accelerates well.

A more useful workflow looks like this: start with the company and ask AI to summarize what the business does, what's changed recently, and what might be relevant based on the role you're targeting. Then move to the individual. Review their public activity and ask AI to surface themes, priorities, and likely points of genuine connection. Use that to build a brief, not a draft email. Something that helps you think more clearly about why this person, why now, and what would actually be worth saying.

Then pressure-test the angle before sending anything. Ask whether the message feels generic, presumptuous, or weak in relevance. Ask what would make it more useful to the recipient. Ask whether the timing even makes sense given what you know.

That's where AI starts earning its place in the outreach process. Not by writing the message, but by helping you earn the right to send one.

When relationships were built more slowly, what made them work wasn't just the time invested. It was the signal that investment sent: that the other person was worth understanding before being sold to. That their context mattered. That the conversation was starting from genuine curiosity rather than a quota.

Those values didn't become less important when the tools got faster. They became harder to demonstrate, which means they became more differentiating for the people who still bother.

The goal of outreach isn't to sound personal at scale. It's to actually be personal. And those are not the same thing. One is a production problem. The other is a judgment problem. AI can help with the first. Only the human behind it can solve the second.

Want more practical approaches like this? Explore my curated library of AI tools, prompts, and workflows at resources.taneilcurrie.com