Building your AI strategy? Stop.

Sales leaders: You do not need an AI strategy to win in 2026.

There, I said it.

AI is not essential for your success.

Before you tune me out - I was building AI products before it was cool. Before anyone wanted to buy them. AI to analyze video from film and TV and correlate it to anonymized household viewer data (second-by-second, millions of households per show) to determine why people stopped watching.

But we never told the media giants we worked with…

“You need AI or you will not know how to make great television anymore”

Maybe we should have…

Maybe the company would still be around.

But that illustrates the point: creating a false sense of urgency only creates the equation that you never could do your job without AI in the first place.

AI startups are pushing you to adopt AI for things you don’t absolutely need it for.

Can you use it for research on accounts? Yes. We do at TitanX.

Can you use it to summarize calls? Yes. We do at TitanX.

Can you use it to automate outreach? Yes (just not cold calling).

Can you use it to build AI nurture sequences that have a real right to exist alongside human touchpoints like Nick Zeckets? Yes.

Can you use it to make mindblowing messaging like Jordan Crawford? Yes.

Can you use it to make targeting lists straight from Claude Code like Joel Martinez? Yes.

Can you use it to have something to talk about at SKO? Or to try to impress leadership? Or be able to sound elite at cocktail parties? Or give great AI strategy workshops at events? Absolutely.

But do you need to?

Dealing with the (broken) elephant in the room

I am convinced that the current subtext in almost every outbound org boils down to…

“Outbound today is broken and doesn’t work.” But no one has a good alternative.

“The unit economics of outbound don’t make sense.” But no one has a way to fix the math.

So if you’re stuck - is AI the only answer?

It’s not a silver bullet.

This is obvious to most of us. Still, we are left with stuck needles that need moved.

Where to look if you need to fix your outbound?

First principles. If you don’t start at the very beginning of outbound theory, you’ll only end up applying a band-aid (with AI, signals, parallel dialing, or some other trendy tactic) over broken fundamentals.

Here is a list of things you can try before deploying AI:

1. Focus as a tactical advantage: split your reps into teams by tactic. Cold calling. Email outbound. Social outbound. Segment lists so they don’t overlap on contacts, then rotate the lists every 2 weeks.
2. Job description technographics and context signals: Sumble.com is amazing for this. Structured outputs of job descriptions you can deliver into your stack. Turn that into the wedge to open up a conversation. No AI needed.
3. Fix deliverability. More of you than are willing to admit are hitting spam on email outbound. Or marked as spam on the phones, or just low connect rates in general.
   1. For email: Make sure you have a strong DMARC/SPF policy for email outbound (don’t leave this unchecked) for any domain you’re outbounding from. You can outbound from your primary domain if you control for volume.
   2. For the phones: reply to get a pilot of TitanX to solve for high quality phone data. Also make sure you do not call from a single number more than 80 times in one day. Rotate numbers. Don’t parallel.
4. If you’re in a niche/vertical, buy data about companies specific to your space. Search for databases and directories in your vertical and you’d be surprised how many websites and vendors are dedicated to that space. From there, enrich those companies with 3rd party data and identify key contacts. Orbital is great for niche contact data.
5. Message like it matters. Study cold call conversations. Study demo calls. Ask yourself what it’s like to be someone in your prospects’ shoes. Heck, ask prospects directly: which of these lines resonates with you most? What problems are keeping you up at night? AI is a great hack here only insofar as it echoes our ability as humans to empathize. You can circumvent AI easily by… empathizing. Attempt hot takes - they work.

Katelyn Bourgoin, as usual, puts it succinctly:

Why wouldn’t the same be true for sales?

Those 5 ideas are just 20 minutes of hand-crafted words. There are more great ideas in the world. Ask ChatGPT Claude Gemini Grok DeepSeek Manus Magistral whichever LLM is most popular today. (I’m sure you’ll get different answers from each).

If you have other first principles for outbound success, reply to this email and I may add them to the next newsletter.

Make this your Christmas gift to yourself:

“I do not need AI to be successful.” But Evan, won’t I fall behind?

Problems with thinking you must rely on AI to succeed:

1. AI is in its infancy.
2. AI does not mean “AI”. It is not artificially intelligent. The wisest in the space question how far LLMs can go, given lack of control, hallucinations, overindulgence, and zero ethical basis whatsoever are baked into the design. Chaos is a feature of LLMs, not a bug.
3. Which specific AI use case is absolutely essential? Name one. Images? Nope. Video? Nope. Copy? C’mon, a bit of work with a marketing counterpart and you’ll be fine. Data analysis? It helps, but also is heavily unreliable. Research? Maybe, but even then, you don’t always need AI, and you should be able to drop into a conversation cold and find your way through.
4. AI comes with its own risks (company data risk, compliance risk, “oh no we all forgot how to write words ourselves” risk).
5. For every AI startup that is worth checking out, there are dozens that are still great software solutions.
6. What AI does REALLY well that’s hard to replicate is read, understand (or mimic understanding of), and analyze unstructured information. So sure, use it to analyze conversations you’ve already had.
7. It may also be the case that, as Geoffrey Colon puts it, “AI won’t be able to make human sense of every pattern.” There will always be an edge to using your brain well.

Instead, hang your hat on your ability to have conversations with prospects in the first place.

Not on AI to research before them or help you follow up after.

You do not need an AI strategy.

You do, however, need a conversation strategy.

Personally, AI has been great for me.

It has made for some hilarious friend deepfakes that I hope no one at Gemini ever sees.

It has made for some lively “debates” about the end of the world.

Most importantly, it has rapidly accelerated the rate at which I create a graveyard of defunct projects.

Not because I’m not good at prompts, or integrations, or whatever.

But because designing robust systems for long-term use is really, really hard.

So yes hire that GTM Engineer. Buy Clay (I use daily). Mess around with Claude Code.

But do not let your discipline become victim to your imagination.

Execution over ideation.

Always.

What should you be focusing on that isn’t a shiny AI object?

Conversations.

If you can execute conversation volume at scale, you will win next year.

Thanks for reading,

Evan Dunn (LinkedIn)

P.S. If you have questions, hit reply. I read every response.