My Take on Sales Autopilots Versus Copilots
Between AI autopilots and AI copilots, which is more likely to dominate the future of sales?
The topic of AI autopilots versus AI copilots has been a hot topic over the past few months.
After having spent almost 2 years building AiSDR for sales teams, I’ve noticed something interesting.
It may come as no surprise that from where I stand, the loudest voices predicting the collapse of autopilots – and championing a future ruled by copilots – are primarily the companies building AI sales copilots.
Let me share why I believe this perspective might be a bit short-sighted, and why autopilots are likely to become the dominant tool despite their current limitations.
Understanding autopilots and copilots
In case you’ve never seen these terms before, here’s the quick rundown of each one:
AI copilots
Copilots are an AI-powered sales tool that helps you research leads and write personalized emails, but won’t send what it writes. You can think of them as an AI assistant that handles the prep work while you maintain control over the final version and delivery.
Some examples of AI copilots are Copy.ai and OneShot. You use them to generate your content, but you’re the one behind the send button.
AI autopilots
Autopilots take this a step further. They are an AI-powered sales tool that researches leads, writes personalized emails, and sends them, as well as reads and replies to incoming emails.
The question, and concern of many, is that AI autopilots aren’t good enough at matching a person’s ability to create high-quality sales emails that convert leads into demos.
Instead, they claim that many AI SDRs rely on “spray and pray” outreach tactics where they flood lead inboxes with low-quality, template-feeling emails and burn your TAM.
Are autopilot critics right?
Now here’s the shocker.
I believe that everything autopilot critics are saying is true… For now.
Sure, today’s autopilots are struggling to consistently match human-quality outreach at scale. They require constant tweaks and fine-tuning, especially when LLMs release their latest updates. And at AiSDR, we take careful steps to make sure AI doesn’t burn through your leads.
At the moment, copilots offer a better balance of AI efficiency and human judgment for high-stakes sales conversations.
But it’s my belief that AI copilots are a product for today. Not for 2-3 years from now.
That said, look how far AI capabilities have come in ~2 years. GPT-4o and Claude 3.5 Sonnet are waaay more advanced that GPT-3 or even GPT-4.
And they’re just getting better with the release of Claude 3.7 Sonnet and soon-to-be GPT-5.
As a result, while today’s autopilots aren’t yet a “cure-all” for all types of broken outreach, they’re already effective at replicating proven GTM plays, such as reaching out to:
- Product champions who change companies
- Companies that are hiring for a role similar to your solution
- Website visitors to high-intent pages who didn’t book a demo
They also solve slow response times by handling incoming emails in a timely manner.
What’s in store for AI in sales in the future?
Given this trajectory, ask yourself this: Do you really believe that in another 2 years, AI won’t be capable to researching prospects and synthesizing data into strategic, well-crafted outreach for multiple channels?
I think they will be.
We’re witnessing billions in investments specifically targeting improvements to AI’s ability to understand context, make nuanced decisions, and create personalized, humanlike communication.
If you’re betting against AI’s ability to run research and outreach, you’re basically saying, “I believe that everything we’ve done in the past 2 years in AI is the best we can do and it won’t get any better, and AI will always require human oversight despite the billions of dollars being invested.”
History suggests this is unlikely.
Building copilots in 2025 is like releasing the Nokia N97 or the Blackberry Tour after the original iPhone came out.
They were probably thinking that big screens were good, but they’d never be as reliable as buttons or physical QWERTY keyboards.
And at the time, they were right.
But flash forward a few years and answer the question, “How did that work out for them?”
Autopilots vs copilots
AI autopilots are poised to become the dominant method of outreach for a simple reason: They align with how humans naturally seek to optimize their work.
After all, when given two options that produce similar results, people inevitably choose the path requiring less effort (unless your name is Robert Frost).
When the iPhone launched, BlackBerry defended physical keyboards as essential for “serious” business communication. Early touchscreens were indeed inferior, but improvements rapidly turned this advantage obsolete.
Similarly, autopilots aren’t quite there yet, but they’re improving quickly.
Forward-thinking organizations are preparing for this shift already, testing autopilot systems alongside current processes and adding GTM engineers to their team who can hone their skill sets and develop frameworks for driving AI results.
With technology like artificial intelligence, you have to think about what will happen 2, 3, 5, or even 10 years forward.
