If you were forwarded this newsletter, join 2,100+ weekly readers—some of the smartest GTM founders and operators—by subscribing here:
Hey, y’all!
We had snow here in Austin overnight, so I got to do snow angels with my 2-year-old (and by snow angels, I mean, there was dirt after he moved his arms and legs a couple of times lol). But hey, he loved it—and so did I—so that’s all that matters!
Today, I wanted to throw out an idea I keep thinking about. I don’t have the answer, so, if you’re building something in this space, or using something, please let me know. I’m excited to see the new crop of sequencing tools evolve in the AI era.
My hope is that it creates more efficiency in the marketplace (in other words - less spammy sales emails).
Alright, let’s get into it.
Sequences 2.0 (AI-First)
Sequences are not the right form factor in the AI era. I think we'll see them slowly die in the coming years.
Sales sequences were created by first movers like ToutApp and InsideSales back in 2011. And then second movers, like SalesLoft and Outreach, came in shortly after that.
The concept was basically: take what marketers had (with tools like Marketo, HubSpot, etc.) and put that in the hands of sellers. But, instead of sending from a shared IP, the emails would go out from the company/reps' gmail server (outbox).
Send an email. If it doesn't get a reply, have another email go out. If it does get a reply, the next email (in the “sequence”) will not go out. It helped sellers do a lot more activity in way less time.
Ten years ago—even five years ago—using a sequencing tool for automation gave sales teams an advantage.
Today, it's clear that sending mass emails doesn't work like it did in 2019. (By the way, I'm not saying you shouldn't use a sequencing tool today).
But, the other thing that has changed is the advent of (Gen)AI/LLMs.
It no longer makes sense to create a new sequence for every new signal that you want to kick off a sequence/play from.
I felt this first-hand at Apollo, where I had 85+ plays/sequences running concurrently. The reporting was brutal. The admin was worse. (But, the level of automation was insanely powerful and effective.)
Here's an example to highlight the problem:
Let's say a contact matches 3 “signals" that you care about:
1. Job change
2. Website visitor
3. Relevant technology recently implemented
Today, you'd have to build logic into the workflow to tell it which sequence you want to add the person into (Job change sequence? Website visitor sequence? Or the technology sequence?).
Instead, here’s what SHOULD happen with AI:
1. Signals that you care about are tracked
2. When one or more relevant signals are detected, bubble up that person or company because the signal(s) indicates it's time to reach out to have a sales conversation
3. All of the data/signals on the person (and on their company), including first-party data from the CRM/DW get thrown into a prompt for more context (eg: closed-lost deal notes, previous contacts we’ve engaged with at the company, product usage data, etc.)
4. An email is generated based on all of the context
With this new method, EVERY single email is different.
Then, "Step 2" in the sequence should be dynamic, based on if/how the person/company engages with Step 1. (This doesn't happen in any sequencers today).
For instance, have new signals been detected? Maybe someone *else* at the company visited the website. Or maybe someone talked about a relevant topic on social since Email #1 went out. Etc. These newly detected signals (since Email #1) should be incorporated into the future steps of the sequence, dynamically, using AI’s reasoning abilities.
•••
The point is, LLMs are incredibly powerful. But, the way Sequences work today limits their capabilities.
I think someone will have to create this new form factor from the ground up to truly innovate (sorry @Incumbents). This is an example where skeuomorphic design thinking currently limits the potential of AI GTM tools.
And I think in the future, we'll look back on "Sequences" as archaic. But, a necessary stepping stone that got us into a magical future where true value can be added at every interaction.
Which will create: more efficiency in the (software purchasing) marketplace aka less spam. And, that’s a future I’m excited to build towards.
Thank you for your attention and trust.
See you next time,
Brendan 🫡
I think Roomie AI from Common Room pretty much cover 1-4 already? Have you tried it B?
https://www.commonroom.io/product/ai/
I dig it, B ... what you're talking about is contextualizing the signals. In today's world, things are pretty linear, as you describe. Get a signal, send a message. But there are so many other factors at play that we now can incorporate to get the rep more context so they can more acutely customize their messages. Eventually, tech will do all of this (sooner rather than later) because I always say prospecting is about timing and relevance. Humans, with all of the activity they have to do, can't manage to stay on top of the timing, and it's just as hard to stay relevant!