Welcome to the bot wars

What’s rapidly coming to change the landscape of b2b outbound prospecting is already here in the job market and it’s…fraught

?

A brief reminder about growth hacking

Not long ago, in a masterful stroke of branding, we saw the emergence of a category of professional service that looked a lot like demand generation but came to be known as “growth hacking.” If you’re unfamiliar, I most like the definition from OptinMonster which defines growth hacking as strategies businesses use “to rapidly increase their user base and revenues with minimal expense.” Nice!

?

As you can see, the concept has waxed and waned over the years but never really recaptured the imaginations of business leaders in search of growth like the epic growth hacker summer of 2015. And the results of a small, informal poll of some colleagues concludes that using the term today elicits eyerolls. Why, I asked: “we’re all trying to rapidly increase revenue with minimal expense, and a growth hacker just sounds like someone doing it in cyber punk outfit.” Fair enough!

?

But while the name may have fallen somewhat out of favor, (I see burgeoning adoption of term “growth ops” in its place) the game is alive and well. In fact, people like Will McCartney have outlined complete tech stacks that allow individuals to identify, prioritize and reach (we’ll return to the problematic nature of “reaching” momentarily) thousands of potential prospects for a negligible cost and time investment.

?

The process is actually quite straightforward, though easier said than done:

1. Use data to identify which potential prospects within your total account universe are in-market (“showing intent”)
2. Engage these prospects across multiple digital and physical touchpoints
3. Craft context-aware messaging to personalize this outreach
4. Sustain this type of personalized, context-aware engagement through the buyers journey unless explicitly asked to stop, the prospect is disqualified or the opportunity closes
5. Scale

?

And what’s capturing the imaginations of the modern growth ops thought leader today is the potential to sequence these different technologies together, managed by AI Agents that do the heavy lifting.

?

Let me give you an example of an out-of-the-box workflow you can acquire from FlashIntel :

1. FlashIntelAgent scrapes the attendees of Linkedin Live Events
2. Attendees data is enriched and an email/cell phone number is appended to the record
3. Based on title, company, etc some number of attendees are enrolled as prospects
4. AI Agent likes and leaves a comment on a relevant linkedin post made by the prospects
5. AI agent sends an email and an SMS and a voicemail drop
6. Human growth ops orchestrator takes action when prospect raises their hand

?

Let’s acknowledge a few things here:

1. This is no different than what the best human SDR/BDR/outbound sales reps are doing today
2. These types of tactics work to push demand and while none of us like cold outreach when it doesn’t solve a problem for us, we don’t mind being bothered at all when it does
3. The AI Agent experience is still pretty poor and the messaging rather ham-fisted
4. Scaled outreach very quickly veers into spam

?

To anyone close to b2b demand generation, the whole infrastructure feels limited by the law of diminishing returns and therefore unsustainable.

This is for two reasons:

• First, because these tactics have a shelf life and while you might be able to get that juicy decision maker with budget to read your voicemail drop now, once they’re getting 10-15 voicemail drops per day, they’re going to tune it right out.
• Second, because we all know when we’re interacting with AI that’s pitching us; it just isn’t contextual enough, especially when trying to connect the dots between complex Enterprise-grade software and the types of problems buyers of those solutions are dealing with.

?

So how this evolves is very interesting, especially as you attempt to move upmarket where the engagement opportunities are fewer and the stakes for each touchpoint are higher.

But we don’t need to speculate too wildly about what’s coming to b2b demand generation because we can already see it in the job market: an arms race of AI-enabled workflows that don’t drive the outcome both parties involved want and actually creates an adversarial relationship.

What's happening now in the job market

Remember what I said before about needing to marry intent, context and personalization, and how that’s hard for today’s AI to do with the same panache as a human, because the problems, solutions and current circumstances are complex? Well contrast that with an open requisition:

1. Intent: (presumably) the hiring company is in-market by virtue of them posting a job description
2. Personalized, context-aware outreach: the job description spells out in painstaking detail the exact success criteria by which a job seeker will be evaluated
3. Ease of engagement: no need to hunt around for contact details, there’s a webform, potentially a recruiter and sometimes a hiring manager explicitly or implicitly named by their title in the JD

?

The above is from the perspective of a job seeker, but the same conditions exist for the ?employer as well. Unless you are a business owner, you’re always passively if not actively “in market” for a better job. Most employees have encapsulated their work experience into a formalized linkedin profile and/or CV, making context-aware, personalized outreach very simple.

?

In other words, this is the type of interaction that current-state AI was born for. And it’s evolved incredibly quickly:

• 2023 state of the art: “ChatGPT, I’m going to paste a job description and I would like you to rewrite my resume and a draft a cover letter that perfectly aligns to this role.”
• 2024 state of the art: “for $129/month lazyapply.com will submit up to 750 job applications per day(!!) across all major job boards.”

?

Aki Ito at Business Insider wrote a great article about using LazyApply , as did Caitlin Harrington at WIRED last fall. But you don’t need these articles for more than the color commentary. You can ask anyone looking for a job or tasked with filling roles: it’s a nightmare because the primary channel for connecting two interested parties has been completely decimated by AI-enabled tools.

?

I’m not even sure you can accurately describe it as a zero-sum game. It’s purely no-win:

?

As recruiters receive more applications largely indistinguishable from the next they can either:

1. Manually review a very small subset of applications
2. “Leverage AI” to help prioritize which candidates are reviewed

?

As job seekers look to meet the moment they can either:

1. Try to get their application submitted as quickly as possible, so the manual reviewer sees it before stopping the review
2. “Leverage AI” to try and game the applicant management platform’s AI filtering through hundreds or thousands of applications

?

In other words, there is a disincentive to thoughtful, well-crafted bespoke job-seeking and recruiting because it’s now just a numbers game.? And anyone who deals with conversion metrics knows, if you cannot improve your conversion rate then you have to increase your at-bats to hit your number.

?

Go and see for yourself. A job seeker has about 1-2 hours from the time a position hits linkedin before it’s received “over 100 applicants.” But if the job is remote, or it’s a name-brand company, the window is even shorter.

?

What does all this mean, practically speaking? In my opinion, that the majority of good jobs are going to get filled the way they always have: by personal referral to a hiring manager. No AI technology assistance required.? Meanwhile, job boards, recruiting platforms and application bots will continue operating a robust ecosystem of AI-to-AI interactions rarely involving the humans that turned them loose on each other.

Back to b2b

?

The “growth ops” tech stack that exists today, and the rapid evolution of AI agents to drive scale exists because the demand for these types of solutions is strong right now. The early adopters understand the use-case and are paying for access because these emerging technologies look a whole lot like what they’ve been using for years, just bigger, faster, more scalable.

?

What’s coming soon is the response to all this scalable outreach: the AI personal assistant that thoughtfully gatekeeps all the cold linked in requests, cold emails, cold texts, cold voicemails. Diligently unsubscribing you and, presumably, occasionally requesting more information, booking discovery calls or sending the email that, “this sounds interesting but the timing’s not right. Can you follow up in about 6 months?”

?

And it’s not hard to imagine your AI agents attending webinars on your behalf and summarizing them, or visiting solution providers websites to chat with their RAG-enabled chatbot as part of a software evaluation.

?

Right now CrewAI , LangChain and various name-brand LLMs are having a moment and one of the more prevalent use-cases is a multi-agent workflow to analyze and respond to your emails. For example, Sam Witteveen recently published this tutorial and Brandon Hancock published this one. ?

?

What’s missing today is the UI that will make configuring these types of multi-agent AI workflows accessible to people without python chops. But it’s a virtual certainty that some sort of no-code or turnkey solution is imminent. In fact, all the major consumer electronics brands have been telegraphing that they want their next batch of devices to be built around the sort of “personal assistant AI” that will make the orchestration of your email, calendar, work and personal files more seamless.

?

What does all this mean, practically speaking? In my opinion, that the majority of b2b deals are going to get done the way they always have: through personal connections between a human buyer and a human seller.? But surrounding those instances of meaningful partnership there is going to be a lot of AI-generated activity being hurtled into the void. Welcome to the bot wars!