Right now you can’t escape the press on GenerativeAI. People who had never heard that phrase 5 months ago are now ‘experts’ on LinkedIn writing lists of companies that are going to replace all your workers. Your VCs are probably asking how you’re making use of AI in your board meetings. And your CEO is looking at where they can grow faster while saving money by automating huge parts of the business.
Is this where the world is going? Yes…but. There are now digital SDRs who are booking meetings with very senior decision makers who may never realise they are speaking with a bot. AEs are using GenAI to create copy–often better and more succinct than what they write themselves (totally separate rant about people needing the ability to write and communicate well for another time and hopefully not immediately outdated). Presentations, marketing copy, website images, SEO management–all are being created faster by AI.
So, with the assumption that we still have a while before it’s all bots talking to bots until the very end signature, and that people still want to buy from people that they trust—where do humans come in and create an edge for companies? How can we make the best of GenerativeAI and get our teams to make us stand out, instead of blend in? And, is the entire premise of the sales motion going to change…back to what it used to be, humans talking to humans?
Thank you to some of the smartest people working in GTM today who contributed extensively and whom I quote and paraphrase here.
Jacco van der Kooij, Founder of Winning by Design
Stanton Cagle, Senior Digital Marketing Manager, Unbabel
Doug Landis, Growth Partner at Emergence Capital
Note–I won’t put my head on the block on if any of these are true, and I don’t think anyone else has a definitive assertion either. Please comment below; I hope to spark a conversation.
Premise One: By using GenerativeAI, many of the low level tasks that junior employees used to handle and start on their career path in GTM will be automated. This will eliminate a crucial step in the career pipeline of people who should be learning from more senior employees and limits the number of qualified individuals for higher level positions.
Stanton Cagle believes that Generative AI in the marketing world is a tool, a time saver. In some cases AI has completely replaced the “manpower” he needs for certain tasks and AI’s value seems to be highest for small, scrappy teams that can’t (or don’t want to) hire team members above the intern level. This is especially the case for those teams that may include only one or two people running an entire brand on their own. For example, Stanton said he uses it for tasks considered tedious or undesirable, such as SEO keyword research and keyword generation. These tasks used to require monthly time blocks in his calendar and weren’t easy for junior team members to latch onto, but often the results from the right AI prompts yield great results.
Jacco thinks this is a good thing: “Entry level sales roles, such as BDR or junior AEs, are stepping stones. This has always been the case.” But it will put more onus on sales leaders to train and develop AEs with core sales skills. What he sees is that “AI will take away a lot of the mundane work (scheduling / following up/ summarizing) and will do some of the hard work. This will give the “human seller” more time to spend with the “human buyer”and the human to human selling will become the biggest differentiator.”
Premise Two: GenerativeAI (as of now) isn’t enough for the whole end-to-end solution and humans still need to be in the loop.
There are bots that outbound. AIs that build marketing campaigns. Companies are building AI solutions to negotiate price. Is that enough?
First and foremost, Doug Landis hits the nail on the head, “ultimately buyers aren’t buying software, they’re buying an outcome.” For low value, transactional activities GenAI might be fine. It might save time in creating proposals and such. However for high value (not just in price, but in value to the business) solutions, it’s often a very intensive and thoughtful process to pull the needs out across multiple buying groups. It’s crucial to not only understand what a company truly needs (not just what they say they want), but also to fit into existing infrastructure and culture of the company, align with internal resources, and potentially radically change internal behaviour. That’s not going to be done by AI. Doug reminds us that “AI won’t be at a point for several years to where it fully understands the buyer’s environment and the things that the buyer is thinking about when they consider solutions.” So there is still the role for very experienced sellers, especially in complex enterprise sales.
Taking this a step further, Jacco van der Kooij, an expert on building GTM frameworks and systems that genuinely work for organisations and who creates content that is heavily validated by experience and data, points out a bigger challenge. As hundreds, or thousands of startups get going, they don’t know many of the ins and outs and rely on content created by ‘experts’. However, as Jacco knows well, “of all the experts out there creating content, only a few have the genuine experience to know what they’re talking about—the rest take that content and repurpose it, often with wrong information and often via GenerativeAI.” Without the humans who have been there, done it, have the scars to prove it—many companies are going to go down a wrong path for years before they figure out their mistakes.
And GenerativeAI is only as good as the data it has to work with. When Stanton comes to things like needing content generated for niche markets or customer bases that are very specific and take exceptional expertise, he finds that it’s going to be challenging or even impossible for AI to generate that and still come off as an expert and authentic.
“I think the perfect example is how much “meme culture” and word-of-mouth tie into modern marketing where companies are piggy backing off of a viral joke on their social accounts or newsletters or webinars. Sure, AI prompts scour the internet for their fodder and can potentially keep up with the breakneck speed that these inside jokes tend to carry.
But what about when those jokes start carrying themselves into physical media and culture? What if a funny or relatable moment happens live during, say, the NBA Finals that is relevant to your product? You need someone plugged into all of that to be creative and, maybe most importantly, sensitive to the many chances you’ll have of that “joke” not landing. I just don’t think AI understands NBA Twitter enough at this point to make a funny joke about Nikola Jokic that will resonate with that community AND your customers at the same time, while also being sensitive to all the bad things that can happen if the tone or language isn’t right.”
Stanton reminds us again of the importance of the human dimension: “Marketing at its core is a psychological field. Humans are unpredictable. Their emotions are unpredictable, their buying habits are unpredictable (or can be unpredictable), and their moods on a given day are unpredictable.”
Premise Three: The entire GTM motion will fundamentally change; what’s currently considered an ‘advantage’ now will actually hasten this change to eliminate the advantage altogether; and we’re all going to have to go back to some basics. Including (gasp) real human-to-human conversations.
As mentioned previously, there’s a growing group of companies that are built on the big LLMs and that automate much of the initial outbound motions of emails, LinkedIn requests and messaging and scheduling. These companies are creating very personalised messages, and are able to do so at scale.
Jacco sees a massive problem in the making here. An analogy he gave is thinking of outbound email being akin to trailer fishing with a net that drags the net over the bottom of the floor. And what we just figured out is that instead of producing the net for $10,000, we can now produce the net for $100. We just lowered the cost of mass outbound emailing both in tools and in labour costs. But in the meantime, the results are not going up. Using these these tools will “wreak havoc and it's going to torpedo the trusted relationship between the seller and a buyer.” If everyone is using these outbound tools for emails and sequences and LinkedIn messages, what will be the ultimate volume of inbound email? And at what point does that make these mediums entirely blocked and inaccessible? Which means, getting through to new buyers directly will be either via old fashioned networking, or picking up a phone and getting through.
Which leads to (this is NOT a paid ad) the premise of Jacco’s company, Winning By Design.
Rather than the typical ‘growth at all costs’ which often means chasing as many net new logos as possible, instead the healthiest and most successful companies will grow based on the success and expansion of their current customers. And the referenceability of this success, and recommendations from these very happy customers. He advises taking a “3 day Field Sales [and CS] Trip” and if you’ve only met one person in your customer instead of multiple, then you’re really missing the point—and pack as many meetings as possible. We all know that people move companies often enough and that they will bring those vendors with them that created success in the past—creating, documenting, nurturing and repeating these successes will be what drives new business, not the outbound sequences or current management methodologies of now. Most businesses spend most of their time, and board presentations, talking about net new growth, and just a few minutes on customer retention and net revenue retention (NRR). This is exactly the opposite of what Jacco has shown drives true healthy growth.
Analogous to this, Doug asserts a new model, the NLG, Network Led Growth, or network referrals, will become the norm, or at least should be. Rather than creating traditional sales territories based on geographies or verticals, thinking about networks to create territories will, or should be, primary. And, this is where AI can play a part—modeling both look-alike companies of those that are currently successful, or using AI to find patterns in LinkedIn networks to create territories based on who you know or are connected to within a certain level that might not be obviously correlated.
Stephen Millard, Operating Partner at Notion Capital, mentioned that in his first sales role, he didn’t make a single cold call–he cracked 8 of the top 10 companies that he was targeting solely based upon making the first client successful, and then asking for references into the next prospect, then the next then the next.
Premise Four: The current GenerativeAI offerings rest on murky ethical or business standards. Basing your business results solely on them and their outputs could lead to serious challenges in the near future.
Stanton sees a critical early challenge: “honestly, marketers are trusting AI with too much agency. You can’t always rely on a prompt to generate an expert take or opinion on an extremely niche subject matter. I don’t think I need to explain the damage it could cause if you’re using your marketing dollars in an ad campaign guided by data generated or organized from a prompt that may have some discrepancies in it” (see Google’s $100 billion valuation drop).
“If you’re using AI as a time saver when it comes to, say, content generation and you haven’t fully vetted that content because maybe you’re not an expert in the subject matter either, you’re going to see marketers getting in a lot of trouble with their product, sales, or legal departments. We’ve already seen plenty of examples happening in the legal field specifically, and I think it’s playing with fire a bit. It all comes back to using AI as a tool, and not as a replacement for quality, relevant, and subjective work.”
Doug also sees challenges for the huge wave of companies based solely on another’s large language model (LLM). More and more companies are blocking the use of these Open LLMs, so if you’re building a company based upon these–you’re limiting your addressable market. Nor are these companies’ solutions anything that isn’t easily copied. The questions he asks are: “what’s unique, what’s defendable, what’s protectable?” Co-pilots are table-stakes according to him—what’s really interesting is the infrastructure: the companies that make LLMs ubiquitous, self-serving, connected to other applications.
There’s also, as Jacco has stated before, the concern that many people are using, but not crediting or monetising the original creators of content, and right now GenerativeAI is basically just regurgitating, albeit with great grammar, content already out in the public domain. True experts are difficult to discern and what might be presented confidently, is not the right information for a particular business’s needs.
And as Stanton states, “Is it ethical to take the experiences of others and claim it as your own work, even if it’s for something as non-consequential as marketing copy? Is it ethical to claim your work is done by an expert if it’s really just a combination of different experts' takes or opinions? I don’t think it is.”
It’s no secret that in many places, even the concept of ‘facts’ are disputed, so it’s no surprise that the extreme polarity between people then impacts if LLMs are considered ‘biased’. We have to sort out our humanity first. From Doug, “technology is only going to exacerbate what’s going on in our human domain.” As GTM professionals, we need to think hard about what we do now, and what’s the chain reaction down the line as a result—are we being penny-wise and time-foolish in the long run, or are we fundamentally changing the game and need to start rethinking what used to be the basic foundations?