Categories
Everyone Is Experimenting With AI. Few Have a System.
A question I hear regularly about AI in go-to-market is some version of “which tool should we be using?” It is the wrong question. Everyone is using something different anyway, whether it is Claude, ChatGPT, Gemini, or a half-dozen others, and the tool matters far less than what you do with it.
So when I recently sat down with a small group of cybersecurity marketing and product leaders for a roundtable on AI in go-to-market, I wanted to push past that. The question I find far more interesting is what happens when AI has to cross the go-to-market motion: field marketing, product marketing, demand gen, product management, sales. Where is it actually connecting the dots, and where is everyone still working in their own silo?
The throughline across the whole conversation was this: the tools are further along than the systems around them. Here are my takeaways from that session.
The Tooling Is Impressive. The Alignment Is Not.
Around the room, people were doing genuinely interesting things. Several had built signal-driven workflows in Zapier with AI layered on top: real-time intent data kicking off a follow-on workflow to generate content, or Clay surfacing a job change or an industry trigger that automatically routed into sales outreach. The plumbing was different in each case, but the instinct was the same, letting a signal set the next action in motion without anyone touching it.
A common thread was competitive intelligence. A couple of teams had taken everything marketing produces, the battle cards and competitive content, and turned it into an internal assistant built on Anthropic or Gemini. Instead of a rep digging through a static battle card in the middle of a live deal, they can now just ask a question and get an answer. One team had done this for their two biggest competitors and said it was working well. Another had trained a similar agent on all their product content so sellers could ask questions directly, and was in the process of moving it from one model over to Claude after an acquisition.
It is impressive work. It is also fragmented. When I asked how all of this rolls up into one connected go-to-market system, nobody had a clean answer. For most companies, it simply does not yet.
One of the leaders put it plainly. He has deployed AI-capable tools that should sit upstream and feed sales, and the tools work fine, but adoption stalls. Marketing on its own is seeing real, successful use of generative AI, mostly in content and demand. The hard part is connecting the first lead all the way through to a closed deal. That gap is not an AI problem. It is an alignment problem.
The same theme came up again and again: marketing is moving faster because of AI, but the rest of the organization has not necessarily sped up to match. Product, sales, and customer success are all engaging customers in different ways, collecting different insights, and working toward different priorities. That mismatch creates friction.
The Voice of the Customer Is Everywhere and Nowhere
One theme ran underneath much of the discussion. Everyone in that room is talking to customers. Sales conversations, support tickets, surveys, advisory boards, win-loss calls, the call recordings sitting in Gong. There is no shortage of customer data anywhere.
The problem is that the output looks completely different depending on who you ask, because everyone has a different motivation. Sales is trying to close. Support is trying to resolve. Product is trying to prioritize the roadmap. Marketing is trying to position and find proof points. As one leader said, each function looks at the same customer interaction and walks away with a different takeaway.
So the issue is not data. It is shared understanding. Without it, everything you try to build on top suffers, including the AI systems. Several people pointed out that this is actually one of the best use cases for AI we discussed all session. Pull the research into one place, feed it in as context, and let the model do the heavy lifting. I shared an example from one of our own engagements, where a senior product marketer used a client’s private model to ingest customer data from many sources. The real skill was interrogating the output, pushing past the generic first answer and making the model cite its evidence, until it got to something visceral and defensible. The art is in the dialogue, not the prompt. We are building a whole library of those dialogues, because AI doing the munching paired with an experienced human steering it is genuinely powerful.
But the human has to be experienced. That caveat came up repeatedly.
Messaging Is Not Something You Set Once
One participant put it bluntly. “If we are not in a room with CISOs on a regular basis, we do not really have a product.” That sounds strong. In cybersecurity, it is just reality. The market moves fast, buyer expectations shift, and technical depth is non-negotiable. You cannot rely on messaging you wrote six months ago and expect it to still hold.
The teams doing this well are constantly talking to customers, running briefings, and refining the narrative. One leader pushes every product marketer on the team to have a few direct customer conversations every month. Not Gong calls, though those help at scale, but direct conversations, because that is where you hear the customer describe the problem in their own language.
We also talked about the temptation to lead with the technology instead of the problem. One leader admitted that a product launched last year leaned hard on the words “autonomous” and “agentic,” and in the end the customers did not care about any of that. What they cared about was that threat hunting was hard, slow, and expensive. They wanted higher-fidelity signals so they would know a hunt was worth running. The AI underneath was fine, but it only mattered because it solved that problem. A lot of teams are afraid to simplify the message because they worry a simple story will not justify the price tag. In my experience, the opposite is true. Clarity is what earns trust.
The other piece that AI consistently misses is emotion. One leader described sitting down with CISOs and sketching out their decision-making by hand, on paper, precisely because AI cannot capture what the executive is actually feeling: the worry about a false positive, the fear that the team’s work is not getting used, the very human concern about looking good in front of the board. Another made the same point from a completely different world, a major consumer hardware product whose entire campaign was built around how a person feels wearing it, with the technology there to reinforce the feeling. Strip the emotion out and you are left with a data sheet.
Skepticism Is Healthy Right Now
There was a fair amount of honest pushback in the room. The phrase that came up repeatedly was “AI slop.” Everyone is writing a blog. Everyone sounds like an expert. And in a field as technical as cybersecurity, content that sounds right but lacks real depth is a serious liability. The moment a buyer realizes you cannot go deep when it matters, you lose credibility, and you do not get it back easily.
The consensus was clear. AI helps with speed. It does not replace understanding. As one leader put it, “all AI tools go back to average,” so if you do not start with real clarity about your customer and your product, average is exactly what you will get back. A few people worried that junior team members are losing the muscle of staring at a blank page and writing something original, because it is so easy now to jump straight to the model. One leader said she forces herself to keep doing it by hand, just so she does not lose the skill. Another reminded us that even the best prototype generated in an afternoon still needs someone with years of product sense to know whether it is actually solving the right problem.
That is the real danger, getting to undifferentiated messaging faster than ever. The differentiation still has to come from somewhere AI cannot reach: primary research, real customer stories, a specific use case, a proof point, a subject-matter expert who has earned their tenure.
What These Leaders Want From a Marketing Partner
We ended on staffing, and the theme there tied everything together. When these executives think about who they want to hire or partner with, three characteristics came up over and over:
First, strong product marketers who understand the full lifecycle, from product conception all the way through to how a salesperson actually sells it. That strategic, end-to-end thinker is the role nobody sees AI replacing.
Second, people who can translate technical depth into clear positioning. The ones who can take genuine complexity and boil it down to the problem the customer actually has, without burying it in buzzwords.
Third, teams that can connect insights across functions. The people who understand that the stream they work on flows into the larger ocean of revenue, and who can move pipeline forward rather than just create noise at the top of the funnel.
In other words, people and companies who can think strategically, not just execute.
The Bottom Line
The tools are not the hard part anymore. Almost everyone in that room is experimenting, and a lot of it is working. The hard part is building a connected system around those tools, aligning the teams who touch the customer, and keeping the human expertise and emotion that AI cannot manufacture at the center of the work. AI is an accelerator. It is only as good as the clarity and the experience you bring to it.
If you are wrestling with any of this inside your own organization, that is exactly the kind of problem we like to help with at Aventi Group. I am always happy to trade notes over a 30-minute working session.
