(Note: this is the second post in a series, go here to read the first post titled “CRM Was Never Built to Deliver Outcomes”)

—–

For the last several years, the conversation around marketing technology has been pretty predictable. First it was more data. Then more automation. Now it is more AI.

On the surface, that sounds like progress. Marketing teams can generate content faster, build segments faster, optimize campaigns faster, and respond to engagement signals faster. Sales teams can get summaries faster, recommendations faster, and more guidance on where to focus.

But from where I sit, that still misses the real issue.

The problem is not that marketing and sales lack activity. The problem is that too many systems still leave them operating from an incomplete picture of the business. They can make the go-to-market engine move faster, but they cannot reliably help it move in the right direction.

That is the difference that matters.

Speed Is Not the Same as Precision

Most competing CRM-centric stacks are still built to help teams manage interactions, workflows, and productivity inside the front office. Now they are layering AI on top of that foundation. That absolutely improves speed. It can make campaign teams more productive and sellers more efficient. But speed is not the same thing as precision, and it is definitely not the same thing as growth execution.

Marketing does not just need to know who clicked, who opened, or who filled out a form. It needs to know where there is real opportunity in the customer base, what kind of motion that opportunity calls for, and whether the business is actually in a position to support it. Sales needs that same level of confidence before outreach ever begins. That is a very different requirement than “give me better lead scoring” or “draft me a follow-up email.”

What Agentic Marketing Actually Requires

This is why I think the industry is still underestimating what agentic marketing and selling actually require.

If you want a system to help drive growth, it has to do more than automate campaign steps after a team decides what to run. It has to help identify the right motion in the first place. It has to continuously interpret signals across the installed base, product usage, contracts and renewals, service history, account activity, and increasingly across the commercial realities that determine whether an expansion motion makes sense now, later, or not at all.

That is what I mean by demand sensing.

Demand Sensing vs. Engagement Tracking

Demand sensing is not just listening for engagement. It is recognizing when something meaningful has changed in an account, a segment, or a book of business, and understanding what that change is likely to mean. It is the difference between seeing activity and recognizing readiness. It is the difference between collecting marketing exhaust and identifying a real growth motion.

A customer might show strong product usage but no adjacent product penetration. Another might be approaching renewal with signs of drift that would never show up in campaign data alone. Another might look like an upsell opportunity until you factor in service friction, billing instability, or some operational constraint that makes the timing wrong. In each of those cases the signal is there, but only if the system can see enough of the business to interpret it correctly.

That is where I believe a lot of current-market AI still falls short.

Where Current AI Still Falls Short

It is very good at generating outputs from the data already sitting inside the front-office application. It is much less capable when the job is to understand growth in context. And marketing feels that gap every day. Teams end up compensating with meetings, spreadsheets, seller backchannels, ad hoc review loops, and manual decision-making to figure out which opportunity is real, which message is right, and which program should actually be launched.

That is not because marketers lack creativity or discipline. It is because the system is not carrying enough truth.

And once you see that, a lot of the noise in the market becomes easier to decode. A faster campaign builder is helpful. A better content generator is helpful. A smarter email assistant is helpful. But none of those solve the critical problem: how to move from signal to coordinated action with confidence.

From Automation to Judgment

That is where agentic marketing starts to become interesting.

To me, agentic marketing is not about turning AI loose to produce more assets. It is about giving the system enough context and enough structure to help assemble the right growth motion. That could mean identifying a cross-sell program that should be initiated now, surfacing the right audience, shaping the right message based on product and account truth, aligning sales outreach to the same context, and moving that motion forward without forcing half the organization into manual coordination.

That is a fundamentally different model than traditional campaign execution.

In the old model, marketing often starts with a plan and then goes looking for an audience. In the emerging model, the system continuously senses where opportunity or risk is forming, helps determine which motion fits, and prepares the work required to activate it. Sometimes that is a marketing-led program. Sometimes it is a seller-led motion. Sometimes it is a coordinated sequence across both. The point is not that AI replaces those teams. The point is that the system becomes much better at helping them converge around the right action.

The Need for a Shared Intelligence Layer

That only works if marketing programs and sales outreach are powered by the same intelligence layer.

This is the part I think many vendors still do not appreciate. You cannot have effective agentic marketing if the system only knows campaign behavior. You cannot have effective agentic selling if the system only knows pipeline stage. Both functions need access to a shared layer of truth that includes what the company sells, how those offerings are positioned, which accounts fit which offers, what signals indicate readiness or risk, what claims are supportable, what proof points are relevant, and how all of that maps to the account’s current commercial reality.

Without that layer, AI can generate language. With that layer, it can help generate judgment.

That is an important distinction.

What Will Define the Next Growth Stack

The future growth stack is not going to be defined by who can produce the most content or who can summarize the most records. It is going to be defined by who can help the business recognize the right opportunity, translate it into the right program or outreach motion, and move that motion forward with shared context across marketing and sales.

That is also why I think the gap between platforms is about to become much more visible.

The vendors that stay confined to front-office context will keep helping teams do more work. They will absolutely create productivity gains. But they will still leave marketing and sales inferring too much of the actual business from fragments. They will still leave teams stitching together signal, message, offer, timing, and follow-up through manual coordination. And they will keep calling that intelligence.

I do not think that is where this market is going.

The Shift Underway Now

The next phase belongs to systems that can sense demand more accurately, initiate programs more intelligently, and ground both marketing execution and sales outreach in a richer layer of truth. That is the real promise of agentic marketing and selling. Not more output. Better judgment. Better timing. Better coordination. And ultimately, better growth execution.

That is the shift underway now.

And it is a much bigger shift than adding AI to campaign software.