When we first started experimenting with AI, I assumed the biggest opportunity would be more efficient demand generation.
Like most demand generation teams, we started with ChatGPT. We used it to draft ad copy, generate email variations, and speed up work that often slows marketers down. It delivered exactly what we expected. Demand programs more efficiently.
Looking back, I think that was the least interesting part of our AI journey.
The more interesting shift came later. Early generative AI showed us it could create content, summarize information, and consistently apply our messaging. That was valuable on its own, but it also changed the questions we started asking. If AI could understand our messaging well enough to evaluate a blog post, could it also understand the buying signals spread across our go-to-market systems and help us make better decisions?
As AI evolved beyond answering prompts and became capable of connecting information across systems and coordinating workflows, that question became more realistic. We weren't just creating our programs faster, we grew into assembling context, interpreting buying behavior, and helping sales understand when to engage, why that moment mattered, and how to prepare for the conversation.
That's when my thinking started to evolve.
For years, my focus has been on building programs that generate and accelerate pipeline. That hasn't changed. What has changed is how I think demand generation can contribute once buying behavior begins to emerge. Now, I believe our role extends beyond creating pipeline. We have an opportunity to intentionally create buying signals over time and help orchestrate that intel into meaningful action for sales.
I think about that journey in four stages: Crawl, Walk, Run, and Sprint. The first three represent capabilities we've built over the past few years. Sprint represents where we're going. It's the future we're building toward, not something we've fully realized yet.
Before AI touched anything customer-facing, we had to answer a simple question: could we trust what it produced?
Our first meaningful investment wasn't in campaign automation or sales workflows. It was in consistency. Our product marketing team trained a custom GPT using our positioning, messaging framework, proof points, and brand voice.
It wasn't designed to replace marketers or generate entire campaigns on its own. Instead, it became a practical tool that helped us move faster while staying consistent.
Yes, it accelerated the creation of ad copy, emails, landing pages, and website content. But one of its most valuable uses turned out to be reviewing work we'd already created.
If we'd written a blog post optimized for SEO or developed messaging for a new campaign, we'd ask a simple question before publishing:
"Does this still sound like Highspot?"
That checkpoint became foundational in our work.Rather than adding another review cycle, we gained a fast way to validate that our messaging remained consistent across channels and campaigns.
This was also the first time we started thinking differently about AI. If it could reliably understand our messaging framework and evaluate content against it, maybe its value wasn't limited to creating assets. Maybe it could also help interpret information, identify patterns, and provide guidance in other parts of our go-to-market. We didn't know exactly what that looked like yet, but now we could ask different questions about what was possible.
Once we felt confident in the content AI was helping us create, our attention shifted to a different challenge.
Content wasn't our bottleneck anymore, context was.
Every demand generation team has experienced some version of this. Someone downloads an ebook, registers for a webinar, or gets a booth scan at a conference. Marketing passes that engagement to sales, often with a follow-up play, recommended content, or an email template tied to the interaction or buyer persona.
Those resources are valuable, but they're only part of the picture.
The rep may not know another seller has already been working with the account. They may not realize multiple people from the same company have engaged with marketing programs over several months, or that specific content has consistently resonated with members of that account’s buying group.
As a result, follow-up often begins with a generic play built around the latest interaction instead of the broader context surrounding the account. That's both a context problem and a conversion problem.
We wanted to close that gap.
At Highspot, our Deal Agent brings together activity across content, emails, meetings, Salesforce and other customer interactions into a single view. Rather than asking sellers to figure it out on their own, we began using Claude to interpret it and present it in a way that was immediately actionable.
We moved away from Slack notifications that simply say, "Here's your new MQL," our alerts now explain why the engagement matters. It summarizes recent activity, highlights existing relationships within the account, recommends who else should be involved, suggests relevant content based on what's already been successful, and provides a draft response grounded in everything that's already happened.
Our sellers weren't starting from zero. They already had plays, templates, and recommended follow-up content. What they often lacked was the broader context spread across multiple systems. Instead of asking a rep to gather that information before deciding how to respond, we can now bring much of it together before they even begin researching the account.
That doesn't replace the seller's judgment. It gives them a better starting point and more space within which to make winning judgements. This was another point where my thinking shifted.
For a long time, I thought of demand generation primarily in terms of creating and accelerating pipeline. I still do. But now our role is beginning to expand.
As organizations capture more buying signals across an expanding ecosystem of GTM systems, and technology becomes better at connecting that information, marketing has an opportunity to orchestrate those signals into actionable guidance for sales. The goal isn't to tell sellers how to sell. It's to make sure they begin every conversation with the best possible understanding of why they're engaging that account in the first place.
Once we had a better way to provide sellers with context, the next question was clear: how could we help them focus on the right accounts before a conversation ever started?
Like most B2B organizations, we had no shortage of buying signals. We could see engagement across LinkedIn campaigns, marketing campaigns, Digital Room engagement, content engagement, website behavior, page-level intent and competitive keyword intent from AdRoll, and a growing number of other data points.
We weren’t data-challenged. We needed to take all that data and determine which signals actually mattered and then turn them into something sales could confidently act on.
That led us to build a propensity model that brings those signals together and groups accounts into four stages: Cold, Warm, Hot, and Fire. The model itself is important, but I don't think it's the most interesting part. The real value comes from what each stage enables marketing and sales to do together.
Cold accounts aren't ready for anyone's attention. We don't spend seller time there, and we don't invest marketing dollars there either.
Warm accounts are where marketing begins to lean in. Through advertising, nurture programs, and other activity, our goal is to create additional buying signals that move the account toward a point where direct sales engagement makes sense.
Once an account reaches Hot, the motion changes. An account development rep (ADR) begins direct outreach, while marketing continues supporting the account with messaging aligned to where that buyer appears to be in their journey.
When an account reaches Fire, it becomes a coordinated effort. Marketing continues creating relevant engagement while the ADR and account executive (AE) work together to move the opportunity forward.
This evolution was yet another step forward in how I view marketing's role in orchestrating pipeline. The propensity score wasn't the outcome. It was the trigger for orchestration.
When an account moves into Hot or Fire, we automatically generate an account research brief that brings together the context a seller would normally have to gather manually.
The workflow pulls together recent marketing engagement, website activity, account history from Salesforce, existing relationships, company research, and the signals that caused the account to surface in the first place. It also recommends the contacts most likely to engage and provides a prompt the seller can use within Claude to draft personalized outreach.
None of those capabilities are revolutionary on their own. The value comes from connecting them into a single workflow that helps sellers understand not just who to engage, but why now is the right time.
That's where I think orchestration becomes tangible.
Marketing isn't replacing the work of the seller. We're helping reduce the time between recognizing buying intent and acting on it confidently.
The faster we can connect those signals and provide meaningful context, the more time sellers can spend doing what only they can do: building relationships, asking thoughtful questions, and moving opportunities forward.
The final stage isn't something we've fully built yet. It's the direction we're actively building toward.
Today, our orchestration largely begins once an account reaches a meaningful threshold of buying intent. We recognize those signals, provide sellers with the context behind them, and help them determine the next best action.
Now, we want to extend that same thinking into marketing activation itself.
Historically, B2B marketing has organized campaigns around industries, personas, and segments. Those approaches have served us well, and they'll continue to have a place. But as more buying signals become available, and technology becomes better at connecting information across systems, I think we have an opportunity to go a step further.
Instead of relying on messaging designed for broad industries or buyer personas, we can begin using the context we've already gathered to emphasize the parts of our messaging that are most relevant to an individual account. The goal isn't to create an entirely new message every time. It's to deliver the right message, informed by what we already know about that account and where it is in its buying journey.
Our vision is to connect those insights directly into our activation platforms. As an account moves from Warm to Hot, or from Hot to Fire, the same signals guiding sales outreach can also influence the marketing experiences that account receives.
That doesn't mean abandoning our messaging framework. It means becoming much more intentional about which parts of our story we emphasize based on what we've learned about a specific account. Instead of delivering generic messaging built for an industry or buyer persona, we have an opportunity to use the context we've already assembled to deliver the part of our messaging that's most relevant to that account at that moment.
For us, capabilities like the AdRoll MCP make that vision increasingly possible by connecting marketing activations with the broader GTM ecosystem. The real opportunity isn't any single integration. It's enabling marketing, sales, AI, and customer data to move more seamlessly across systems so messaging, account insights, propensity signals, and activation can work together as part of one coordinated go-to-market motion.
There's still plenty of work ahead before that becomes reality. We'll continue refining the technology, the workflows, and, most importantly, how our teams work together. But the direction feels clear.
The future of demand generation is not simply producing more campaigns faster. It’s becoming better at creating buying signals, recognizing when those signals converge, and responding with increasingly greater relevance, better timing, and stronger coordination.
Over the course of our evolution, our biggest breakthrough wasn’t using AI.
It was realizing that marketing's contribution to revenue doesn't stop once pipeline has been created.
Generating and accelerating pipeline will always be at the core of what demand generation does. That doesn't diminish marketing's traditional role. It expands it. We still create demand, and we have a new opportunity to shape what happens after demand has been created by recognizing when buying signals converge and helping sales successfully act on them.
I don't think of buying signals as individual actions like attending a webinar or downloading a whitepaper. I think of them as the accumulation of behaviors that, taken together, tell us an account is moving closer to a buying decision. Marketing's opportunity is to intentionally create those signals through our programs, recognize when they begin to converge, and orchestrate that intelligence into winning sales moves.
Orchestration isn’t about replacing the judgement of the seller or automating relationships. What it is about is helping the right information reach the right people at the right time so they can make better decisions.
One thing this journey has reinforced for me is that technology rarely changes an organization on its own. Every capability we've built required marketing, sales, operations, and product teams to rethink how we work together. AI made new workflows possible, but people still decide which signals matter, when to act, and how to create meaningful conversations with customers. In many ways, technology has been the easier part. The harder, and more rewarding, work has been advancing how our teams operate together.
If your team is primarily using AI to create content today, that's a great place to start. It was for us too. But I'd encourage you to ask new questions:
Where are your buying signals today?
How intentionally are you creating them?
How connected are they?
And what would change if demand generation became responsible not only for creating those signals, but also for orchestrating them into meaningful action for sales?
I don't think we've answered all of those questions yet because we're still learning ourselves.
But I do think that's where demand generation's next job begins.
Generative AI didn't just change what marketing could produce. It changed what marketing could contribute. And, for me, that's the next evolution of demand generation.
Last updated on July 29th, 2026.