Abacus Insights Partners ads

We’re using AI to analyze bad data faster instead of fixing the data itself. Most of the excitement around AI in communications measurement is focused on what happens at the end of the process. Those are real improvements. No one is going to complain about getting a first draft in seconds instead of hours. But it’s worth pausing on a simpler question. What exactly is that AI analyzing? In many cases, the underlying data hasn’t changed all that much. It’s still built on the same familiar shortcuts: a. Boolean queries standing in for meaning b. Word proximity standing in for intent c. Article-level sentiment standing in for what actually landed AI hasn’t fixed those, just made them easier to process. So what you end up with is something that looks like progress i.e. faster, cleaner, more scalable but is still anchored in data that was never particularly precise to begin with. Anyone who has spent time inside comms measurement has seen a version of this before. - A new tool arrives. - The interface improves. - The charts get sharper. And yet the same question shows up in the executive meeting: “𝘚𝘰 𝘸𝘩𝘢𝘵 𝘴𝘩𝘰𝘶𝘭𝘥 𝘸𝘦 𝘢𝘤𝘵𝘶𝘢𝘭𝘭𝘺 𝘥𝘰 𝘸𝘪𝘵𝘩 𝘵𝘩𝘪𝘴?” That question doesn’t go away just because the report was generated faster. If anything, it becomes more important. It’s in how the data itself gets created. A few practical shifts I’ve seen make a difference: → 𝗔𝘂𝗱𝗶𝘁 𝘁𝗵𝗲 𝗶𝗻𝗽𝘂𝘁𝘀 𝗯𝗲𝗳𝗼𝗿𝗲 𝗰𝗲𝗹𝗲𝗯𝗿𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗼𝘂𝘁𝗽𝘂𝘁𝘀. If your measurement still depends on keyword lists and proximity rules, AI will only make those limitations more efficient. → 𝗣𝘂𝘀𝗵 𝘃𝗲𝗻𝗱𝗼𝗿𝘀 (𝗼𝗿 𝗶𝗻𝘁𝗲𝗿𝗻𝗮𝗹 𝘁𝗲𝗮𝗺𝘀) 𝗼𝗻 𝗱𝗮𝘁𝗮 𝗰𝗿𝗲𝗮𝘁𝗶𝗼𝗻, 𝗻𝗼𝘁 𝗷𝘂𝘀𝘁 𝗿𝗲𝗽𝗼𝗿𝘁𝗶𝗻𝗴 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀. Better tagging, better entity recognition, better theme detection, this is where the real leverage sits. → 𝗧𝗲𝘀𝘁 𝘄𝗵𝗲𝘁𝗵𝗲𝗿 𝘁𝗵𝗲 𝗱𝗮𝘁𝗮 𝗰𝗵𝗮𝗻𝗴𝗲𝘀 𝗱𝗲𝗰𝗶𝘀𝗶𝗼𝗻𝘀. If the output looks better but leads to the same conclusions, the underlying data likely hasn’t improved. None of this diminishes what AI can do. It just reframes where the value actually is. Because in the long run, communications teams won’t be differentiated by how quickly they can generate reports. They’ll be differentiated by whether the data those reports are built on is actually worth trusting. Which is why I’m curious how others are approaching this. I've been analyzing how companies show up compared to their competitors, and where they have to prioritize their earned media. Want me to run it for you? → DM me "AI" → I'll run the study for you. → I'll share where you're showing up well, where you're underrepresented, and what earned media messages have to land for your next. DM me "AI"

GEO is an AR Opportunity so analyst relations should treat AI-generated answers as a first-class channel. Not long ago, an enterprise buyer sizing up a cloud platform or SaaS category did something AR teams understood deeply: they read and talked to analysts. A Magic Quadrant, a Wave, an IDC MarketScape, a run of inquiry calls. Analyst influence was hard to measure, much of it happening in one-to-one conversations, but AR knew how to shape it. These markets are still complicated, and buyers still seek answers. Increasingly, they ask a different intermediary first. “Who leads this category?”, “What should my shortlist be?”, “Is this vendor credible?” gets typed into Claude, Perplexity, or Copilot and into analyst firms’ chatbots. The answer comes back synthesized. A major ingredient in that synthesis is analyst content. That single fact makes Generative Engine Optimization, the shaping of how a brand shows up in AI answers—an AR concern, not a marketing curiosity to file under SEO. Generative engines favor sources that are authoritative, structured, and comparative: material that ranks and adjudicates vendors. That describes what analyst firms produce. These markets are also a mess of competing vocabularies with vendors using different names for the same function and the same name for different ones and analysts normalize it into categories a buyer, or a model, can reason across. Independent analysts publishing on LinkedIn and their own sites get ingested. In many technology markets, analyst content matters as much to models as earned media coverage racing to track. Here AR has an edge. AR is better equipped to interpret it than the SEO teams behind most early GEO playbooks. What’s new for AR is that the signal is now visible. Tools can track where a brand surfaces in AI answers, how it’s framed, and which analyst voices appear to carry that framing. So whose desk does this belong on? The reflexive answer is SEO, but this isn’t keyword optimization. The focus instead is on tracking and understanding the authoritative third-party intermediaries whose framing shapes buyer perception. The first AI answer can set the consideration set before a buyer reads a Magic Quadrant or places an inquiry. It doesn’t go away after that. GEO doesn’t replace analyst relations; it is the new surface on which analyst influence is expressed, and AR is built to work it. If you're responsible for communications at a large enterprise, you've probably noticed a shift. Instead of going to media outlets directly, buyers are asking AI platforms for recommendations. AI synthesizes what those outlets are saying about you. I've been analyzing how companies show up compared to their competitors, and where they have to prioritize their earned media. Want me to run it for you? → DM me "AI" → I'll run the study for you. → I'll share where you're showing up well, where you're underrepresented, and what earned media messages have to land for your next. DM me "AI"

If the agency you hire spends $500K of your $10m agency budget (or $250k of your $5m) to measure your performance, you don’t really have an independent view of performance. Over time, that shows up in the results. As AI becomes more embedded across the comms landscape, governance becomes more important, not less. Questions around system setup, terminology, permissions, auditability, source transparency, reproducibility, and decision rights all become more complex when work is orchestrated across multiple systems rather than contained within a single tool. That complexity is manageable. The harder problem is that the underlying incentives do not change. Comms organizations and their agencies have historically operated under pressure to present results in the most favorable light. The familiar dynamics like grading your own test, or the fox guarding the hen house, do not disappear simply because AI is introduced into the system. If anything, AI can make this more complicated. There is enough flexibility in how these systems are configured, what data is included, how models are tuned, how outputs are framed, that it becomes easier to “stack the deck” without it being immediately obvious. And this is not always explicit or insidious: sometimes implicit biases about market-framing and terminology serve the same purpose, unknowingly. And because the outputs often appear more structured, more consistent, and more data-driven, they can carry a level of credibility that masks those underlying choices. That creates a subtle but important shift. The risk is not just biased reporting, but biased systems producing results that feel objective. As AI-enabled orchestration improves and more of the workflow becomes automated, the question is less about whether you can generate analysis and more about whether you can trust it, and explain how it was produced. If comms does not keep pace in how it governs and validates its data, there is a real risk that the function loses control of its own narrative to: external vendors, internal stakeholders, or adjacent teams that are more disciplined about measurement. Once that happens, it becomes much harder to reclaim. How are you thinking about data governance as AI becomes more embedded in your company’s comms and marketing workflows? I’ve written a more detailed version about this in a short memo on AI and comms measurement. Comment “𝗠𝗘𝗠𝗢” and I’ll send it over.

In comms measurement, “good enough” often means fast, cheap, and directionally right. That may win the budget. It doesn’t mean it should. There is a growing tension between the speed and cost advantages of AI and the need for data quality and accuracy. As AI adoption accelerates, decisions are increasingly being made at the executive level, sometimes over the legitimate objections of practitioners who are closer to how the data is actually produced and used. Those raising concerns can sound like Cassandra, correct, but inconvenient. In that environment, even imperfect systems can start to win. If something is: - Fast. - Scalable. - “Good enough”. Then, it can be attractive from a procurement standpoint, especially when the alternative is slower, more expensive, and harder to justify. The issue is that “good enough” in this context often means reinforcing the very limitations comms has been struggling for decades to move past. Clip counts, impressions, and surface-level summaries become easier to produce but they remain difficult to translate into meaningful decision drivers. That creates a real risk that AI-driven comms measurement moves backward before it moves forward. The outputs improve in speed and presentation, but not necessarily in substance. If teams do not start moving beyond those legacy constructs, cheaper systems will fill that gap. Executives will replace your was-never-good-enough system with one that stays not good enough. And when that happens, comms risks being pushed into a more tactical role, with less influence over how its impact is defined and understood inside the business. The opportunity with AI is real, but it depends on where it is applied. Used well, it should shift effort away from repetitive production work and toward interpretation, validation, and decision-making. Used poorly, it simply accelerates existing patterns. The difference is not the technology itself, but how deliberately it is integrated into the workflow, how discriminative the technologies are deployed and how much emphasis is placed on judgment. How are you thinking about that trade-off? I go deeper on this in a short memo on AI and comms measurement. Comment “𝗥𝗘𝗣𝗢𝗥𝗧” and I’ll send it your way.

A few years ago, a Comms’ exec and one of my first clients, said something that stuck with me: “Right now, nobody’s asking how I’m being measured. But when they do, I want to have the answer.” He understood something most leaders miss, that sooner or later, someone 𝘸𝘪𝘭𝘭 ask. “How do we know comms measurements are working?” “What’s the ROI on all this communications activity (PR, AR, Internal Comms, Owned content)?” And when that moment comes, you either shape the narrative or someone else shapes it for you. We built a framework that helped him get there: • KPIs tied to business outcomes, not activity. • Weighted impact scores instead of volume charts. • Market relevance and message resonance over clip counts and sentiment. When the question finally came, he didn’t defend the spend or scramble to justify his team’s work. He didn't fake his way through the meeting with clip counts and impression numbers. He walked in with the data, showed clear contribution, and let the results speak for themselves. There was no debate about budget or value, just a calm demonstration of impact that reframed the entire conversation. That’s what happens when you define success before the spotlight finds you. In comms, the silence of vanity metrics doesn’t protect you, it just leaves someone else to eventually tell your story. Happy to help you create the ROI, my DMs are open to chat. If you're responsible for communications at a large enterprise, you've probably noticed a shift. Instead of going to media outlets directly, buyers are asking AI platforms for recommendations. AI synthesizes what those outlets are saying about you. I've been analyzing how companies show up compared to their competitors, and the messages they have to prioritize in their earned media. Want me to run it for you? → DM me "AI" → I'll run the study for you. → I'll share where you're showing up well, where you're underrepresented, and what earned media messages have to land for you next. DM me "AI"

AI, analytics, narrative intelligence. Put together, they raise an interesting question for me: is comms entering a golden age? Everyone from Gartner to Scott Galloway seems to think Comms is poised to become more important in the age of AI. I think that is directionally right. But I also think the role will look meaningfully different than it does today. In a zero-click world, large language models are becoming a new layer between brands and audiences. People are increasingly getting synthesized answers instead of clicking through ten blue links. And at least for now, those models appear to triangulate across sources, looking for patterns and consistency in how a company is described. That matters. It suggests that the old model of pumping out hyperbolic owned content and hoping it overwhelms the conversation may become less effective. The breathless press release and the “category-defining, revolutionary, game-changing” product page may not disappear, but they may increasingly look like outliers against the broader media record. If the earned narrative says one thing and the owned narrative says something much more inflated, I suspect LLMs will learn to discount the latter. That creates an opening for Comms. As Generative Engine Optimization matures, I think smart CMOs are going to realize that owned content works better when it is closer to earned perception. Not identical. Still favorable. Still strategic. But grounded. Credible. Internally consistent with what the market, media, analysts, and customers are actually saying. That kind of storytelling has historically been closer to the instincts of Comms than to the instincts of demand marketing. So: golden age? I’m not there yet. But I do think we are heading toward a world in which Comms has more influence over the company narrative, and more influence over the owned content that teaches AI systems who a brand is. That is not a golden age. But it may be the start of a more important age. If you're responsible for communications at a large enterprise, you've probably noticed a shift. Instead of going to media outlets directly, buyers are asking AI platforms for recommendations. AI synthesizes what those outlets are saying about you. I've been analyzing how companies show up compared to their competitors, and the messages they have to prioritize in their earned media. Want me to run it for you? → DM me "AI" → I'll run the study for you. → I'll share where you're showing up well, where you're underrepresented, and what earned media messages have to land for you next. DM me "AI"