{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreibil5bghnk72y25p4wzqwlinj7lashx2fbwegkpltbrncjwafapra",
    "uri": "at://did:plc:tja4igwnrdfjk3m3yqcchfku/app.bsky.feed.post/3mnf2qf2okve2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreifekpzwsqclzn3373nng2fu2ehwmttllunmwtben5dft534m7egp4"
    },
    "mimeType": "image/png",
    "size": 264177
  },
  "description": "On rebuilding a marketing team around AI without losing what matters.",
  "path": "/where-ai-belongs-in-marketing-and-where-it-doesnt/",
  "publishedAt": "2026-06-03T12:00:21.000Z",
  "site": "https://www.cmoalliance.com",
  "tags": [
    "tool recommendations",
    "ChatGPT",
    "Campaigns",
    "Messaging",
    "outreach",
    "product narrative",
    "marketing leader",
    "localization",
    "trust"
  ],
  "textContent": "There’s a version of this article I could have written six months ago.\n\nIt would have sounded more certain, more polished, and probably more useful in all the wrong ways. A neat framework, a few tool recommendations, and a closing line about embracing the future. The kind of piece that makes it sound as if the strategy was obvious from the start.\n\nIt wasn't.\n\nWhat I've learned since is more useful than a framework: the real question is not how to use more AI, but where it actually belongs, where it adds value, and where it starts to weaken the very work marketing is supposed to do. That distinction matters more than most people will admit. And in the markets we operate in, it matters more than most.\n\n## It started with a gap, not a strategy\n\nBefore I get into what changed this year, one thing needs to be said: AI is not new for us. We started using it as soon as it became accessible. ChatGPT, the early tools, the first wave of integrations. We were in it from the beginning.\n\nThe question was never whether to use AI. It was always how, and where, and at what cost to the work.\n\nThat framing matters, because what I'm describing here is not a team discovering AI. It is a team that has been living with it long enough to know where it helps and where it doesn't.\n\nEarlier this year, someone on my team left. The work they were responsible for didn’t leave with them. Campaigns still had to run, reporting still had to move, and the operational coverage that keeps a marketing function functioning could not simply wait for the hiring process to be complete.\n\nAt that point, AI wasn’t a strategic choice. It was a practical one.\n\nThat constraint forced a more honest conversation than any strategy session would have. When you stop asking \"how do we adopt AI?\" and start asking \"what actually needs to get done, and what's the best way there right now?\", you make better decisions.\n\nAI is scaffolding, not the building. It can help you move faster, structure the work, and reduce friction, but it cannot create the thing itself.\n\nSome work immediately benefited. First drafts, research synthesis, content adaptation across markets, reporting summaries, repetitive analysis. A campaign brief that would have taken an hour to structure could be shaped in minutes. A performance report that used to require manual extraction could be condensed into a cleaner first pass.\n\nThat is useful. But usefulness is not the same as leadership.\n\nOther work did not belong anywhere near AI. Positioning. Messaging. Prioritization. Trade-offs between brand and performance. The judgment calls about what one market needs versus another. The decisions that depend on context, trust, and experience.\n\nThat distinction became the basis for how we started rebuilding.\n\n## What the noise actually costs\n\nWhat surprised me most was not the technology. It was the amount of attention it consumes.\n\nNot long ago, I received a cold outreach email from a vendor. The subject line told me that the future of my job was being discussed in a room I wasn't in. The message made clear that if I wasn't using AI more aggressively, I was at risk of being left behind. Possibly losing my job.\n\nI found it extraordinary. Not because the pressure isn't real, but because that kind of fear-based selling tells you something important about the current moment. The noise isn't just loud. It is designed to be destabilizing.\n\nAnd it works, even on people who should know better – including me.\n\nThe AI landscape for marketers is relentless. Every week, there’s a new platform, a new integration, a new promise that this one will transform everything. Because marketing teams are expected to be early adopters, there’s pressure to test, compare, and have an opinion on all of it. That pressure comes from vendors, from leadership, from peers on LinkedIn, and honestly, from inside your own head.\n\nI’d be lying if I said I’d never felt it. The field moves fast enough that the anxiety is legitimate. I’ve sat in many meetings wondering whether I was behind, whether there was a tool someone else was using that was quietly making my approach obsolete.\n\nBut I’ve also watched that anxiety, unmanaged, do real damage. Not to me personally. To the team.\n\nEvery new tool comes with a learning curve. Every workflow change creates friction before it creates efficiency. Every \"quick test\" adds another layer of distraction to a team that already has a full operating rhythm to protect. And everyone has an opinion. Leadership asks why we aren't using a particular platform. A colleague forwards an article about something that transformed their pipeline. A vendor gets through on a call and plants a seed of doubt.\n\nOver time, the effect isn’t innovation; it’s noise.\n\nMost conversations about AI focus on productivity. Very few focus on attention. Attention is a scarce resource in a marketing team. If people are constantly interrupted by new tools, new claims, and new requests to evaluate the next thing, they’re not being enabled. They are being fragmented.\n\nI've become much more intentional about filtering on behalf of the team, and about being willing to push back, including on senior stakeholders, and say: I know what we are doing, and I know why we are not doing that. Even when I'm not entirely sure I'm right.\n\nWe tested a prospecting content assistant that promised highly personalized outreach across verticals and markets. In practice, it produced output that was fast but generic, and required so much manual correction that it added more work than it removed. We didn't roll it out. That decision protected the team's time, but it also protected quality.\n\nIn short, not every useful-sounding tool is actually useful.\n\n## Fixed and flex\n\nThe most useful model I've found is borrowed from how I think about brand voice across markets. Some things stay fixed. Some things flex.\n\nFor AI adoption, I apply the same logic.\n\n**Fixed** is the work that depends on judgment, accountability, and context. Campaign strategy. Positioning. Messaging in regulated markets. The framing of a product launch.\n\nAnd at least fifty percent of marketing leadership is stakeholder management. Aligning product, sales, legal, and the market itself is deeply human work. AI can help you prepare for those conversations, but it can’t have them. It can’t read the room in a tense meeting. It can’t sense that a sales leader is worried about pipeline risk but isn't saying it directly.\n\n**Flex** is where AI removes friction without removing value. Drafting support. Data summarization. Reporting automation. Localization assistance. Reformatting content for different channels. Pulling together campaign data so the team can focus on interpretation instead of assembly.\n\nThe challenge is that pressure always moves in one direction: toward faster output, more scale, more AI. Leadership is knowing where that instinct helps and where it starts to erode the work.\n\n## What I'm not automating\n\nPriceHubble supports clients across financial and real estate markets in ten countries. Our buyers are regulated, risk-averse, and deeply attuned to credibility. They do not respond to content that feels generated. They respond to precision, evidence, and the sense that the company they are evaluating understands the specific pressures they operate under.\n\nThat context shapes every AI decision we make.\n\nEarlier this year we were refining a product narrative for insurers. The stakeholder landscape was complicated: compliance requirements pulling in one direction, commercial messaging pulling in another, and a relationship with a key account that had taken two years to build. No prompt was going to navigate that.\n\nThe final narrative came from a series of conversations, a lot of redrafting, and someone on my team who understood the client well enough to know what they were actually worried about. AI played no part in that process, and it wasn't a gap.\n\nI don't want AI deciding how we position ourselves in a market where trust matters more than novelty. I don't want AI writing final messaging for a regulated segment where a single poorly calibrated sentence can undermine months of relationship building. I don't want AI replacing the thinking behind a product launch, a competitive response, or a market narrative.\n\nAnd I don't want AI becoming the default answer for an under-resourced team. When teams are stretched, the instinct is to automate everything possible. But automating the wrong things doesn't create leverage. It creates distance between the team and the quality of the work.\n\nThe real job of a marketing leader is not to remove humans from the process. It’s to make sure they are spending time where they add the most value.\n\n## How I lead differently now\n\nSome of this I learned the hard way – and it was harder than I'd like to admit.\n\nEarlier this year I pushed the team to trial a new AI content workflow during a particularly busy campaign period. The timing was wrong, the tool needed more setup than we had bandwidth for, and a few weeks in, one of my team members told me it was becoming a burden rather than a solution.\n\nShe was right.\n\nI’d been too focused on the potential upside to see what was actually happening around me. We paused, regrouped, and reintroduced the workflow properly once the campaign had landed. It worked much better the second time.\n\nThe lesson stayed with me: enthusiasm for a tool is not a substitute for reading the room, including your own team.\n\nSince then, I've stopped talking about AI as transformation and started treating it as infrastructure. It is part of the operating model, used where it improves quality, consistency, or speed. Not a separate initiative.\n\nI've also been more explicit about where we don't use it. That clarity matters more than I expected. Teams work better when they know the boundaries, not just the opportunities. People don't need endless encouragement to experiment. They need permission to use judgment.\n\nTwo examples that stuck with me:\n\n  * We considered AI-generated copy in a regulated market segment, but a team member flagged that the tool was flattening nuance in a way that could create compliance risk. They were right. We kept the idea generation support and removed AI from the final drafting process in that context.\n  * Separately, a draft localization workflow looked efficient on paper, but a market lead pointed out that it was stripping culturally relevant references in one region and over-explaining in another. The output was technically fine but strategically weak. We changed the process so AI could support adaptation, but not replace market review.\n\n\n\nGood leadership now requires a different kind of confidence. Not confidence that the tool will work because it is new, but confidence to say no when it doesn't improve the work.\n\n## What I'm trying to protect\n\nThe question I keep coming back to is not \"how do we use more AI?\" It’s \"what are we trying to protect?\"\n\nProtect thinking time. Protect the judgment behind the work. Protect the context that informs decisions. Protect the conversations that happen between functions before they become problems. Protect the trust that gets built when marketing is seen not just as a content engine but as a strategic partner.\n\nThe sign that it’s working isn’t a metric. It’s a feeling in the room. A team that has enough space to think, that isn’t burning through its energy evaluating tools, that’s doing the work rather than managing the overhead of the work.\n\nThat said, I want to be honest about the other side of this. Some tools have genuinely surprised me. Not because they promised transformation, but because they quietly disappeared into the work. The ones worth your time don't demand attention or create another layer to manage. They just make something that was slow, faster, or something that was fragmented, cleaner.\n\nWhen a tool does that, you notice it by its absence, not its presence. Those are the ones we keep.\n\nIn finance and real estate, trust is not a soft metric. It’s the reason a procurement team shortlists you, a compliance officer approves the relationship, and a client renews. It’s built slowly and lost quickly. No AI tool accelerates that. Only people do.\n\nSo no, this piece doesn't end with a framework. It doesn't end with a list of tools or a call to embrace the future. What I've learned is less tidy than that, and probably more useful:\n\n**Know what you're protecting, and don't automate it**.",
  "title": "Where AI belongs in marketing – and where it doesn't",
  "updatedAt": "2026-06-03T12:00:20.868Z"
}