cotera
Articles

AI for Marketing: The Part Nobody Talks About Is the Part That Actually Works

Ibby SyedIbby Syed, Founder, Cotera
8 min readFebruary 9, 2026

AI for Marketing: The Part Nobody Talks About Is the Part That Actually Works

AI for Marketing

If you've been to a marketing conference in the last two years, you've heard the pitch. AI will write your emails. AI will generate your social posts. AI will generate your ad creative, personalize every visitor's experience, and — this was actually said on a panel I attended — "eliminate the need for junior marketers entirely." The speakers looked very proud of themselves.

Then you get back to your desk, fire up the tool, and spend forty minutes watching it produce blog post drafts that sound like they were assembled from the recycled phrases of every marketing blog ever indexed. Technically correct. Strategically vacant. The emails could have been written by a particularly well-read chatbot, which — to be fair — is exactly what happened. The social posts are so professionally bland they could belong to any company in any industry. The ad copy nails every keyword and resonates with no one.

I've watched this cycle play out at company after company. The hype promises execution. The reality delivers mediocre first drafts. And marketing teams end up spending as much time editing AI output as they would have spent writing from scratch, except now they also feel vaguely guilty about using the tool incorrectly.

Here's what I think went wrong: the entire AI-for-marketing conversation has been dominated by content generation. Write faster, publish more, produce produce produce. But the part of marketing where AI delivers genuinely transformative value isn't production at all. It's the part that happens before anyone writes a single word.

The Intelligence Layer

Every good marketing decision starts with research. What do our customers actually care about? What are competitors doing? Which keywords have real volume? What's the sentiment around our brand? Where are the gaps in the market?

This research phase is where marketing teams spend (or should spend) enormous amounts of time. It's also where they consistently cut corners — because the research is tedious, because the tools are fragmented, and because the pressure to ship content is relentless. "We don't have time for competitive analysis" is something I've heard from marketing leaders at companies of every size. They're not wrong about the time. They're wrong about the trade-off.

What AI does brilliantly — genuinely, game-changingly well — is compress the research phase from days to minutes. Not by guessing or hallucinating, but by systematically pulling data from the same public sources a human would check, then synthesizing it into something actionable.

Intelligence Layer

I call this the Intelligence Layer. Picture flipping the typical AI workflow on its head. Instead of parking AI at the tail end of your process to crank out words, you station it at the front to gather intelligence. The prompt isn't "write me a blog post about X." It's "before I write anything, tell me everything about this market — who the competitors are, what keywords actually have volume, what customers are complaining about, where the content gaps are." AI handles the research grind. You handle the judgment calls. You write (or direct) the content yourself. The output is marketing that's both faster and genuinely better, because it stands on real data rather than whatever the team brainstormed over Thursday's lunch.

This is the unsexy application of AI for marketing. Nobody makes a splashy demo about competitive research automation. But it's the one that actually changes outcomes.

Where the Intelligence Layer Compounds

The magic of AI-powered marketing intelligence isn't any single research task. It's the compounding effect of doing all of them consistently.

Competitive intelligence. Not "what were your competitors doing last quarter when someone finally got around to the battle card update." What are they doing this week? Which keywords did they start targeting? Which ads are new? What are G2 reviewers specifically complaining about? A full competitive analysis that used to consume a week per competitor now takes an afternoon for your whole competitive set when AI is pulling LinkedIn, SimilarWeb, G2, job boards, and news in parallel. The companies running this monthly have a structural edge that compounds every single cycle.

SEO and keyword research. What should you write about? Not what sounds interesting in a brainstorm. What terms have actual search volume, manageable competition, and commercial intent? Competitor keyword analysis and SEO competitive analysis tell you exactly which content opportunities exist — backed by data, not intuition. The difference between teams that grow organic traffic predictably and teams that hope for it is almost always the quality of their keyword research.

Brand and social listening. What are people saying about you? Not in surveys where they give you the polite answer. On Reddit, Twitter, and G2 where they say what they actually think. Brand monitoring and social listening used to require $1K+/month tools that most teams abandoned within three months. AI-powered listening that delivers Slack digests instead of dashboard logins actually gets used. And teams that listen to their market make better positioning decisions. Shocking, I know.

Customer sentiment. What do customers love and hate? Not about you in general — about specific features, specific experiences, specific touch points. Sentiment analysis across reviews and social mentions surfaces patterns that no quarterly customer survey will ever capture, because surveys ask your questions and reviews answer theirs.

Each of these individually is useful. Run all four consistently and something interesting happens: your marketing strategy starts being driven by evidence instead of assumption. You know what competitors are doing, what keywords to target, what the market is saying, and what customers care about — before you write a single word. That's a fundamentally different starting point than "let's brainstorm what our next blog post should be about."

Why Content Generation AI Disappoints

Let me be precise about why content generation AI consistently disappoints people, because the obvious explanation — "the models aren't good enough" — is actually wrong. The models are disturbingly good at producing fluent, grammatical, well-organized prose. And that fluency is the trap.

Good marketing content isn't fluent text. It's text with a point of view, informed by specific knowledge that the reader doesn't have. When you read a great marketing article, the value comes from the author knowing something you don't — a data point, an experience, a counterintuitive insight, a framework born from practice. The writing quality matters, but it's secondary to the thinking quality.

AI content generation tools produce writing without thinking. They pattern-match on millions of existing articles and produce something that looks like marketing content but contains nothing new. Every sentence is defensible. No sentence is interesting. The output reads like someone fed Wikipedia into a blender with a HubSpot blog and set it to "approachable." Readability score: excellent. Would you forward it to a colleague? Absolutely not.

The teams getting the most value from AI for marketing aren't using it to write. They're using it to research, analyze, and synthesize — then writing (or directing the writing of) content that's informed by AI-gathered intelligence but shaped by human judgment and actual expertise.

Put differently: AI is an incredible research assistant and a mediocre copywriter. Most teams have that backwards.

The Research Stack That Actually Works

Here's the specific intelligence stack I'd build for any B2B marketing team. The order matters — each layer builds on the one before it.

Layer 1: Competitive landscape. Monthly. Who are your competitors, what are they doing, and what's changed? Market intelligence covering hiring, reviews, keywords, traffic, leadership, and news for each major competitor. This is your strategic foundation.

Layer 2: Content opportunities. Monthly. What keywords should you target? Competitor keyword analysis plus traffic data tells you where the gaps are. Combine with search volume and difficulty scores to build a prioritized content calendar that's driven by data.

Layer 3: Audience intelligence. Weekly. What is your market saying right now? Brand monitoring and sentiment analysis across Twitter, Reddit, reviews, and news. This surfaces the language your customers use, the problems they're frustrated by, and the competitive comparisons they make naturally. All of this feeds into better positioning, better messaging, and better content angles.

Layer 4: Campaign intelligence. Ongoing. What are competitors running in paid? PPC competitive analysis covers Google Ads, Meta, LinkedIn, TikTok, and Reddit ads. This tells you what messaging competitors are testing, which platforms they're investing in, and how their funnel is structured. You learn from their ad spend without spending a dollar yourself.

The compounding effect of running all four layers is substantial. You're not just "doing marketing." You're doing marketing with better information than 95% of teams in your space, because those teams are still brainstorming in a conference room while you're working from actual market data.

The "So What?"

The AI for marketing conversation has been captured by content generation tools that promise to replace the writing process. They won't — not because AI can't write, but because writing without insight produces content without value.

The real opportunity is AI as an intelligence layer. Research, competitive analysis, keyword analysis, brand monitoring, sentiment tracking — the work that informs strategy. This is work that most teams know they should do but don't, because the manual process takes too long. AI makes it fast enough to actually happen. And marketing teams that operate on intelligence instead of intuition make better decisions, produce better content, and grow faster.

Stop asking AI to write your marketing. Start asking it to make your marketing smarter.

Try These Agents

For people who think busywork is boring

Build your first agent in minutes with no complex engineering, just typing out instructions.

Get started