AI and Marketers: What's Actually Happening and What to Do

The honest assessment

Marketing is in a weird spot with AI. Half the industry is pretending it changes everything. The other half is pretending it changes nothing. Both are wrong.

Here's what AI can already do in marketing, and do reasonably well: generate first drafts of blog posts, social captions, and email sequences. Build audience segments from CRM data. Produce dozens of ad copy variations for A/B testing in the time it used to take to write three. Create passable images for social media. Summarise campaign performance data and suggest optimisations. Repurpose a webinar transcript into ten LinkedIn posts. Tools like ChatGPT, Claude, and Jasper are doing this work right now, in real marketing teams, at real companies.

What it can't do... and this matters... is strategy. AI doesn't know that your CEO hates anything that looks too "salesy." It doesn't know that your biggest customer found you through a niche subreddit three years ago and that's why organic community matters more than paid for your specific business. It can't read the room in a brand workshop and realise the CMO and the head of product have completely different visions for the positioning. The human context stuff, the politics, the taste, the instinct for what will land with a specific audience... that's still yours.

The parts that are actively shifting: content production is getting radically faster and cheaper. Coca-Cola, Unilever, and dozens of other major brands are already using generative AI for ad creative. HubSpot's AI tools are writing marketing emails for hundreds of thousands of businesses. The volume of content being produced is exploding, which means the bar for what cuts through is going up. More noise, same amount of attention. That's the real challenge.

Where it gets properly interesting is in data and personalisation. AI can now analyse customer behaviour patterns across channels and surface insights that would take a human analyst days to find. If you're a marketer who understands both the creative side and the data side, you're about to become extremely valuable. If you only do one... it depends which one.

Your exposure level: Medium

Medium doesn't mean "relax." It means the disruption is real but uneven. Some marketing roles are being hit hard right now. Others are barely affected. The difference comes down to what percentage of your job is production versus strategy.

If you spend most of your day writing copy, creating social posts, building email campaigns, and producing reports... your exposure is closer to high. These are the tasks AI handles well enough that companies are already reducing headcount. A McKinsey report found that marketing and sales roles had the highest potential for automation by generative AI of any business function. Not accounting. Not admin. Marketing.

But if your role is more about campaign strategy, brand positioning, client relationships, market analysis, and creative direction... you've got more breathing room. Not immunity. Breathing room. The strategic layer of marketing requires understanding context, making judgement calls under uncertainty, and persuading actual humans in rooms. AI is a long way from doing that well. The marketers i see doing best right now are the ones using AI to eliminate the production grind so they can spend more time on the thinking. They're not faster at writing. They're faster at everything else, so they have time to write well.

The 90-day action plan

1. This week: generate a full campaign brief with AI. Take a product or service you know well. Open ChatGPT and prompt it to create a complete campaign brief: target audience, key messages, channel strategy, content calendar, and KPIs. See how close it gets to what you'd produce. The bits where it's generic or wrong? That's where your expertise lives. Know those bits intimately.

2. Week two: build an AI-assisted content workflow. Pick one recurring content task. Weekly social posts, monthly newsletter, blog content. Use Claude or ChatGPT to generate first drafts from a detailed brief you write. Edit them into your brand voice. Time the whole process. If you can prove you've cut content production time by 40% with no quality drop, that's a conversation worth having with your manager.

3. By day 30: learn AI-powered audience analysis. If you use HubSpot, Salesforce, or any major CRM, check what AI features they've shipped in the last year. Most marketers are sitting on AI tools they're already paying for and not using. Copilot in your analytics dashboard can answer questions like "which customer segment had the highest LTV growth last quarter" in plain English.

4. By day 45: create a prompt library for your brand. This is a genuine competitive advantage and most marketers haven't done it. Write 10-15 detailed prompts that capture your brand voice, audience, and typical content needs. Include examples of good and bad output. Save them somewhere the team can access. You've just made everyone faster and more consistent.

5. By day 60: experiment with AI-generated creative. Use Gamma for a presentation deck. Use an image generator for social media concepts. You don't have to publish any of it. The point is to understand what's possible so you can brief it, evaluate it, and direct it. Art directors don't paint. But they need to know what good looks like.

6. By day 75: run a real A/B test. Take one campaign element. Email subject lines, ad copy, landing page headlines. Generate 10 variations with AI and test them against your human-written version. Document the results. This gives you actual data about AI performance in your specific context, not someone else's case study.

7. By day 90: present your findings. Put together a short deck (Gamma will help) showing what you tested, what worked, what didn't, and what you'd recommend the team adopts. Concrete recommendations backed by your own data. That's how you become the person who actually knows about AI rather than the person who just talks about it.

AI tools you should be using this week

• ChatGPT for Work — The Swiss army knife. Use it for copy drafts, brainstorming campaign angles, repurposing long-form content into social snippets, and generating audience persona documents. Best when you give it detailed context about your brand and audience upfront.

• Claude for Work — Better than ChatGPT for longer strategic work. Paste in a full marketing plan or competitor analysis and ask it to identify gaps. Excellent for synthesising large amounts of research into actionable briefs. i prefer it for anything over 1,000 words.

• Gamma for Presentations — Creates presentation decks from a text prompt or brief. Won't replace a proper designer for client-facing work, but for internal decks, stakeholder updates, and strategy presentations, it saves hours. The templates are surprisingly decent.

• Grammarly AI — Goes beyond spell-check now. It can adjust tone, rewrite for clarity, and check that your copy aligns with brand voice guidelines you set up. Useful for reviewing AI-generated drafts before they go out. Catches the subtle "this sounds like a robot wrote it" issues.

• Perplexity for Research — When you need to pull together market data, competitor information, or industry trends quickly. Gives you cited sources, which matters when you're building a strategy doc and need to back up your recommendations with something more credible than "AI told me."

What to say in meetings

When the quarterly planning meeting comes round and someone inevitably says "should we be using AI for our content?"... don't do the thing where you either hype it or dismiss it. Try: "I've been testing it on our email campaigns. It's good for first drafts and variations, but it still needs a human pass for anything client-facing. I can show you the workflow if it'd be useful." Practical. Measured. Shows you've done the work.

If your agency or client brings up AI-generated content and there's anxiety in the room, this works well: "The tools are getting good at volume. Our job is making sure we're focused on the stuff that actually performs, not just producing more of it." That reframes the conversation from "are we being replaced" to "are we being strategic."

And if you're in a budget conversation where someone suggests replacing a team member with an AI tool: "We tested that. The output needs about 40% editing time to hit our quality bar. What works better is giving the team the tools so they can do more." You want to be the person with the data, not the opinion.

If the worst happens

If you're made redundant from a marketing role, the good news is that marketing skills are genuinely transferable in ways that some other professions aren't. You understand audiences, messaging, data, and how to make people care about things. That's valuable everywhere.

The most natural adjacent moves: content strategy (which is growing as companies try to make sense of AI-generated content), product marketing (which is more strategic and less exposed), customer success (which uses the same audience understanding and communication skills), and brand consulting. If you've built real AI skills during your 90 days, "marketing consultant who actually understands AI implementation" is a real niche with real demand. Small and mid-sized businesses are desperate for someone who can tell them what to actually do rather than just showing them another demo.

Here's what i'd say from experience. The marketing job market is noisy right now. Lots of people with "AI" in their LinkedIn headline who've done nothing more than write a few prompts. If you've got a portfolio that shows real campaigns, real results, and real AI integration... you stand out. Bring your A/B test results. Bring your workflow documentation. Bring the prompt library you built. That's not a CV. That's proof you can do the job that marketing is becoming, not the job it used to be.