How I Actually Use AI in Marketing (The Honest Version)
Let me be direct about something first: AI does not write my clients’ copy. I do.
What AI does is dramatically accelerate the research, planning, and analysis stages that happen before I write a single word. The actual copywriting — the hooks, the persuasion architecture, the emotional resonance, the voice — that is human work. AI helps me do the preparation 10 times faster so the human work is better informed and more precisely targeted.
Most articles about AI marketing tools are written by people who have tested them in a sandbox, not people who rely on them to produce real ROAS for paying clients. I manage campaigns for businesses spending $1,000 to $50,000 per month on ads. The tools I recommend here are the ones I would not remove from my workflow. Everything else has been tested and cut.
AI Tools for Audience Research
Audience research is where AI provides the most immediate value for conversion copywriters and ads managers. Understanding your ideal customer’s language, fears, desires, and objections used to take days of manual research. AI compresses that into hours — with the right prompts.
ChatGPT — Persona Research and Language Mining
I use ChatGPT for deep audience research prompts that surface the exact language my ideal customer uses when describing their problem. Specifically, I ask it to role-play as the customer persona and describe their situation in their own words — frustrations, fears, desires, previous failed attempts, and what they would say to a friend.
The output gives me raw language I can mine for copy hooks, landing page headlines, and email subject lines. It is not always accurate, but it is fast — and it gives me a starting point I then validate against real customer reviews, Reddit threads, and support tickets.
The Audience Research Prompt
Copy and use this prompt directly. Replace the brackets with your specific details:
“You are a [specific customer type] who has been struggling with [specific problem] for [time period]. You have tried [previous solutions] but they did not work. Describe your situation in detail — what keeps you up at night, what you wish existed, what you have already spent money on trying to fix this, and exactly what you would say to a close friend about this problem. Use your own natural language, not marketing language.”
Claude (Anthropic) — Deep Persona Development
Claude is my preferred tool for nuanced audience analysis. Where ChatGPT sometimes defaults to generic marketing language under pressure, Claude tends to maintain the persona voice more consistently across longer research sessions. I use it specifically for building out detailed customer journey maps — mapping what the customer thinks, feels, and does at each stage from problem awareness to purchase decision.
AI Tools for Ad Copy Generation and Testing
AI is genuinely useful for generating multiple copy angle variations quickly — which is the most time-consuming part of the creative testing process. The key is using AI for volume and speed, then applying human judgment for selection and quality.
Claude — Hook Angle Generation
My primary tool for generating multiple Facebook ad hook variations. I feed it the audience research output, the offer details, and the target market, then ask for 10 different hook angles using different emotional entry points: fear, curiosity, desire, frustration, social proof, and specificity.
I then read through the 10 outputs, select the 2 or 3 strongest angles, and rewrite those completely from scratch using proper direct response principles. The AI provides the angle hypothesis. The human does the actual writing.
Jasper — First-Draft Ad Copy at Scale
Useful specifically for generating first-draft Facebook ad copy when I need to produce a large number of creative variations quickly for a new campaign launch. The output quality is consistently mediocre — but mediocre first drafts are faster to edit than blank pages. I use Jasper to reduce blank-page paralysis and generate raw material, never to produce final copy.
Generate ad copy with AI → Copy and paste directly into Facebook Ads Manager → Run without editing → Wonder why CTR is 0.4% and ROAS is 1.1x
AI copy reads like a template. Audiences have seen thousands of AI-generated ads. They know. They scroll.
Use AI to generate 10 angle variations → Select the 2 strongest emotional angles → Rewrite completely in human voice with direct response principles → Test against each other → Iterate based on real performance data
AI gives you speed. Human judgment gives you conversion. The combination gives you 4x+ ROAS.
AI Tools for Performance Analysis
This is where AI provides the second biggest value for paid advertising — surfacing patterns in campaign data faster than manual analysis. Most advertisers stare at spreadsheets and miss the signal buried in the noise. AI finds it faster.
Meta Advantage+ Creative Insights
Facebook’s own AI surfaces which creative elements are performing best across your ad sets — which headlines are driving clicks, which copy angles are generating leads at the lowest CPL, which visual styles are getting the highest CTR. I check this weekly for every client account and use it to identify patterns that inform the next creative iteration cycle.
This is free, built directly into Meta Ads Manager, and most advertisers completely ignore it. It is one of the most underused tools in paid social advertising.
Google Performance Max Asset Insights
Google’s Performance Max campaigns provide asset-level performance labels (Best, Good, Low) for every headline, description, and image you are running. This tells you exactly which copy elements Google’s AI is selecting most often — which is strong signal for what is resonating with your audience at the search and display level.
ChatGPT for Data Pattern Analysis
I copy raw performance data — CTR, CPL, ROAS, conversion rate by ad set and creative — and ask ChatGPT to identify patterns, anomalies, and hypotheses worth testing. It does not replace human strategic judgment but it consistently surfaces angles and correlations I might miss on a manual review of the same data set.
I use its output as a list of hypotheses to evaluate, not as conclusions to act on directly. The judgment about which hypotheses are worth testing remains human.
AI Tools I Tried and Stopped Using
Honesty about what does not work matters more than a comprehensive list of tools. Here are the AI tools I tested, used in real client campaigns, and removed from my workflow — and exactly why.
Fully Automated AI Ad Copy Tools
The copy quality is consistently mediocre. It reads like a marketing template — generic, safe, predictable. Audiences have seen thousands of AI-generated ads. They recognise the pattern instantly and scroll. CTR drops. ROAS follows.
AI Image Generators for Ad Creatives
Inconsistent quality, frequent uncanny valley aesthetics, and — most importantly — the output lacks the strategic visual thinking that makes ad creative actually work. Good creative decisions require understanding of what stops a scroll. AI does not understand that yet.
AI Email Subject Line Optimisers
Less useful than simply writing three subject line variations by hand using proper curiosity, specificity, and pattern interruption principles. The AI output tends toward safe, generic phrasing that gets ignored in crowded inboxes.
AI Landing Page Builders
The copy architecture is structurally sound but emotionally flat. Landing pages need to make the reader feel something — recognised, understood, hopeful, urgent. AI landing page copy makes readers feel like they are reading a brochure.
AI for Final Sales Page Copy
Sales pages require the highest level of persuasion precision — every sentence earns its place by moving the reader toward a specific emotional state. AI output at this level is consistently too generic to do the job. The stakes are too high for template thinking.
AI Chatbots as Customer Service Replacements
In the marketing context specifically, AI chatbots handling warm leads who have questions about a purchase convert at dramatically lower rates than a human follow-up sequence. The handoff from AI to human needs to happen earlier than most businesses think.
The Right Way to Think About AI in Marketing
AI is a force multiplier for skilled marketers. It makes fast things faster and turns research that used to take a week into an afternoon task. But it does not replace the fundamental skill of understanding what makes people take action — which is a human psychology question, not a technology question.
The businesses getting the most from AI marketing tools in 2026 are the ones pairing AI speed with strong human judgment. They use AI to:
- Research audiences in hours rather than days
- Generate 10 creative angle variations in minutes rather than hours
- Surface performance patterns across large data sets automatically
- Reduce blank-page paralysis at the first-draft stage
- Test more creative variations per month than was previously possible
And they use human expertise to:
- Select which angles are actually worth testing
- Write the final copy that will run in front of real audiences
- Make strategic campaign decisions based on business context
- Position the offer in a way that feels genuinely different
- Build the emotional connection that turns browsers into buyers
Use AI to go faster. Use your brain to go better. The combination beats either on its own by a significant margin.
AI vs Human: Where Each Wins in 2026
Here is the honest breakdown of where AI outperforms human-only processes, where humans outperform AI-only processes, and where the combination is the only correct answer:
Want AI-Powered Marketing Applied to Your Business?
Book a free 30-minute audit with Umer Khan. He will show you exactly how AI research and systems can improve your ad performance — and what that looks like applied to your specific campaigns, audiences, and offer.
Book Your Free Audit → uk7539214@gmail.com · umerkhan.online · No pitch. No commitment.Frequently Asked Questions — AI Marketing Tools 2026
The best AI marketing tools for 2026 that Umer Khan uses daily are: ChatGPT and Claude for audience research and copy angle generation, Jasper for first-draft Facebook ad copy, Meta Advantage+ creative insights for performance pattern analysis, and Google Performance Max insights for asset-level data. The key is using AI for research and generation — not for final copy output, which always requires human rewriting.
AI can generate useful first drafts and copy angle variations for Facebook ads, but it cannot write genuinely persuasive direct response copy on its own. The best process: use AI to generate 10 angle variations, then have a skilled human copywriter select and completely rewrite the strongest 2 to 3. AI saves time on generation; human judgment determines what actually converts at 4x+ ROAS.
ChatGPT is excellent for specific marketing tasks: audience persona research, generating copy angle variations, interpreting campaign performance data, and brainstorming offer positioning. It is less effective for final copy output, email subject line optimisation, and anything requiring genuine emotional nuance or strategic persuasion judgment. Use it as a research and ideation tool, not a writing tool.
Umer Khan at Umerix Growth uses ChatGPT and Claude for audience research prompts and copy angle generation, Jasper for first-draft ad copy at scale, Meta Advantage+ creative insights for performance analysis, and Google Performance Max insights for asset-level data. All final copy is written by Umer Khan personally — AI accelerates the research phase, never the writing phase.
No. AI is a force multiplier for skilled copywriters — it makes research faster and creative generation cheaper. But it does not replace the human judgment required to understand what makes a specific audience take action. Businesses that replace copywriters with AI see lower conversion rates and worse ROAS because AI copy lacks the emotional nuance and strategic persuasion that direct response requires to convert at scale.
Use AI for: audience research and persona development, generating multiple copy angle variations quickly, interpreting campaign performance data, identifying patterns across large data sets, and first-draft content generation. Use human expertise for: selecting the strongest angles, writing final copy, strategic campaign decisions, offer positioning, and anything requiring genuine persuasion. AI goes faster. Humans go better. The combination produces 4.8x ROAS at Umerix Growth.