The AI tools that have had the biggest impact on my daily work are Claude (for strategy and writing), Claude Code (for structured repeatable workflows), and Perplexity (for research). The combination that actually works is not about using every AI tool available — it is about using three to five well-chosen tools deeply, and building a prompt library around them.
There is no shortage of AI tools marketed to marketers. Most of them are either too narrow, too expensive for what they do, or built for problems that are not actually painful enough to need a dedicated tool.
Here is an honest account of what I actually use, what I dropped, and why.
What Does My Core AI Stack Look Like?
My daily AI stack has three components.
- Claude — for strategy, writing, analysis, and complex reasoning tasks
- Claude Code — for structured, repeatable marketing workflows via skill commands
- Perplexity — for research and finding current, sourced information quickly
That is the core. I have tried more tools than I can count. These are the ones that have earned a permanent place in how I work because they solve real, frequent problems.
How Do I Use Claude Every Day?
Claude is my primary AI tool for anything that requires extended reasoning or nuanced output. Campaign strategy, messaging frameworks, content reviews, performance data analysis, and working through complex positioning questions.
What separates Claude for me is its ability to hold complex context and produce structured, thoughtful responses without unnecessary hedging or filler. The output requires less editing than most alternatives I have tested.
I keep a context document for each client and project. At the start of any session involving that client, I paste the context. This removes the need to re-explain the brand, audience, and constraints every time, and the output quality is noticeably better for it.
How I use it in practice: Campaign briefs, messaging frameworks, competitive positioning analysis, content strategy, and reviewing and improving first drafts. Anything that requires thinking, not just generation.
What Is Claude Code and How Do I Use It?
Claude Code is the command-line version of Claude, extended with skills — pre-built workflows triggered by slash commands. I use it for structured, repeatable tasks where I want consistent output format every time.
The skills I use most often:
- /market-audit — full marketing audit of a website or campaign
- /market-competitors — structured competitive brief from a competitor URL
- /market-seo — SEO content audit for blog posts and landing pages
- /blog write — full blog post from outline to draft, SEO-optimised
The difference between Claude chat and Claude Code is the difference between asking a question and running a workflow. For tasks I do regularly, Claude Code produces consistent, structured output faster than a chat session would — and the output structure is always the same, which makes review faster too.
You can download the skills I use from the Skills page. Both the AI Marketing Suite and the Blog Content Engine are free.
How Do I Use Perplexity for Marketing Research?
Perplexity is my research tool. When I need current information — recent stats, competitor news, industry developments, or market context — Perplexity gives me sourced answers faster than any other tool I have tried.
The key advantage over Google Search is that Perplexity synthesises information from multiple sources and cites them clearly. I can ask "what is the current state of AI Overviews in Australia" and get a structured, sourced answer in seconds rather than spending 10 minutes on manual searching and source verification.
I use it to:
- Find recent statistics for content and client reports
- Research competitor activity and industry news
- Get quick background on topics before deeper work
- Verify claims before including them in content or recommendations
What AI Tools Did I Try and Drop?
Quite a few. Here is the honest summary.
Jasper and Copy.ai — useful for simple ad copy generation at one point, but the output quality did not justify the cost compared to using Claude with a well-written prompt. Claude produces better marketing copy when prompted correctly, without the separate subscription.
Notion AI — useful for quick in-document tasks like summarising or rewriting short sections. Not powerful enough for complex marketing work. I use it occasionally, not as a primary tool.
Various "AI marketing platforms" — most are wrappers around ChatGPT with a specific interface built on top. They add friction without adding real capability. I prefer working directly with the underlying model using my own prompts.
AI image tools — the quality has improved significantly, but I do not currently use AI image generation regularly. Stock photography platforms cover most of what I need without the prompt engineering overhead.
What Makes the Difference Between AI Tools That Stick and Ones That Do Not?
The tools I kept solve real, frequent problems. The tools I dropped either solved problems that were not painful enough to warrant a tool, or required more setup than the value justified.
The pattern is consistent:
- Tools that stick are used multiple times per day on tasks that are genuinely time-consuming
- Tools that get dropped solve one specific problem that comes up occasionally
- Tools that were never useful are solutions looking for problems
The best AI tools disappear into the workflow. You do not think of them as "AI tools" — you just work faster. That is the standard worth holding any new tool to before adding it.
What Should a Marketer's AI Stack Actually Look Like?
For most growth marketers, the right stack is simple:
- One strong LLM for general work — Claude or ChatGPT. Test both and pick the one whose output you prefer editing.
- Claude Code with skills for structured, repeatable tasks
- One research tool — Perplexity works well for current information
- A well-maintained prompt library — the most underrated component of the whole stack
The prompt library is what makes everything else faster. Once you have reliable prompts for your most common tasks, the per-task overhead drops to near zero. That is when AI actually starts to feel like genuine leverage rather than an interesting experiment.
If you want a starting point, the free AI Marketing Suite and Blog Content Engine on the Skills page are the skills I use in my own work. Download them, install them, and run one task. You will see the difference quickly.
Frequently Asked Questions
Should growth marketers use Claude or ChatGPT?
Both are strong for marketing work. Claude tends to produce more structured, direct output with less hedging — which makes it faster to edit for most marketing tasks. ChatGPT has a larger plugin ecosystem and wider third-party integrations. The practical answer: test both with the same prompt on a real task and use whichever output you prefer working with. Most marketers end up with a clear preference after a week.
How much does a useful AI marketing stack cost?
Claude Pro is $20 per month. Perplexity Pro is $20 per month. Claude Code charges per token used — typically a few cents per complex task with an Anthropic API key. You can build an effective AI stack for $40 to $60 per month. That is less than most SaaS marketing tools and replaces hours of work per week.
Is it worth learning AI tools as a marketer right now?
Yes — not because AI will replace marketers, but because marketers who use AI effectively will outproduce those who do not. The competitive advantage is real and it compounds over time. The marketers building AI into their workflow now will be significantly more productive in 12 months than those who are not. The time to learn it is before it becomes table stakes.
