The AI-Powered Marketing Stack: Tools We Actually Use in 2025
After testing 60+ tools across client campaigns, here’s the exact stack we use — and why most AI marketing tools are oversold, overpriced, and underdelivering.
Let me be straight with you: the AI tools market is a mess right now. Every week there’s a new “game-changing” AI marketing platform promising to 10x your output and halve your costs. Most of them are ChatGPT wrappers with a £99/month price tag.
We’ve spent the last two years testing over 60 tools across real client campaigns — from DTC ecommerce brands spending £30k/month on ads to B2B SaaS companies running complex lead gen. Here’s what actually stayed in our stack, what got cut, and why.
🔍 The honest filter we use: Does this tool save us more than 3 hours per week per client, or improve campaign performance by more than 15%? If the answer is no after a 30-day real test, it’s out.
Layer 1: The Intelligence Core (Research & Strategy)
Before you create a single ad or piece of content, you need to understand your market better than your competitors do. These tools are how we do that at scale.
In Our Stack
Our primary AI assistant for long-form strategy work, brief writing, and research synthesis. Claude’s ability to hold context across a long conversation and produce nuanced, brand-specific copy makes it consistently better than ChatGPT for agency work. The new Projects feature is genuinely useful for keeping client context organised.
In Our Stack
We use Perplexity for real-time research — competitor intelligence, market trends, campaign benchmarks. Unlike static LLMs, it pulls from live web data and cites every source. Invaluable for client strategy decks that need current data.
In Our Stack
Still useful for rapid ideation, ad copy angles, and structured output (when you need JSON or specific formats). Custom GPTs for specific client voices are genuinely time-saving. But for nuanced writing, Claude wins.
Layer 2: Creative Production
Creative is the biggest performance variable in paid social. Our creative production stack has changed more in the last 12 months than in the previous five years combined.
In Our Stack
Generates ad creative variants at scale and scores them before you spend a penny on media. We use it to rapidly produce static ad variants for Meta campaigns — particularly useful when a client needs 20+ creatives for a launch without the budget for a full shoot.
In Our Stack
AI video generation and editing. We use it to extend UGC clips, create transitions, and generate B-roll. Not for replacing proper video production — but for extracting more creative variants from existing footage. The quality has improved dramatically in 2025.
In Our Stack
Still the best text-to-image tool for conceptual work and mood boarding. We use it for lifestyle imagery, abstract brand visuals, and ad concept development before committing to a production shoot.
Cut After 30 Days
Expensive for what it does and the output quality isn’t materially better than Claude or ChatGPT with a good prompt. If you’re paying for Jasper and already using Claude, you’re paying twice for the same capability.
Layer 3: Automation & Workflows
This is where the real leverage is. The tools below aren’t “AI” in the traditional sense — but when combined with AI models, they turn repetitive marketing tasks into automated systems.
In Our Stack
Our automation backbone. We use n8n to connect everything — ad platform data to reporting dashboards, lead forms to CRM to email sequences, content briefs to publishing workflows. It’s more powerful than Zapier, cheaper at scale, and can be self-hosted. The learning curve is real (expect 15-20 hours to your first solid workflow) but the payoff is enormous. Marketing automation saves an average of 2.3 hours per campaign — multiply that across 20+ clients.
In Our Stack
We use Make alongside n8n for specific integrations where its native connectors are stronger. Its visual workflow builder is more approachable for clients who want to manage their own automations without technical knowledge.
Layer 4: Analytics & Attribution
The measurement layer is where most agencies are still flying blind. These tools have transformed how we understand what’s actually driving results.
In Our Stack
The foundation of every client analytics setup. GA4 has become genuinely powerful once you get past its initial complexity — especially for multi-touch attribution. We build every client a custom Looker Studio dashboard that pulls from GA4, ad platforms, and CRM data in one view.
In Our Stack
Session recordings and heatmaps remain invaluable for CRO work. When a campaign is driving traffic but not converting, Hotjar tells you exactly where the friction is on the landing page. AI-powered summary features added in 2024 make synthesising session data much faster.
The Honest Truth About AI Marketing Tools
Here’s what the vendor marketing won’t tell you: AI tools are multipliers, not magic. They make good marketers significantly more productive. They make bad marketers produce bad work faster.
The biggest mistake we see brands and agencies make is stacking tool upon tool without a clear workflow connecting them. You end up with a £2,000/month tool subscription bill and no measurable improvement in results — because the tools aren’t talking to each other and there’s no systematic process around them.
Start with one AI layer at a time. Master it. Integrate it into your workflow. Then add the next. The agencies winning with AI right now aren’t the ones with the most tools — they’re the ones who’ve built the tightest systems around the fewest, most powerful tools.
Want to see our full stack in action?
We’ll audit your current marketing toolset and show you exactly where AI automation can save you time and improve your results.