I fell in love with complexity and it almost cost me a quarter of margin. 📉
In 16 years running a digital agency I have watched every technology wave roll in — from the arrival of Facebook Ads to the Big Data explosion. None blinded me as much as the wave of Generative AI. Just a few months ago, at GMC (Global Marketing and Commerce), we fell into a dangerous trap: the infinite-subscription trap.
We ended up with 25 active AI tool subscriptions. We looked like NASA. We had a prompt generator for every social network, three different meeting-transcription tools, four video generation platforms and a never-ending pile of plugins promising to “automate your life in one click”.
The reality was different. Our processes were slow, the team was mentally exhausted, and — most worrying — our clients' ROI was not moving at the same rate as our software bill.
The hidden cost of “Tech Shine”
Having 25 tools did not make us faster; it made us heavier. Every new tool meant a learning curve, another open tab, and a brutal fragmentation of attention. The team was spending more time deciding which tool to work with than actually working on the client's strategy.
That is when I decided to apply what I now call the “AI Diet”.
Step 1: The Value Audit
We sat down with the team and analyzed every subscription under a brutal filter: Does this tool save us at least 5 hours per week or measurably improve the client's ROI?
We discovered that 70% of our tools were redundant. We had tools doing the same thing as other tools, just with a slightly prettier interface. We were paying for “visual comfort”, not real efficiency.
Step 2: The 16-Year Filter
This is where a decade and a half of experience kicked in. Someone with two years of experience is dazzled by an AI that writes acceptable copy. Someone with 16 years knows that “acceptable” copy does not sell in a saturated market.
We decided that AI should not replace human judgment — it should amplify it. If a tool tried to “think” for us in a mediocre way, it was out.
Step 3: Consolidation into 3 Pillars
We trimmed our stack down to just three fundamental pillars that actually move the needle in a modern agency:
- Autonomous Execution & Operations: Manus handles the complex execution work across social and ads that used to require hours of manual work.
- Thinking & Code Architecture: Claude became our architect for structuring strategy and building custom solutions.
- Real-Time Research & Data: Perplexity replaced hours of manual Google searches, giving us immediate context for every campaign.
The Results: Less is a Lot More
When we simplified, something unexpected happened: creativity went up.
Without the noise of 20 different platforms, the team could focus on what actually matters — storytelling, consumer psychology and the human touch no AI can replicate. The result was a 40% improvement in our average campaign ROI, simply because we stopped “playing” with technology and started “using” it with purpose.
Cutting tools didn't just lower our operating costs (we save thousands of dollars a month in SaaS) — it eliminated “AI Brain Fry”, that mental fatigue of jumping between interfaces.
Lesson for agency owners and founders
If your tech stack looks like a Product Hunt catalogue, you probably have a bottleneck, not a competitive advantage. AI is a bicycle for the mind — but if you try to pedal 25 bikes at the same time, you will not get anywhere.
My advice today: Pick your three non-negotiable tools and master them. The rest is noise.