Getting Your Ducks in a Row with AI

Anthony McGrath • February 23, 2026

AI hasn’t replaced marketing strategy. It’s just made weak strategy more expensive — faster.


Automation doesn’t understand your business. It doesn’t know your margins, your sales cycle, or what a good lead actually looks like. All it does is scale the decisions you give it.


Get those decisions right, and AI is a growth engine.
Get them wrong, and it simply accelerates waste.

AI Is an Accelerator, Not a Brain

AI only performs as well as the thinking behind it. Before automation ever goes live, there needs to be clarity on:


  • The single primary business objective
  • What success looks like beyond platform metrics
  • Which outcomes actually drive revenue


When AI is pointed at a clear destination, it performs exceptionally.
When it isn’t, it optimises noise.


Strategy → Inputs → Automation → Outcomes






A B2B SaaS company switched on automated bidding without tightening conversion goals. Performance looked strong — demos everywhere.


The demos came from students, competitors, and people “just having a look”.


After switching optimisation to revenue-qualified leads only:


Cost per lead increased slightly (+7%)

Revenue per lead jumped +39%

Sales stopped asking, “Why are we calling these people?”


That’s automation working properly — once the inputs were fixed.

✅ Automation Readiness Checklist

✔ One primary business objective defined (not multiple competing goals)


✔ Success defined in business terms (revenue, pipeline quality, ROAS)


✔ Conversion actions mapped to actual intent, not activity


✔ Low-value signals removed or deprioritised


✔ Budget aligned to the objective (no “test budgets” expecting scale)


Rule:
If you can’t explain what “good” looks like in one sentence, automation isn’t ready.



AI Optimises What You Reward.


AI doesn’t “figure things out”. It follows signals.

If you reward volume, it delivers volume.
If you reward low-intent actions, it finds more of them.


High-performing accounts are ruthless about inputs:


  • Only high-intent conversions are optimised
  • Low-value signals are removed
  • Platform goals reflect real business outcomes, not vanity metrics


Better signals = better optimisation. Every time.


 



A service business removed micro-conversions and optimised only for revenue-driving enquiries.


The result:

Revenue per lead increased 41%

CPA nudged up slightly (+6%)

Sales complaints dropped dramatically

(a rare but meaningful KPI)

If you optimise for junk, you don’t get better junk.
You just get it faster.


Where Automation Wins — and Where Humans Do

🧠 Human Oversight Checklist



(What should never be left fully automated)


✔ Weekly review of search terms and placements


✔ Weekly creative fatigue and message relevance check


✔ Monthly conversion mix and intent review


✔ Monthly budget distribution check


✔ Quarterly account structure and strategy review


Rule:

Automation removes busywork, not responsibility.

Creative Is Still the Biggest Lever


Automation can scale creative.
It cannot create a strong idea.


Performance stalls rarely come from bidding or budgets — they come from creative fatigue or weak messaging.


Winning brands:


  • Test concepts, not just formats
  • Refresh messaging before performance drops
  • Use real customer language, not internal assumptions


"AI amplifies creative impact. It doesn’t replace it."

Guardrails Aren’t Anti-AI — They’re Pro-Profit


When AI spends at scale, guardrails protect performance.


Strong accounts consistently apply:


  • Clear account structure to protect intent
  • Account-level negatives to stop bad spend early
  • Brand exclusions where cannibalisation exists
  • Regular insight reviews to correct learning

Automation without guardrails doesn’t mean freedom.


It usually means drift.


Here’s what profitable AI setups still rely on:


🛑 Guardrails & Risk Control Checklist


(Stops AI learning from noise)


✔ Clear campaign and account structure


✔ Account-level negatives applied


✔ Brand separated or excluded where necessary


✔ Placement exclusions maintained


✔ Learning signals reviewed regularly


Rule:

Guardrails aren’t anti-AI — they’re pro-profit.

Automation without supervision isn’t smart. It’s risky.

AI Scales Good Decisions — and Exposes Bad Ones


When strategy, inputs, creative, and guardrails are aligned, AI becomes a genuine growth lever.


When they’re not, it doesn’t fail quietly.
It fails fast.


  • More spend.
  • More noise.
  • More frustration.

That’s why AI success isn’t about trusting the platform more — it’s about being clearer, more disciplined, and more intentional than ever.

The Bottom Line -

AI isn’t the strategy. It’s the accelerator. If your foundations are solid, automation helps you grow faster. If they’re not, it simply gets you to the wrong outcome sooner.

Ask yourself:



Do I trust what AI is optimising towards?

Would I be happy if spend doubled tomorrow?


Can I explain current performance to a stakeholder clearly?


If the answer to any of those is no, automation needs intervention.

Want to Know What AI Is Really Optimising in Your Account?


If automation is spending your budget, you should know why — and whether it’s scaling the right decisions.


If you want results, not promises or excuses, we should talk.

Recent Posts

By Anthony McGrath April 16, 2026
What Exactly is Zero-Party Data? The term refers to data that a customer intentionally and proactively shares with a brand. To understand its power, let's look at the data hierarchy: Third-Party Data: Data bought from aggregators (often messy and unreliable). First-Party Data: Data you collect via your own channels (purchase history, site behavior). Zero-Party Data: Data the customer tells you (preferences, sizes, tastes, and intent). "Zero-party data is the holy grail of personalization. It’s the difference between guessing someone likes blue because they hovered over a link, and knowing they like blue because they told you it's their favorite color." Why the Shift is Happening Now The "Data Privacy Revolution" isn't just a trend; it's a structural shift in how the internet functions. Regulation: GDPR and CCPA have put the power back in the hands of the consumer. Tech Giants: Apple’s App Tracking Transparency (ATT) and Google’s phasing out of cookies have made traditional tracking nearly impossible. Consumer Expectations: Modern shoppers want personalization, but they want it without the feeling of being watched.
By Anthony McGrath April 16, 2026
Mastering Omnichannel Synchronization in 2026
By Anthony McGrath April 15, 2026
In the high-stakes world of business, there’s a flashy obsession with customer acquisition. We love the thrill of the hunt—the new lead, the signed contract, the "Welcome!" email. But if you’re focusing all your energy on filling the bucket while ignoring the giant holes in the bottom, you aren't growing; you’re just sprinting in place.  The real heavy hitters know that the real money isn't made on the first sale. It’s made on the tenth.
By Anthony McGrath April 15, 2026
Why Vertical Specialization is the Ultimate Growth Lever for Digital Agencies
By Anthony McGrath February 23, 2026
Digital marketing is one of those terms that feels obvious until someone asks you to explain it properly.  At that point, most explanations collapse into: “Running ads” “Posting on social media” “Doing SEO” Or the vague classic: “being online” None of those are wrong — but none of them are the whole picture either. Digital marketing is how businesses attract, convert, and retain customers using digital channels. Simple definition. Complicated execution.
By Anthony McGrath February 19, 2026
By 2026, Performance Max mistakes are rarely technical. They’re strategic. The platform does exactly what it’s told — sometimes a little too well. Here are the most common Performance Max optimisation mistakes we still see, with examples that explain why results go sideways.