What We Actually Use AI For Inside a Google Ads + HubSpot Agency (And What We Don't)

People ask us a version of the same question all the time. If you run on AI, what is actually left for the humans? It is a fair thing to wonder, and most agencies answer it with a fog of buzzwords. We would rather be specific. This is what AI touches inside our work every week, and this is what it never gets near.

We are not precious about either list. Being specific is the point. It is the only way you can judge whether the way we work is the way you want your accounts handled.

Where AI earns its place

The work AI is genuinely good at tends to share a shape. It is repeatable, it is high in volume, and the right answer is either obvious or only needs a person to confirm it. That is most of the grind in any ads or CRM account, and it is exactly where we lean on it.

A few of the everyday jobs:

• Sorting search terms. A Google Ads account throws off hundreds, sometimes thousands, of search terms. AI sorts them into intent buckets fast, so we are looking at a clean shortlist of what to act on instead of reading rows one by one.
• Tidying and enriching data. Deduping records, standardising messy fields, filling gaps from public sources, flagging what looks off. The unglamorous data work that makes everything downstream more reliable.
• First-draft copy. Ad headlines, email outlines, the bones of a landing page. AI gives us a starting block to shape, not a finished piece. The draft is never the deliverable.
• Prep for reporting. Pulling the numbers together, spotting what moved week on week, surfacing the lines worth a closer look before we write anything up.
• Pattern spotting across accounts. When you run a lot of accounts, AI is useful for surfacing where something looks unusual and deserves a human to go and check.
• Analysis at scale. Given a clear framework to work inside, AI can chew through a large volume of data, structured numbers and unstructured text alike. Account exports, search query text, reviews, call notes, competitor pages. It reads far more than a person could in the time and hands back the parts worth a closer look.

In all of these, AI is handling the repetitive work that used to eat hours. None of it is the thinking. It is the legwork that clears the way for the thinking.

Analysis deserves a special mention, because it is where AI has come furthest. The reason it works for us is not the tool on its own. We bring the skills, the frameworks and the strategy, and we give the AI the context to work inside. With that in place it can take on a lot of data, structured and unstructured, and turn it into something we can actually act on. Hand the same tool no framework and no context and you get confident output that means very little. The framework and the read on what it means stay ours. The horsepower is the machine's.

Where it stays human, full stop

Then there is the work we keep on people, and we are deliberate about it.

• Strategy. What to do, in what order, and what to deliberately ignore. That is judgment about your business, your margins, your goals. AI will produce a plan if you ask it to. It will not know which plan is right for you.
• Diagnosis. Working out why an account is underperforming is detective work. It depends on context that is not in the data, like a seasonal quirk, a pricing change, a competitor who just moved. A person connects those dots.
• Client calls. The conversation where we work out what actually matters to you this quarter is not something we would ever hand off. It is where the real brief gets written.
• The judgment about which task is even right. This is the quiet one. Before any work happens, someone has to decide it is the correct work to do at all. AI is happy to do the wrong task beautifully. Catching that is a human job.
• Accountability. When the numbers need to move, someone owns that. AI does not stake anything on whether it was right, and it does not answer the phone when something needs to change.

The line is consistent. AI handles the execution. People decide whether it is the right work and stand behind the result.

The part most people miss: where the time goes

Here is the question underneath the question. If AI is handling the repeatable parts, are you just getting less of our attention for the same fee?

The opposite, actually. The time AI frees up does not get pocketed. It gets reinvested straight back into your account. The hours that used to disappear into sorting search terms and cleaning data now go into more analysis, more testing, more of the deeper work that genuinely moves results. You get more of the thinking, not less, because the grind is no longer eating the day.

That is the whole trade we are making. We let AI take the repeatable load so our people spend their hours on the parts that need a person. A client who works with us is not getting a thinner service dressed up in tech. They are getting more attention on the work that actually changes the numbers.

Why we are this transparent about it

We could wave a hand and say we are an AI-powered agency and leave it there. Plenty do. We would rather you know exactly which jobs the machine touches and which ones a person owns, because that is how you judge whether your accounts are in good hands.

The version of AI we believe in is not the one that replaces the team. It is the one that takes the grind off the table so the team can do the work you actually hired humans for. Speed on the repeatable, judgment on everything that matters.

If you have been wondering where AI fits in how your own marketing gets run, or you have inherited a setup and you are not sure what is being handled by a person versus a tool, that is a conversation worth having properly. It usually brings a lot of clarity, and clarity tends to be the thing that was missing.