How Media Buyers Can Win in the Age of AI

AI is making media buying easier. The harder questions remain: what deserves attention, what creates value, and did the advertising actually make a difference? That is where a good media buyer earns their place.

For years, knowing your way around an advertising platform was a meaningful advantage. It still matters. But if your entire proposition is knowing which settings to change, it is worth asking how much of that work a client will need to buy from you in future.

How is AI changing media buying?

AI is automating parts of campaign execution and helping buyers process information faster. For example, Google’s Smart Bidding uses AI to optimise for conversions or conversion value in individual auctions. That is a specific capability, not evidence that a platform understands everything about a business.

Someone still has to decide which outcomes matter, whether the data represents them accurately and what the result means. Our view at Second Order is that this moves the buyer’s value closer to the business: its customers, its economics and the decisions that shape growth.

What skills do media buyers need in the age of AI?

The most useful combination is commercial understanding, measurement literacy, creative judgment and practical AI capability. These skills help you give automation better direction and challenge its output.

1. Understand what a conversion is actually worth

A cheap enquiry is only useful if it has a reasonable chance of becoming a valuable customer. Speak to sales. Understand qualification, close rates, margin and the time between an initial enquiry and revenue. Review the leads that went nowhere as carefully as the ones that converted.

Consider a hypothetical comparison: 100 enquiries at $20 each produce two customers, while 40 enquiries at $50 each produce eight. Both campaigns cost $2,000. The first has a lower cost per lead; the second has a lower acquisition cost, at $250 rather than $1,000. Customer margin and retention still need checking before deciding which campaign to scale.

Where the setup supports it, feed reliable downstream outcomes back into campaign measurement. If sales volume is too sparse, validate an earlier event against later outcomes. An easily measured action is not automatically the right objective.

2. Develop a point of view on creative

AI can help explore hooks, angles and variations. Your job is to decide which ideas deserve attention. What problem does the audience recognise? Why should they believe the promise? Does the work feel relevant to their world?

Read comments. Listen to customer conversations. Look at the landing page on your phone. Taste becomes commercially useful when it helps you recognise an unclear proposition, an unconvincing claim or an experience that makes the customer work too hard.

Fifty versions of a weak idea are still a weak idea. Test distinct hypotheses: convenience versus confidence, for example, or a product demonstration versus customer proof. Give each test a clear question. This is closely connected to the human skills AI cannot buy.

3. Question what the dashboard is telling you

Attribution assigns credit to advertising touchpoints. Incrementality asks whether advertising caused an outcome that would otherwise not have happened. They answer different questions.

Start by checking platform reporting against business results. Account for reporting delays, duplicate conversions and the time customers need to decide. When budget, volume and setup support it, use a properly designed holdout or geographic experiment to investigate additional impact.

These tests need care: small samples, seasonal shifts and differences between markets can distort the comparison. If the evidence is inconclusive, say so. Explaining uncertainty is part of the job.

4. Use AI to create more time for thinking

Start with tasks that are repetitive and easy to review: organising research, drafting a reporting narrative or flagging anomalies for investigation. Check calculations against the underlying data and verify any factual claims before sharing the result.

Keep approval over live budget changes until you have a tested process, explicit permissions and clear spending limits. Use approved environments for client information. The useful question is whether the workflow frees up attention without losing accountability.

How can media buyers show their value to clients?

Bring a recommendation that connects evidence to a commercial decision. Explain what to scale, stop, fix or investigate, why you recommend it and what would change your mind.

A report saying “cost per lead fell” leaves the client to work out whether that matters. A recommendation explaining that lead costs fell while qualification deteriorated, and proposing a change to the conversion signal, is much more useful. It gives the client a decision and a reason.

That is the thinking behind our Human Intelligence Index: making campaign signals useful to the next decision, with human attention and context shaping the interpretation.

A practical 30-day plan for media buyers

1. Week one: audit the signal. Trace an important conversion from the ad to the business outcome. Check duplicates, missing events, qualification and reporting delays. Agree who owns data accuracy.
2. Week two: automate one repeat task. Choose reporting preparation or research organisation. Review the output against a manual version and record the errors before expanding its use.
3. Week three: launch a meaningful test. Define the audience problem, creative hypothesis, success metric and decision rule in advance. Allow for the conversion cycle rather than calling a winner from early results.
4. Week four: make a commercial recommendation. Explain the evidence, trade-offs and next action. A month may be enough to improve the process without being enough to establish a reliable campaign winner.

Will AI replace media buyers?

AI can automate tasks within media buying. That does not establish that it can replace every responsibility of the role. The practical risk is building a career around execution that becomes easier to automate. Developing commercial, creative and measurement judgment gives buyers a broader contribution to make.

Should media buyers learn to code?

Coding can help with data handling and workflow automation, but it is not the only route to becoming more useful. Start by understanding the problem, the data and the decision. Learn the technical skills that help you solve that problem reliably, whether through code or existing tools.

What should you change first?

Check whether the outcome your campaigns optimise for reflects what the business values. If that connection is weak, faster execution can simply repeat the wrong decision more efficiently.

At Second Order, we combine paid media planning and execution, commercial advisory and human judgment. We use AI to expand what we can do, while staying accountable for what we choose to do.

Automate the execution. Own the judgment. If you want to understand what that could look like for your brand, let’s find 20 minutes to talk.