By Zach Lenchner, Sr. Director, Product Management at QuantumPath | JWX
There is a lot of noise around AI in advertising right now.
Some of it is exciting. Some of it is useful. A lot of it is still too vague.
When people talk about AI in programmatic, the conversation often jumps straight to big ideas. Autonomous buying. Fully automated optimization. AI agents running campaigns end to end. Machines making every decision.
Maybe some of that happens over time. But that is not where most traders need help first.
Traders do not need AI that sounds impressive in a pitch. They need AI that makes the actual work easier. They need fewer clicks, better QA, cleaner setup, useful alerts, clear recommendations, more control, and less time spent managing repetitive tasks that should not require as much manual effort as they do today.
That is where AI can start creating real value.
A lot of programmatic operations is not strategy. It is the work around the work.
Taking a brief and turning it into platform setup. Applying naming conventions. Checking required fields. Building insertion orders and line items. Assigning targeting. Making sure budgets and flights are correct. Confirming that brand safety, inventory, geo, frequency, and device settings match the plan.
Then checking it all again.
Then making changes.
Then checking those changes.
That is where a huge amount of time goes.
Not because traders do not know what they are doing. They do. The problem is that too much of the workflow still depends on them manually carrying information from one place to another and remembering every rule along the way.
AI should help reduce that burden. Not by taking away control, but by doing a better job of organizing the work, spotting issues, and helping traders move faster with more confidence.
Before AI starts making big optimization decisions, it should help make sure the campaign is set up correctly.
That sounds basic, but it matters.
A lot of campaign problems start before the campaign ever goes live. The wrong budget gets entered. A targeting setting is missed. A naming convention is applied inconsistently. A required field is left blank. A campaign launches with a setup that does not fully match the approved plan.
These are not glamorous problems. They are operational problems, and they are exactly the kinds of problems AI should help solve.
If a media brief says one thing, the campaign setup should reflect it. If a naming convention is required, the system should enforce it. If budget control needs to happen at the IO level or line item level, the workflow should make that decision clear. If required inputs are missing, the trader should know before the campaign reaches the DSP.
That is useful AI.
It is not trying to be clever for the sake of it. It is helping the work get done correctly.
Most teams already have QA processes.
The issue is that QA is often manual, inconsistent, and dependent on how much time people have. A manager or another trader checks the setup against the brief. They review budgets, targeting, naming, brand safety, and other settings. They try to catch anything that looks wrong before launch.
That process is important, but it has limits. AI can make QA more consistent by checking the setup against the approved plan, flagging gaps, identifying conflicts, and surfacing anything that looks unusual. It can help separate real risks from acceptable variations. It can also explain why something is being flagged, so the trader understands the issue instead of just seeing another generic warning. That last part matters.
Traders do not need more noise. They need better signal.
A useful AI agent should not flood the team with alerts that nobody trusts. It should help prioritize what actually needs attention and make it easier to understand what changed, what is risky, and what needs to be fixed.
Every operations team has seen alerts that eventually become background noise.
Too many notifications. Too many false positives. Too many warnings that do not explain what actually needs to happen. That is not helpful.
For AI to matter in programmatic operations, alerts need to be tied to real workflow decisions. A campaign is pacing behind. A line item stopped spending. A deal is not delivering. A setting changed directly in the DSP. A campaign is spending against inventory that does not match the approved plan. A budget shift created a new risk. A performance issue is getting worse and needs attention.
Those are the kinds of alerts traders can use. But the alert itself is only part of the value.
The system should also help explain what happened, why it matters, what the likely cause is, and what action the trader may want to take next. That is where AI becomes much more practical. Not just “something changed.” More like: “This changed, here is why it matters, here is what we think caused it, and here are the recommended next steps.”
That is the difference between a notification and useful operational support.
One of the biggest mistakes in the AI conversation is assuming that traders want to give up control.
Most do not. They want better tools. They want software that handles repetitive work, catches mistakes earlier, recommends smart actions, and helps them manage more complexity without losing visibility. But they still need to understand what is happening. They still need to approve important decisions. They still need to override the system when needed. They still need a clear audit trail when changes are made.
That is especially true in programmatic, where small setup decisions can have real budget, client, and performance implications.
AI should not feel like a black box.
It should feel like a strong operator sitting beside the trader. One that understands the plan, watches the details, flags risk, recommends action, and keeps a record of what happened. The trader should stay in control. The system should make that control easier to manage.
At QuantumPath by JWX, we think AI should start with the actual operational problems traders face every day.
Cleaner setup. Better QA. Stronger governance. Useful alerts. Clear recommendations. Less repetitive work. More visibility across platforms.nThat is the foundation.
The opportunity is not just to make AI sound impressive. The opportunity is to make campaign operations meaningfully better.
For traders, that means less time spent clicking through repetitive steps and more time focused on the decisions that actually need their judgment.
For managers, it means more confidence that campaigns are being built, checked, and governed consistently.
For the business, it means a more scalable way to operate across platforms without depending on every detail living in someone’s head.
That is what traders actually need from AI.
Not magic.
Not a black box.
A better way to do the work.