By Zach Lenchner, Sr. Director, Product Management at QuantumPath | JWX
There is a lot of noise around AI in advertising right now. Most of the conversation is focused on what AI can generate: copy, creative, recommendations, media plans, and strategy.
Some of that is useful. A lot of it also creates risk.
Advertising is not a blank page problem. A campaign setup, media plan, PMP, reporting view, or DSP integration already has structure. The work may be complex, but it is not undefined. There are platform rules, required fields, budgets, dates, targeting decisions, approval paths, and business logic that determine whether something is usable or not.
That is an important distinction for how we think about AI at QuantumPath by JWX.
We are not asking AI to invent how advertising should work. We are using it to make existing advertising constructs easier to access, faster to operate, and more consistent across teams and platforms.
AI has to be grounded in the work
A lot of bad AI experiences come from asking a model to answer without enough context. The prompt is vague, the response is confident, and the output may or may not reflect how the business actually works.
That is not the approach we are taking.
QuantumPath by JWX starts with the operating objects that already exist in programmatic advertising. Campaigns, advertisers, IOs, line items, DSP objects, supply partners, PMP packages, reporting metrics, and governance rules are not abstract concepts. They are the objects teams work with every day.
The value of AI comes from helping users move through those objects with less manual effort and more consistency.
A client brief can be turned into a structured campaign. A campaign objective can be connected to the right inventory path. Setup issues can be caught before anything is pushed into a DSP. Reporting can move beyond a static performance table and help explain what needs attention.
None of that requires the system to guess. It requires the system to understand the structure of the work.
The source of truth matters
The most important question is not whether AI can produce an answer. The better question is what the answer is based on.
In QuantumPath by JWX, the answer should be tied to the actual operating layer of the business. That means campaign hierarchy, platform mappings, approved vendors, setup logic, reporting data, performance signals, user permissions, and workflow status all matter.
This creates a very different product than a general AI chat experience.
When the line item structure is already defined, AI is not being asked to make one up. When the DSP integration determines which fields matter, the system is not relying on a generic interpretation of platform setup. When supply recommendations are based on approved SSPs, PMP rules, performance data, and client constraints, the recommendation becomes grounded in the business instead of generated in isolation.
That is how QuantumPath by JWX reduces hallucination risk. The AI is operating inside a controlled system with known objects, known rules, and known constraints.
Making complex systems easier to use
The real opportunity for AI in ad tech is not only automation. It is accessibility.
Programmatic advertising has a lot of operational knowledge trapped in people’s heads, spreadsheets, naming conventions, screenshots, QA docs, and platform workflows. Experienced traders know what matters, but that knowledge does not always scale cleanly across teams. New users need training. Different teams interpret setup rules differently. Quality often depends on who is doing the work and how much context they have.
QuantumPath by JWX is designed to turn that complexity into a structured operating layer.
A user should not need to know every DSP field, every PMP rule, or every reporting dependency to make progress. The product should guide them through the right steps, expose the right decisions, and make it clear where human judgment is still required.
AI becomes useful when it helps people understand the system faster. It can reduce repetitive setup work, explain tradeoffs, identify what changed, and surface issues that would normally require manual review.
The product still needs rules, controls, and a source of truth. Without that, AI just becomes another place for ambiguity to enter the workflow.
Efficiency comes from removing unnecessary work
Efficiency does not come from AI being clever. It comes from removing work that should not have to be repeated manually.
Today, too much time is spent translating the same intent across briefs, naming conventions, line items, inventory decisions, QA checks, and platform-specific setup. Every handoff creates room for inconsistency. Every manual rebuild creates room for error.
QuantumPath by JWX reduces that effort by making the workflow more structured from the beginning.
The same logic applies to performance. Pattern recognition only matters if it connects back to something a user can act on. A recommendation needs to map to a setting, a supply path, a budget shift, a targeting change, a PMP decision, or a reporting insight.
Otherwise, it is just commentary.
The goal is not to create more opinions. The goal is to help teams make better decisions faster, with the right context in front of them.
Human oversight is part of the design
There are places where automation makes sense. There are also places where approval, review, and judgment matter.
QuantumPath by JWX is not built around the idea that AI should fully replace the operator. It is built around the idea that the operator should not have to fight the system to do high-quality work.
Recommendations should be explainable. Setup should be auditable. Decisions should tie back to inputs. Users should understand why something was suggested and what will happen if they accept it.
This is especially important in advertising because the stakes are real. Budgets are real. Client requirements are real. Platform differences are real.
AI should help manage that complexity, not hide it.
The practical future of AI in ad tech
The companies that get AI right in advertising will not be the ones that add a chat box and call it innovation. They will be the ones that understand where the work actually happens.
Planning, setup, activation, governance, optimization, reporting, and communication all happen across platforms, teams, and constraints. That is the environment AI has to support.
QuantumPaths’s approach is to use AI inside that reality.
We take the constructs that already drive the business and make them easier to use. The workflow gives AI something reliable to operate against. Automation is used where it reduces manual work. Recommendations are valuable when they are tied to data and rules. Human review stays in the places where judgment matters.
That is how AI becomes useful in programmatic advertising.
Not by guessing.
By making the operating model smarter, faster, and more accessible.