To evaluate an agentic advertising platform, develop scoring criteria before demos, confirm the platform handles campaign reporting and activation (e.g., pacing, QA, creative swaps, etc.), and ensure deterministic rules govern every action that touches live spend. Goodway Group applied this approach when vetting 70+ agentic software providers prior to partnering with Fluency, and is now automating and scaling multichannel execution.

Agents aren’t just changing the advertising workflow; they’re changing how organizations evaluate software and systems. Instead of traditional evaluation criteria like ease of use and core features, the most important component of an agent-run digital advertising system is its ability to advance broader business objectives, like reducing the overall cost of media activation. Increasingly, that means governance: how everything works together, safely, at scale. 

At Advertising Week New York 2026, this concept was a recurring theme and a key focus of an AI & Innovation session track panel featuring Mike Wolk, SVP of Growth and Partnerships at Goodway Group. 

According to Wolk, “A lot of people are doubling down on the tech architecture they built five years ago, trying to make it work for where the future's going. On the other extreme, people are chasing the latest tools but not making meaningful progress."

Wolk and the Goodway Group team knew they needed a new way to evaluate agentic advertising solutions if they were going to find a scalable system that could grow with the company and evolving agentic tech.

Having used agentic advertising workflows within their organization since early 2026, Wolk shared his recommendations for how ad agencies evaluate different agentic advertising solutions. He also shared how Goodway Group ultimately chose Fluency’s Agentic Advertising Operating System to help manage hundreds of client campaigns across three channels. 

Establish your strategy first: decide what to build and what to partner on before evaluating any agentic advertising platform

According to Wolk, start by deciding what to build in-house and what to partner on before evaluating vendors. For instance, Goodway Group chose to own the work that gave them strategic differentiation. They wanted to hire a software partner to help them manage cross-channel task execution and campaign activation. This approach keeps your people focused on what sets your agency apart while using software to handle routine high-volume tasks. 

Like many large agencies, Goodway had already built some in-house workflow automation. But with the market moving this fast, the opportunity cost of continuously building their agents was high. Maintenance, governance, and system integration would divert resources and focus away from Goodway’s capacity to deliver client value. A partner with the right architecture already built represented an opportunity to integrate the entire workflow and see value quickly.

"We've always been really deliberate about what we want to build where the market's not offering it, or where we could bring incremental value, and then where we partner," said Wolk.

Starting with a vision of how the company could split their advertising workload made it easier to build their agentic software evaluation criteria. At the top of their must-have software list was a tool that would own more day-to-day task execution so their teams could orchestrate campaigns better, from a higher level

Create clear agentic advertising vendor evaluation criteria before a single demo

Wolk suggests writing your agentic advertising software scoring criteria before you schedule any vendor demos: what you need to accomplish, by when, and at what cost. Assign points to each criterion so each vendor is scored the same way. Using this methodology, Goodway Group scored more than 70 agentic advertising partners as a cross-functional group.

Here is an example of what an evaluation criteria checklist could look like: 

‍

In their research, Wolk and the team found that most AI and agentic software partners fell into one of three categories: 

1. Specialized tools that solve an individual offering or empower a larger capability, such as research, planning, or creative work.
2. Agents and AI dashboarding tools that report on results and surface insights.
3. Generalist or catch-all solutions that lacked the sophistication the team needed to take over day-to-day campaign execution across channels without sacrificing performance.

When it came to media infrastructure, the team prioritized solutions that had been proven or scaled past the incubation and early start-up stages. It was important to them to find a solution that balanced governance and rigor with a clear roadmap for incorporating more agentic and automation capabilities over time. Operationally speaking, the Goodway Group wanted a solution to reduce the repetitive work taking up too much time, especially for tasks requiring very little strategic judgment. 

Pacing, QA, and creative swaps are all clear candidates to hand off to agents because these AdOps tasks follow strict rules that can be documented, require little judgment, and are most valuable at high volume: rapid, constant changes improve the odds of the right outcome.

But this also raised a new question in the team’s agentic advertising solution evaluation: once an agent recommends a change to a live campaign, what decides whether or not the recommended change goes through?

Prioritize governance: put a deterministic layer between AI recommendations and live spend

Agentic platforms must apply strict governance standards to support campaign execution at scale. That means deploying LLMs strategically with governance between AI recommendations and live spend. 

LLMs are probabilistic by design, varying their output even when given the same input. Assigning AI to autonomously manage tasks that should have set, predictable, or repeatable outcomes is risky because it’s very hard to get AI to take the same exact actions every time. Deterministic execution, on the other hand, follows explicit rules: the same input produces the same output every time, and every action is auditable.  

With its “own less, orchestrate more” philosophy, the Goodway Group team stressed governance as a critical component of their evaluation. They wanted to keep AI in an advisory role and rules-based agents in charge of spend. 

Iterate and learn at scale: pilot on your highest-volume accounts

In an effort to learn and scale agentic workflows quickly, Goodway Group started with their largest multi-location clients. Piloting agentic workflows on a small, safe account gives you forgiving margins that hide the kinds of problems that matter at scale. Piloting AdOps agents to run key workflows across your highest-volume accounts, however, shows how they respond under real pressure.

When Goodway Group adopted Fluency’s Agentic Advertising Operating System (AAOS) to manage specific workflows, they kicked off this new approach with a large, complex multi-location client.  The team built confidence in these agentic workflows through early due diligence and monitored early launches closely: running a first batch of campaigns, confirming the system does exactly what it’s set up to do, then handing over more.

The team worked with Fluency to build agentic advertising workflows to help them manage budget pacing for a key client. Fluency’s automated budgeting agents ensured accurate spend across these campaigns, instead of relying on Goodway Group’s teams to check pacing once or twice a week.

Treat adoption as a discipline: change management is critical

In the long run, using agents means building something better and more efficient for human teams. In the short run, though, it can mean asking experienced practitioners to change deeply ingrained habits, such as moving where the team stores its creative assets. 

That’s why Goodway Group took a “move deliberately before scaling” approach, according to Wolk. They understood that a successful agentic automation integration required a willingness to break their existing processes. 

The team put time into building their agentic workflows correctly from the very beginning and setting best practices before moving everything over to agents. This allowed them to build on a stable foundation, so each new launch and test took less effort than the one before.  

Get key practitioners involved first and early

Getting the right people in the room, right off the bat, was another critical component of their shift to agentic advertising operations. Even with mandatory training, not everyone is on board right away, so Goodway started by involving its best practitioners from the beginning.

Wolk described the company’s approach during the session at Advertising Week 2026: build a cross-functional group that is informed (subject matter experts) and empowered (able to move past legal and IT), with an executive sponsor to make sure the work gets done.

The company’s practitioners reviewed the tool, confirmed it would deliver significant value, and became internal champions for the rollout. They also went through Fluency Academy, a self-service onboarding education course designed to help new Fluency users get up and running with confidence. Using the tool in a guided environment was more useful than sitting through a demo, enabling Goodway Group to get rolling quickly. 

Optimize around outcomes, including operational time savings and client performance

Goodway Group launched a national fitness franchise’s campaigns across 143 locations in one day via Fluency, with zero manual campaign builds. According to Wolk, Fluency’s agentic advertising solutions “inverted” how the team spends its time. Now, they spend about 90% of their time on strategy and 10% on implementation.

With agents managing routine, repetitive, and high-volume tasks, the Goodway Group team has more time and energy for strategic client work. 

“Testing and learning becomes so much easier once you have this infrastructure in place,” Wolk said. “Our team can have different kinds of conversations with clients because of how we architected our operations.”

The company can now roll out winning tests across every campaign with just one click instead of implementing the change in one-off campaigns. Fluency’s bulk editing and multi-platform management capabilities, all powered by agentic workflows, have helped the team take on twice the workload per “power user” compared to a year ago, according to Wolk. 

How to build your own agentic advertising strategy

Turning these evaluation and rollout decisions into a plan for bringing agentic advertising into your own advertising operations takes time and a clear framework.

The Agentic Advertising Playbook walks you through how to carry out each of these steps within your own advertising organization. It covers the necessary architecture to keep AI recommendations separate from live campaign execution, the true cost models behind the build-vs-adopt decision, eight workflows to hand to agents first, real-world production case studies, and how to prepare your team for the shift.

‍

Frequently asked questions about choosing an agentic advertising system

What is the difference between deterministic and probabilistic agents in advertising?

Deterministic agents follow fixed rules, so the same inputs produce the same outputs and every action is auditable. Probabilistic agents, powered by LLMs, interpret data and generate recommendations; their outputs can vary. 

Which advertising operations tasks should you automate with agents first?

Assign agents to manage repetitive, rule-based work first: AdOps workflows like budget pacing, campaign QA and naming, creative swaps, and recurring multichannel reporting are excellent first use cases. These tasks follow rules that can be easily documented for agents to read, and they require little strategic judgment. 

Which advertising operations tasks should an LLM handle versus deterministic software?

Let LLMs handle interpretation, recommendations, and creative AdOps tasks. Deterministic software, like agentic automation, should execute high-volume workflows that require precise, repeatable outcomes. LLMs are expensive per transaction, and millions of campaign changes a month can make the economics unworkable. 

Will agentic advertising replace media buyers and traders?

Agentic advertising changes the kind of work media buyers and traders do, but agents will not take or replace their jobs. For example, agents can execute repetitive tasks (within strict rules set by human strategists) for workflows like campaign/account setup, budget pacing, and reporting. Fluency’s agents helped Goodway Group's team invert their workload, with the team now spending about 90% of their time on strategy and 10% on implementation.

What should an agency fix in its processes before adopting an agentic platform?

Before you adopt an agentic advertising platform, audit your entire campaign launch and management workflows. Fix any broken steps (agentic automation solutions will scale errors rather than resolve them), standardize campaign naming, centralize creative assets, and set clear goals so you have a strong foundation for agentic workflows to take over seamlessly. 

What should you ask agentic advertising vendors before buying?

Ask agentic advertising vendors where the platform sits within the campaign management workflow (e.g., planning, change execution, reporting), how its agents decide what actions to take, and who stays in control. You should also ask about which channels integrate into the single system and if teams can override or shut off agent workflows.

How do you roll out agentic advertising automation without disrupting live campaigns?

Before using agentic advertising solutions on live campaigns, have your best practitioners audit your campaign launch and management workflows for any naming convention issues or gaps in the workflow. Rule-based tasks, like budget pacing, are a good starting point for agentic advertising automation pilots. Run an initial campaign batch and confirm the system does exactly what it was set up to do before handing over more work to agents.

‍