Everyone in media is talking about AI right now – usually somewhere between “this will change everything” and “please don’t let it break the workflow.”
Atlas AI, launched through Foxcatcher, is trying to land in a more practical place. Built on top of the company’s proprietary WorldView platform, the system serves as an intelligence layer spanning media trading and campaign analytics, unifying campaign data, audience behaviour, engagement signals, and performance modelling into a single operating environment.
The goal is less about replacing traders and more about removing the spreadsheet-heavy grunt work that slows teams down.
The company says Atlas AI has already reduced some operational inefficiencies by up to 70%, while giving teams faster reporting, deeper optimisation insights and real-time visibility across an increasingly fragmented advertising market.
Still confused? Don’t worry. Mediaweek sat down with Atlas AI’s Andrew Molan and Foxcatcher’s Varun John to get the low-down.
Foxcatcher’s Varun John and Atlas AI’s Andrew Molan
Mediaweek: Just to start, explain to me in the most basic terms (aka like I’m a five-year-old) what Atlas AI is?
Andrew Molan: Atlas AI is the intelligence layer that sits atop our proprietary media and analytics platform, WorldView.
In simple terms, it’s like having a really smart assistant sitting next to our trading teams. Instead of people manually digging through dashboards, spreadsheets and reports, Atlas AI helps surface the important information quickly so teams can focus on higher-level tasks.
Mediaweek: What has been a main focus for you?
Andrew Molan: Streamlining and automating legacy, repetitive, and time-consuming tasks that traditionally consume a huge amount of operational time. Things like reporting, campaign setup workflows, tracking implementation checks and identifying optimisation opportunities can now happen much faster.
Importantly, though, Atlas AI is designed to support our teams, not replace them. Human expertise, strategy and relationships are still at the centre of what we do.
Mediaweek: You say Atlas AI has reduced some operational inefficiencies by up to 70%. What specific functions are seeing those gains – and how are you measuring them?
Andrew Molan: The biggest efficiency gains so far have come from campaign setup workflows, tracking implementation, reporting, data analysis and campaign monitoring.
Historically, many of these processes relied heavily on our teams manually pulling together information from multiple systems. Atlas AI has made that information significantly more accessible across the business, meaning teams can get answers faster without relying on one or two specialists.
Reporting and optimisation workflows are another major area. Traders often spend a great deal of time navigating multiple platforms to piece together a complete picture of campaign performance. Atlas AI helps surface that information in a much more connected and actionable way.
We’ve been actively logging and measuring operational workflows internally. For example, in some cases, tasks that previously consumed a good amount of time on a Monday can now be completed within minutes. The gains are not just about speed, either; they’re about giving back the time to allow teams to focus on higher-value strategic thinking.
Mediaweek: The market is flooded with agencies and adtech businesses claiming to have “AI-powered optimisation”. What genuinely differentiates Atlas AI from the dozens of automation layers already sitting inside DSPs and martech stacks?
Andrew Molan: The biggest difference is that Atlas AI hasn’t been bolted onto an off-the-shelf product. It has been built on top of proprietary technology and data infrastructure that we’ve spent years developing, and we can truly call our own.
Most platform-level AI tools operate within a single ecosystem or optimise around a relatively narrow set of signals. Atlas AI operates across multiple data sources simultaneously, including media activation data, website analytics, audience behaviour, and our own internal finance processes.
That broader context is important because modern media performance can’t be properly understood within isolated platform silos anymore.
Another key difference is that Atlas AI has been designed around operational workflows inside a real agency and trading environment.
Many AI tools focus purely on optimisation outputs, whereas we’ve focused heavily on streamlining day-to-day operational processes that consume significant time across our programmatic team.
Importantly, Atlas AI is also designed around augmentation. The goal is not to hand everything over to AI. The goal is to give teams better visibility, faster access to information and more time to focus on strategic thinking.
Mediaweek: There’s been growing anxiety across agencies about AI replacing junior talent in the pipeline. How do you avoid hollowing out the next generation of traders and analysts?
Varun John: There’s no doubt that skillsets across the industry will evolve over time. Highly repetitive, manual tasks will increasingly be automated, and that’s already happening across many parts of the industry.
That said, we still see people as the most important asset inside the business. The role of AI should be to remove low-value repetitive work so teams can focus more on strategic thinking, analysis, creativity and client outcomes.
Junior traders and analysts still need to learn how the media works, how to interpret data, how to think critically and how to communicate insights effectively. Those skills become even more important in an AI-enabled environment.
The difference is that the next generation will likely spend less time manually downloading reports and extracting data and more time learning how to use intelligent tools effectively. We see AI as something that can accelerate capability, not replace it.
Mediaweek: Atlas AI is making optimisation decisions based on multiple data signals and proprietary modelling. Can advertisers actually interrogate how the system reaches its recommendations, or does it risk becoming another “black box” optimisation engine?
Andrew Molan: Transparency has been a really important focus for us. We’ve already been building proprietary models and analytics frameworks internally for years, so in many ways, that part of the process hasn’t fundamentally changed.
For us, AI is the ability to surface insights, connect signals, and make information more accessible across the business.
Atlas AI is designed to support traders and strategists, not replace human oversight. Recommendations are surfaced alongside contextual signals and supporting information so teams can understand why something may be happening before making decisions.
Ultimately, the human trader is still responsible for strategy and decision-making. Atlas AI helps accelerate the process of identifying opportunities, risks, and trends, but human judgment remains key to how recommendations are interpreted and applied.
Mediaweek: If Atlas AI works as intended, what does the media agency of 2030 actually look like – and which parts of media buying do you believe should never be automated?
Varun John: The media agency of 2030 will almost certainly be far more connected, automated, and AI-assisted than it is today, particularly from an operational perspective.
Many repetitive workflows that currently consume significant time across agencies will become increasingly streamlined. Things like reporting, campaign monitoring, anomaly detection, campaign setup and certain optimisation processes will become far more intelligent and proactive.
We also expect AI to become much better at identifying patterns and surfacing opportunities across fragmented datasets in an autonomous and methodical way.
However, there are parts of media and advertising that should remain very human. Strategy, creativity, client relationships, commercial judgement and understanding cultural nuances are incredibly difficult to automate properly.
For us, the future is not about replacing teams with AI. It’s about building technology that helps talented people operate with greater speed, visibility and confidence while allowing them to focus more on the areas where human thinking adds the most value.