The Media Image, a premier global independent acquisition marketing agency, built a Meta co-pilot multi-agent system to boost operational efficiency and drive results for their clients.

The vision

The Media Image wanted to develop a multi-agent system that simplifies how marketers, business intelligence (BI) teams and partner managers access and act on insights from Meta.

The system would provide a single conversational interface that retrieves a user’s query, fetches the required data through Meta APIs, and transforms the results into clear, actionable insights or recommendations.

The agency built a multi-agent co-pilot system that allows users to retrieve data from the Meta API specifically for performance statistics, CAPI troubleshooting and recommendations. It consisted of:

• Chat UI. Accepts user inputs, sends queries to the orchestration agent, displays structured outputs and collects user feedback.
• Orchestration agent. Routes queries to specialised sub-agents, coordinating multi-step query planning and response consolidation.
• Performance statistics co-pilot. Pulls campaign, ad and ad set KPIs from the Insights API, calculates relevant metrics and produces structured summaries.
• CAPI troubleshooter co-pilot. Pulls dataset quality metrics and identifies areas for improvement.
• Recommendation co-pilot. Pulls recommendations from Meta and prioritises them by impact — including various Advantage+ options that will improve performance, as well as increasing opportunity score across the ad account.

The co-pilot also integrates with The Media Image’s other pre-existing tools, such as the Integration Quality Dashboard, which pulls all scores from client’s event managers into one view. From there, the agency could identify lower scores at a broader level, then ask the co-pilot to pull the recommendations from the CAPI troubleshooter and implement to improve CAPI integration effectiveness.

Driving efficiency and performance

The system was a success and resulted in meaningful improvements in key areas, including:

• Analyst productivity. Automating the retrieval of KPIs, CAPI issues and recommendations saved analysts hours per account per week, freeing staff up to spend more time on higher-value business tasks.
• Operational efficiency. Reduced time-to-insight and ticket-to-resolution for CAPI issues, significantly lowering operating costs.
• Revenue potential. By improving campaign performance and operational efficiency, the agency is able to handle more accounts or higher-spend clients without increasing headcount, effectively increasing margins and capacity for new business. Through increasing adoption of tools such as Advantage+ placements, audience and budget from Meta across the agency, they also see revenue increases for their clients as performance improves.
• Client retention: Enhanced service quality through faster insights and optimisation drove both higher client satisfaction and retention.

The numbers

In-depth A/B testing found that the co-pilot system drove the following results in December 2025:

A view to the future

The system is currently being rolled out across all clients at The Media Image, and the agency will continue to invest in it moving forward. Now armed with a tool that can pull significant insights, the next step will be continuing to feed the co-pilot system context, allowing it to read beyond just numbers (for example, the content of ad creative).

In an industry where in engine processes will become more and more automated, tools like these help The Media Image differentiate themselves from their competitors and provide additional value to their clients.

The goal is to continue improving the system’s efficiency and accuracy, allowing the agency to source and implement recommendations quickly and therefore invest more time in improving the performance of their clients’ accounts, including increasing spend on Meta technologies to help deliver better results. By increasing the adoptions of new and upcoming Meta best practices, The Media Image anticipates further improved performance, which will lead to incremental budget allocated to Meta technologies. Furthermore, the success of tools like these will lead to continued investment from the agency as they continue to widen their offering to clients, and go outside of the in-engine work to provide value elsewhere.