What are the real considerations and impacts of AI on agency fees?

An increasing number of agencies are including promises of AI benefits in their pitches. Typically, these are framed in the context of improved productivity through automation, allowing the agency to provide more for less. A.K.A. – savings.

This is a valuable opportunity for any marketer facing heightened demand, a tight budget, or any procurement team focused and measured on cost reductions.

However, the claims of savings are becoming increasingly unbelievable, with some boasting of up to 60% more in savings than what the marketer is currently paying. It’s not just content and creative agencies involved; media agencies also offer lower fees and, in some cases, no fees at all, arguing that automation has substantially reduced their labour costs and, therefore, the fees required to provide those services. Additional factors beyond automation efficiency come into play with media agencies, including non-transparent payments from media sellers and principal-based trading. For this discussion, let’s concentrate on content creation.

ChatGPT was launched in November 2022, and in this brief period, agencies have raced toward the shiny new offering. Major holding companies and groups have announced significant investments in AI technology amounting to millions and even, in some cases, billions of dollars.

While many herald this integration of Artificial Intelligence (AI) and automation within advertising agencies as a pathway to increased efficiency and reduced operational costs, the underlying investments required for AI adoption necessitate a re-evaluation of agency compensation models. This often means that the anticipated cost reductions may not materialise or may even lead to new cost structures.

While AI undoubtedly offers the potential to streamline workflows, enhance creativity, and provide deeper insights, the assumption that its application will automatically translate to lower client costs is a significant oversimplification.

Embracing change within the agency

There is an early mover advantage for any agency in adopting AI, through the faster generation of insights, more agile strategy iteration, and adaptable marketing processes. However, implementing AI solutions in an advertising agency involves several key considerations, including:

1. Strategic Planning and Roadmap: Agencies must establish a clear roadmap for AI adoption in marketing. This entails setting up a cross-functional AI council to provide direction and drive strategy. The council should concentrate on risk management, upskilling talent, and fostering trust. It also needs to review existing client contracts and collaborate with clients to identify and mitigate risks.
2. Data Readiness: Many agencies neglect data governance and metadata management, which needs to be implemented at the enterprise level. Identify key marketing use cases for AI and designate a marketing data champion to oversee and maintain data quality and security. This may require many agencies to recruit essential skills in data and technology to manage this process. Data management, a critical component for effective AI utilisation, demands robust systems and expertise, adding another layer of expense.
3. Change Management: Successful implementation requires change management across every element of the current workflow within the agency, the clients and the agency suppliers. This includes training and upskilling employees to prepare them for AI integration.
4. Financial Costs: While the holding companies have announced significant investments in this area, the financial cost of implementing AI for an independent agency or a local office of a network agency can vary widely depending on the scale and scope of the project. Costs will include software and hardware investments, data management, training, ongoing maintenance, improvements and upgrades. The budget for these expenses must be considered against potential cost savings from increased productivity and efficiency. We explore these costs further below.
5. Timeline: The timeline for implementing AI solutions can also vary. Initial pilot projects may take a few months, while full-scale implementation could take a year or two or more. It’s crucial to have a phased approach, starting with pilot projects and gradually expanding to widespread implementation. We look at these in more detail below.

The timeline and cost of fully automating an advertising agency using AI can vary significantly based on several factors, including the agency’s size, the complexity of the tasks, the technology stack chosen, and the level of customisation required.

Financial Investment

While the financial cost of designing and implementing an AI-enabled transformation of an agency will vary wildly depending on the size of the agency, the range of services and processes impacted, the level of complexity of integrating into existing client business and more, for this exercise, we have developed costs for a medium size independent creative agency offering strategic, creative and production across digital advertising.

These cost estimates are broken down into the following four categories:

1. Technology Costs: This is based on purchasing existing software and developing custom solutions. It includes designing, training and testing the AI platform with costs ranging from $250,000 to several million dollars.
2. Consulting Fees: There are already a plethora of AI and Tech Consultants hawking their services in marketing and advertising. Hiring one or more of these experts to guide the automation process may cost $100,000 to $200,000 per year.
3. Training and Talent Costs: The human cost is often overlooked when discussing technology-driven transformation. Investing in training your existing employees on a new system or recruiting subject matter experts could cost between $50,000 and $100,000 per year, excluding additional salary costs.
4. Ongoing Maintenance and Enhancements: Technology is not a set-and-forget investment. Keeping up with ongoing maintenance, updates, and upgrades could cost 10% to 20% of the initial cost annually.

Depending on the scale and scope of automation, the total investment might range from $500,000 to several million dollars per agency or office. Since agencies are at best reporting a 15% EBIT on income, the agency would need to sacrifice almost three and a half million in revenue to fund even the entry-level investment, and it just goes up from there.

Investment in time.

Many estimates put the entire AI transformation process at anywhere from one to two years for a typical independent agency. While some argue that they are further developed in less time, it is unlikely that they have embraced a total business transformation and instead have looked at piloting a smaller, more defined implementation of generative AI rather than a total business transformation.

While it may be argued that the race to AI integration started in 2022, many agencies have been slow to start. This is because of a number of key issues.

1. Decision paralysis – caused by a lack of knowledge or understanding of the opportunities.
2. Risk mitigation – hoping that others will lead the way and reduce the inherent risks of mapping their own transformation.
3. False starts – with successive small projects that have led to poor outcomes or abandoned due to loss of momentum and commitment.

Considering that the market is allowing the following timelines for each of the transformation steps identified above across the one-to-two-year transformation process:

1. Initial Assessment (1-3 months): Evaluate current processes and identify areas for automation.
2. Development and Integration (6-12 months): Build or implement AI tools, integrate them into existing systems, and customise them to fit the agency’s needs.
3. Testing and Iteration (3-6 months): Test the systems, gather feedback, and make necessary adjustments.
4. Training and Transition (2-4 months): Recruit new staff capabilities and train existing staff on new systems and transition workflows.

Impact on Agency Fees

With the agencies facing a two-year, multimillion-dollar investment to transform their business model to leverage the advantages of AI and automation technology, it is worthwhile to consider how this will be funded, not just for the upfront transformation but the ongoing maintenance and upgrade costs.

The overwhelmingly standard agency fee model continues to be a resource-based cost model based on hourly rates, billable hours, overhead, and profit margins. How will an agency build the cost of the AI transformation process into the current fee model, including the external and internal cost of resources? A simplistic approach would be to increase the overhead of their resource fees. However, these fees and rates have been under significant competition and negotiation processes for a number of years, and many agencies would struggle to have their existing clients agree to pay more for the same services.

Instead, after years of agencies clinging to traditional fee models, we see more agencies embrace new fee approaches. These evolving pricing models highlight a crucial shift for agencies moving away from billing for their time and towards pricing based on the value and results delivered, often enabled by AI.

Efficiency-Based Enhanced Profitability Pricing: Agencies that use AI to produce more output with less time and cost may maintain fixed-fee pricing without reducing revenue. While the agency benefits from increased profitability due to AI-driven efficiencies, the client does not necessarily see a direct cost reduction.

Tech Fees Surcharge: Agencies are introducing a surcharge, typically 1-5% of agency fees, to account for their technology investments. This direct pass-through of costs ensures that clients contribute to the agency’s AI infrastructure.

Increased Traditional Pricing Models: With increased expertise and quality of agency staff, particularly in tech and data, AI can increase existing labour-based or output-based models. For example, agencies may charge for additional, higher-level staff required to manage AI tools or to generate revenue from new, AI-enhanced creative deliverables. This won’t necessarily lead to lower costs and could potentially increase costs.

Other Non-Traditional Pricing Models: Agencies that develop specific and customised AI applications and platforms, such as LLMs or workflow bots, are exploring subscription or licensing models, treating their AI as intellectual property.

These approaches indicate that while AI can drive efficiencies, agencies’ primary focus is leveraging it to enhance their offerings and improve their financial performance, which doesn’t automatically equate to lower client costs.

The bottom line for agencies and marketers

While AI and automation hold immense potential to transform advertising agencies, the narrative of automatic cost reduction overlooks the significant underlying investments in technology, talent, and training required for successful implementation.

To recoup these costs and capitalise on AI’s value, agencies must adapt their fee models through surcharges, efficiency-based pricing, and the monetisation of AI-driven intellectual property.

Consequently, the application of AI in advertising agencies is more likely to result in a restructuring of fee models and a shift towards value-based pricing rather than a straightforward decrease in advertisers’ costs, as agencies often propose in the highly competitive pitching world.

To understand how technology is impacting your agency fees find out more on our approach to agency commercial evaluations and fee models or contact us for a confidential discussion.