“AI agents,” “multi-agent systems,” and “service-as-software” – is this the future of advertising?
I’ve spent the last few weeks on this topic. Let’s unpack:
🤖 AI agents: Autonomous digital entities performing specialized tasks, using LLMs to simulate human behavior and interactions. Instead of simply aiding human workers, these agents draw insights from both structured and unstructured data to make decisions and complete tasks with minimal human oversight.
⚙️ Multi-agent systems: Networks of AI agents working together, effectively serving as a digital workforce. Complex problems are broken down into discrete subtasks handled by specialized AI agents (vs. CRM or ERP systems that help digitize business processes but rely on structured data and manual inputs).
🛠️ Service-as-Software: From traditional software-as-a-service (SaaS) to service-as-software, where software is no longer just a tool but becomes the worker itself. From tools to end-to-end services, this paradigm shift enables outcome-based pricing models and lowers opex by automating/replacing workflows.
Here are two potential impacts we can expect on the advertising industry:
1️⃣ Advertising to/with agents: Tools like ChatGPT have changed the way consumers explore and learn about problems.. and brands. When someone is researching a new TV, a hotel option, or what anti-aging cream to buy, Brand Managers should be able to answer questions like, « What is this LLM saying about my brand? How do I educate it? How do I compare with competition? »
2️⃣ Workflow automation: Agents could replace and streamline multiple workflows, reducing human workloads and enhancing performance. Think of areas like creative automation (production, serving, personalization), media buying (planning, optimization, payments), and overall advertising analytics (from data collection to insights).
Market Opportunity: The transformation to service-as-software may represent a $4.6 trillion market opportunity over the next five years (Foundation Capital). For companies, this model is highly scalable and agile, empowering even small teams to achieve results that were previously labor-intensive and costly.
In the first comment are two helpful resources: an article from Foundation Capital (credits for the image) and a discussion between the CEOs of NVIDIA and Salesforce
As always, keen to hear thoughts and/or experiences in this space.