When we publicly launched Gigi in June of last year, Gigi emerged as the only company to harmonize two foundational Amazon ad tech products—its DSP and Amazon Marketing Cloud—to help marketers buy and measure Streaming TV Ads. Since then, we’ve managed tens of millions of Streaming TV ad spend for customers.

Every customer and every prospect over the past year, though, has shared the same request: we need you to do more than just Streaming TV. They love how we’ve removed the bifurcation between the Amazon DSP and Amazon Marketing Cloud into a single workflow. They recognize the value of the Gigi 1P Platform as a private and secure way to collaborate first party data with Amazon Ads. And they trust us to use the full scope of Amazon’s ad tech to enable them to grow their business. They want us to do more, and it's time for us to listen.

We recognized it was essential to support our customers with full funnel strategies and all ad formats supported by the Amazon DSP. Still, we asked, “How could we do so in a manner that offered a demonstrably better customer experience than the DSP itself?” At the same time, we’ve found ourselves amid modern history's most significant technological transformations: LLMs and generative AI. 

The AI Media Manager 

If one looks broadly at the most successful AI application layer companies, they transform the paradigm of interacting with B2B software. One no longer “buys” a product, but instead, they are “hiring” an AI agent to autonomously execute tasks and perform a job’s function. 

The first wave of AI agents achieved meteoric success by tackling horizontal categories of work, such as Cursor for software development, 11x for sales, and Decagon for support. The next wave of AI agents will be more verticalized, tackling specific jobs in the workforce and share the following characteristics: 

• Jobs that perform many repetitive and manual tasks

• Complement an LLM's general intelligence with deep vertical-specific expertise, and

• Rely on a set of vertical-specific tools to perform autonomous tasks 

All of this describes the media manager at an agency or brand. Media managers spend most of their days performing a set of prescribed tasks based on internal best practices, including optimizing campaign performance against a predefined set of rules and following a script to analyze results and present them to clients or internal stakeholders. These tasks are often painstakingly repetitive and manual, preventing them from performing the strategic work that’s accretive to both them individually in their careers and the growth of their company.  

Over the past year, we've been working with media managers at retail media agencies to understand how they spend their time. We then deconstructed all the tasks that consume the most time and are the least strategic, and created a set of agentic workflows and automations so media managers can focus on developing expertise in the service of their clients. 

We are building the AI media manager. Our first product is an AI media manager for the Amazon DSP, and it is purposely built for agencies.

Why Agencies

This recent AI wave has created yet another chapter in the pronouncements of the “death of the agency”.  Sam Altman believes 95% of what agencies do today will be automated with AI. Mark Zuckerberg wants to “kill the agency” by allowing Meta’s customers to push a button and receive an outcome. We see a different future. We believe agentic AI will present a new wave of opportunities for media agencies, and we want Gigi to be the catalyst that enables agencies to seize the agentic AI opportunity before them. 

Our team has worked with and worked for the leading agencies in retail media. We’ve learned the secret sauce of the best agencies is not data and technology, but instead, scaled processes and standard operating procedures institutionalized across their employee base to achieve a consistent standard of client service. These operational investments of agencies can be perfectly applied to train an AI agent in the same manner that an agency leader would train a new employee. Except that the AI agent has all the advantages of software: always available, error-free, and scalable up and down with demand. 

Perhaps most important, though, is that this process enables an agency to create a bespoke software product specifically tailored to their agency and its unique approach. In our experience, agencies will often ask software providers, “How can my off-the-shelf version of your software be different than any other agency that uses this product?” In the world of legacy SaaS, those differences were minimal at best and non-existent in most cases. With agentic AI, an agency can now “hire” software and train it in a manner that is unique to them. If implemented correctly, a virtuous feedback loop can be created where agencies spend their time developing expertise and thought leadership, codifying that expertise to their AI agent, have the AI agent execute the expertise across all clients, and work with their AI agent to uncover new insights and expertise.  

We also think agentic AI will transform the operating model of agencies. Currently, an agency needs to linearly scale headcount as customers and revenue grow. A fixed ratio of FTE to clients is formed.  But, as agentic AI performs many of the time-consuming and repetitive tasks of a team member, an agency should exponentially increase its customer base without the need to grow its existing team.  Agencies that adopt agentic AI at scale will, in turn, benefit from margin expansion and, with an improved operating margin, will transform how agencies are valued.  

Gigi is here to unleash the power of agentic AI to agencies.  If we succeed in this mission, we can elevate the role of media managers, improve performance and client satisfaction, and strengthen the agency's financial position.  

What’s Next

We began onboarding early agency partners to a closed beta of the new Gigi at the start of May and will continue to onboard a select group of agency partners throughout the summer. While we are admittedly still early, it has been remarkable to see the time savings our customers have achieved, the improved advertising performance, and the richness of the product experiences we can create as an AI-native company.  

If you’re an agency that manages endemic advertisers on the Amazon DSP and is interested in being part of this closed beta, please reach out (adam@gigico.ai). 

We are incredibly excited to introduce you to the new Gigi. We have so much to do and are only getting started.