How dentsu LATAM’s AI Playground Will Help Add Value for Clients

This summer, dentsu launched its ‘AI Playground’ across Latin America, a set of customised solutions that combines emerging technologies with audience data to bring more efficiency and agility to companies’ day-to-day. Built on dentsu’s Azure OpenAI architecture, the AI Playground comprises more than 100 AI technologies and can generate more than 400 optimised processes.

Designed to democratise emerging technologies and improve the sustainability of the business, the toolset integrates martech, media and creative solutions, managed via the proprietary Nawel® platform and Navegg, a dentsu company specialising in data intelligence. The Playground provides a framework for companies that adapts to their structures, needs and available technologies, thus allowing them to identify use cases for - and create strategies around - various new AI tools and other tech.

To discover more about the AI Playground and dentsu LATAM’s approach to the future of AI in marketing, LBB’s Ben Conway spoke with the agency’s CEO CXM and chief solution officer, Juan Camilo Suarez.

LBB> So, tell us about this ‘AI Playground’! It includes over 100 AI technologies - what are some of the most vital ones, and what solutions will it provide for your agencies and their clients?

Juan> We are integrating three types of technologies to accomplish different microtasks from 14 capacities we have at dentsu.

External capabilities: The Playground can be connected to any technology that has API available. We have some technologies by default like OpenAI, Microsoft, Google, Semrush and RapidAPI among others that can be chosen to adapt to the request.

Data integrations: We have Playground connected with analytics platforms and data cloud services like GA4, Bigquery and MSSQL. All technologies that pass through our proprietary prompting technology and API call generator can bring information in an easy way to the user.

Internal technologies: We have developed internal technologies to consolidate audience (Navegg), first-party and third-party data - and integrate it with marketing cloud platforms and DSPs.

The benefits for the clients are:

• Taxonomy and templates to generate consistency.
• Integration with Workfront to facilitate collaboration.
• Guidelines and manuals to streamline training.
• Democratise the use of technology to enhance the development of talent and generate efficiency.
• Connect company assets and tools with a powerful AI engine to orchestrate inputs and generate insights.
• Audience intelligence engine integration (Navegg).

LBB> When did this project begin? How long has this area been a serious target for investment for you?

Juan> The planning and execution of the stage has been a 12-month journey where we have over 30 people in the region working toward consolidating processes, microtasks and tools to make this robust framework for our operations. 

[Above: Juan Camilo Suarez, CEO dentsu CXM LATAM]

LBB> Why is now the right time to develop and deploy this toolset?

Juan> The advances in AI, the challenges of growing organisations and efficiency focus, and the complexity of tool access have been our drivers to move this project forward.

LBB> How does the AI Playground build on the existing Azure OpenAI architecture? What new opportunities does it open?

Juan> Our Playground connects with different technologies. In this ecosystem, we are deploying process repositories in Sharepoint, data models in Azure and task-oriented API integrations with OpenAI.

LBB> The tool is managed alongside dentsu’s data intelligence company Navegg - how important has this relationship been, and how has it helped to incorporate audience data into this new toolset?

Juan> Audience data is a key in the future of the Playground. For many microtasks in our daily operations, we need to consider how audiences and individuals get to the right target and create concepts with the audience in mind. We are working on integrations for API connections to facilitate the process integrations and include audience analytics in each microtask that makes sense.

LBB> The Playground helps to create personalised frameworks for individual organisations, simplifying the use of emerging technologies. Why is this so necessary? What are the biggest challenges in the AI/emerging tech space that make it so complex and unwieldy?

Juan> Think about the different challenges organisations are going through right now. Their employees are getting free access to any AI technology without rules and restrictions; specialists have training for each tool they have to interact with; processes across countries and regions are different; challenges to generate efficiencies due to a lack of consistency; and a lack of a clear path to integrate AI into microtasks.

All these reasons are why we are pushing very hard into AI-responsible involvement in microtasks, to support our professionals in manual operations.

LBB> What will be some of the first concrete examples of the AI Playground at work? What types of projects or brands can we expect to see making use of this toolkit?

Juan> We are starting to operate many microtasks through the Playground across multiple brands. This is a matter of starting from the bottom-up, where the Playground is organising our teams, standardising our deliverables and generating efficiencies in tool management and training. Now, for those brands that we are supporting in their microtasks with AI, we have an additional step which is to work together on the scope [of the AI usage] and the rules, regulations and restrictions for its use.

LBB> What excites you most about the future of AI tech in advertising - and equally, what fears (if any) does it bring? How can the industry overcome these concerns?

Juan> In advertising, like in any other industry, we have many manual tasks that keep our people from having time to think and add value to the client's strategy. Having access to tools that generate efficiencies and facilitate daily operations for our people is the key part of this effort.

AI should be a tool to support and not to replace. Any deliverable made by AI has to be curated by a person and brought as a source for insights and strategic discussions.