Everyday: Reinventing the conversation between brands and their clients with Google Cloud's Generative AI

About Everyday

Everyday is an agency that specializes in advising brands on the quality of their conversations with their audiences throughout the year. The agency offers content with high added value by making it more direct, more meaningful, more creative, and more emotive. Agile and innovative, the agency has adopted an integrated approach to strategy and creation, while relying on semantic data, influence, and artificial intelligence tools.

Everyday is part of Publicis Groupe in France, one of the world leaders in public relations with 5,000 employees.

Industries: Media & Entertainment

Location: France

Products: Vertex AI, Cloud Storage

Tell us your challenge. We're here to help.

About Publicis Sapient

Publicis Sapient is a digital business transformation company. We partner with global organizations to help them create and sustain competitive advantage in a world that is increasingly digital. We operate through our expert SPEED capabilities: Strategy, Product, Experience, Engineering and Data & AI, which combined with our culture of curiosity and deep industry knowledge, enables us to deliver meaningful impact to our clients' businesses through reimagining the products and experiences their clients truly value. Our agile, data-driven approach equips our clients' businesses for change, making digital the core of how they think and what they do. Publicis Sapient is the digital business transformation hub of Publicis Groupe with more than 20,000 people and over 50 offices worldwide. For more information, visit publicissapient.com.

A specialist in conversations between brands and their audiences, Everyday adopted Google Cloud's generative AI so it could adapt to an industry undergoing transformation, as well as innovate by offering its clients more relevant and more unique content that reinforces the quality of the conversations between a brand and its audience.

Google Cloud results

Optimizing low-added-value tasks, such as data collection
Strengthening creativity—teams have more time to focus on the creative aspects of public relations
Improved content quality and impact thanks to more comprehensive knowledge about the brand
Stronger and more relevant connections between brands and their audiences thanks to more personalized content

Personalized AI for each client

Part of the Publicis Groupe in France, Everyday is known as an innovative and agile digital social media agency. It offers high-added-value content to improve the quality of conversations between a brand and its audience throughout the year. The agency is convinced brands can deepen their relationships with consumers by developing their daily presence on social media platforms and online, making them more meaningful, more creative, and more emotive. "We've always been passionate about designing proprietary and definitively unique content strategies for our clients. We get the chance to work on a fascinating subject—connecting brands with their audiences every day of the year. And to plan for the future of the agency, we continue to challenge our talents so they can work with us to invent a new way of expressing ourselves," explains Cécile Leprince, Managing Director at Everyday.

"We've always been passionate about designing proprietary and definitively unique content strategies for our clients. And to plan for the future of the agency, we continue to challenge our talent so they can work with us to invent a new way of expressing ourselves."

—Cécile Leprince, Managing Director at Everyday

In order to achieve its ambitions, the agency has increased its creative talent pool, strategic intelligence, and data processing capacity. Its latest recruit was technological and part of this quest for innovation so its clients can have the biggest positive impact on social media platforms and online, with Everyday deciding to rely on Google Cloud's generative AI.

Industrializing the use of AI

"The objective is not to replace our employees by offering our clients AI-generated content," clarifies Cécile Leprince, Managing Director at Everyday. "We're convinced generative AI will have an impact on our businesses and the way we work, as well more generally on the value proposition of agencies in our line of business. More specifically, it will relieve our employees of tasks with no added value for our clients, such as gathering knowledge before being able to write text highlighting what makes a brand stand out. In that sense, AI is an amazing assistant that saves time and allows humans to fully express their creativity."

Before launching, Everyday took the time to reflect. The agency not only wanted to explore the potential of AI and come up with use cases that could be integrated into its business processes, but also to think about its approach from an industrialization perspective. "The implementation of AI must follow a strict process. You must classify your data, develop the AI, train it, and so on, and it's the quantity and quality of data that will determine performance," explains René Thomas-Nelson, Senior Director of client Experience & Innovation at Publicis Sapient, a digital business transformation company that has supported Everyday with its project. "Based on this principle, everything we do must be able to benefit several clients."

Personalized, ethical, and secure AI

Nevertheless, to be truly effective and offer relevant content, generative AI must be trained with appropriate data. In other words, if the use case is generic and the generative engine is shared, the training itself must be specific to each of the agency's clients. "Each client has their own AI, trained with data they entrust to us. When choosing tools and infrastructure, it was therefore essential to opt for a partner capable of ensuring data security and keeping our clients' databases isolated. Google Cloud is highly advanced when it comes to generative AI with pre-trained models that have saved us considerable amounts of time, and it's also delivered on these guarantees," explains Ludovic Michel, Creative Director at Everyday.

"Google Cloud stands out for its especially responsible policy. While making AI more accessible to all companies, Google Cloud sets clear boundaries, such as refusing to produce faces with its models. It's a responsible approach implying an excellent mastery of these innovative technologies, which are capable of bad as well as good."

—René Thomas-Nelson, Senior Director of client Experience & Innovation at Publicis Sapient

Another fundamental point, whatever the use case, was that the AI had to offer impeccable ethics. "We must of course avoid bias, which is our responsibility, but also be able to rely on dependable technology. Again, Google Cloud stands out for its especially responsible policy. While making AI more accessible to all companies, Google Cloud sets clear boundaries, such as the refusal to produce faces with its models. It's a responsible approach that is perfectly aligned with Publicis Sapient's own positioning, implying an excellent mastery of these innovative technologies, which are capable of bad as well as good," says René Thomas-Nelson.

AI innovation at the heart of a three-way collaboration

With a strong technological base that meets its technical, ethical, and business needs, Ludovic Michel invented three use cases for generative AI.

The first model assists writers with creating text for websites, social media platforms, and even newsletters. The objective is to examine and maximize brand knowledge by asking AI to consolidate relevant content to write text.

More oriented toward the creation of visual content, the second model generates illustrations or modifies existing illustrations.

Finally, the third use case plans to develop AI capable of verifying that the created content is in line with the brand's DNA and complies with the client's corporate identity, ethics, etc.

"In fact, these three use cases were the result of a close collaboration between Everyday, Publicis Sapient, and Google Cloud," explains Ludovic Michel. "We brought the business expertise, Publicis Sapient added its mastery of Google Cloud infrastructures, and Google Cloud not only shared its technical knowledge but also its experience, specifically helping us to align use cases with business objectives. Typically, from the very beginning, Google Cloud alerted us to the replicability of our use cases, so as to make our work and our investments profitable by developing models that can be used by several clients, while keeping everyone's informational assets completely isolated."

Simplified and secure implementation thanks to Google Cloud

Launched in July 2023, the prototype development of the first use case was carried out by Google Cloud teams in less than two weeks. Combining LangChain's open source framework features with those of Vertex AI, it unites the linguistic capabilities of Google Cloud's PaLM 2 large language model (LLM) with client-specific AI analytics. Its implementation was simplified by the wealth of tools and managed services provided by the Google Cloud, which offers all the essential building blocks to carry out such a project from pre-integrated generative AI models (such as PaLM) to the AI workload management and deployment platform, Vertex AI.

Last September, the prototype entered the pre-production phase. This is when Everyday launched its first tests. Data has been injected into a Cloud Storage bucket that guarantees both security (encryption) and isolation (data in a container). It's used to personalize the model and factualize the generated texts. With the technological appropriation phase fully underway, Publicis Sapient has already started to take over the operation of the model thanks to the transfer of skills carried out by the Google Cloud teams.

"It's an exciting experience and we're having a wonderful time. Generative AI will enable us to seek more accuracy, more relevance, more specifications, and ultimately more intelligence and creative intelligence."

—Ludovic Michel, Creative Director at Everyday

At the same time, Everyday has begun training its employees in the practice of "prompting," which is the principle of giving the "right" instructions to the generative AI to obtain the most relevant responses possible. "It's an exciting experience and we're having a wonderful time," beams Ludovic Michel. "Generative AI will enable us to seek more accuracy, more relevance, more specifications, and ultimately more intelligence and creative intelligence."

In the meantime, the enthusiasm of Everyday's creative director for generative AI and this first successful use case has sparked significant interest within Publicis Groupe, as Cécile Leprince points out: "The relationship between a brand and its audience is very complicated and, with changing consumer expectations, brands can no longer tell their stories in the same way. Generative AI will allow us to be more responsive, generate more relevant content, and ultimately promote better symbiosis between our clients and their own clients. It's part of a broader transformation process and plays an essential role in our repositioning around the conversation between a brand and its clients. There's no doubt our experience will benefit the entire Group."

Tell us your challenge. We're here to help.

About Everyday

Industries: Media & Entertainment

Location: France

About Publicis Sapient

Vertex AI Cloud Storage