How agencies are using AI to innovate for clients and work faster

Agencies are getting savvier at using artificial intelligence as data comes into greater focus as marketing and information technology are becoming more entwined.

Some studies back this up. For example, 81 percent of marketing and IT leaders surveyed by data platform Lytics said their two departments will become more involved in marketing over the next five years, while 66% of marketers plan to integrate AI into their marketing stack.

Experts say the use of AI is increasing as ad targeting with first-party data has become even in more demand. AI can help automate some of those processes, from creative to segmentation and potentially help agencies optimize their ad spend by cutting down on tedious tasks.

“I believe the true power of these technologies and tools are not to replace creatives, but rather make us more efficient,” said Ben Williams, TBWA Worldwide global chief creative experience officer. “There will always be a need for and value in creative direction, human curation, human refinement of an idea and decision making in terms of what is right for the brand we’re working with.”

But organizations are still experimenting with AI as the space grows, using it in a range of ways from moderating content and brand safety to implementing it in their communications strategy and creative products.

Generating content and creativity

One of the key ways agencies are testing AI is in the creative process, using data points to drive content. With these insights, AI generators are able to create relevant content in minutes or even seconds. But this doesn’t replace the work of content creators, acknowledged Nadia Gonzalez, CMO of AI marketing firm Scibids.

“It instead gives [creative teams] a rich tool to use, ensuring more diverse content is seen at just the right time and place,” Gonzalez said. “With companies of all sizes using data science and AI-driven analytics, we’ll also see greater dynamism in things like pricing, personalization and recommendations.”

At creative consultancy Codelab303, founder Anthony Chavez said understanding their clients’ desires is the first step in knowing how to use AI for the organization. Besides using automation for time-intensive tasks, they are utilizing AI tools for producing writing, music and other visuals. The firm has worked with brands such as Ulta Beauty and Carvana.

“There are activities that machines can do much better than humans,” Chavez said. “Whether the objective is providing real-time pricing, streamlining sales bookings or automating content marketing, there is likely an AI candidate within every organization, and the question is more about revealing the role than the resource.”

Similarly, TBWA uses AI content generators to produce assets and products for some of their clients, including Nissan and Corona. Williams described AI as part of a current “creative revolution” that can inspire teams and clients to think differently.

“We’re in a creatively exciting time that reminds me of the ’90s when people, brands and the world were exploring and experimenting how new technologies, tools and services can enhance and amplify their brand or mission,” Williams added.

Over time, these automated creative tasks can make the process more efficient, he said, though he also acknowledged the need for the human touch in this process specifically taking those assets and applying the creative direction of those ideas for their clients.

Training algorithms for brand safety

Many organizations have been using this technology to moderate content or identify harmful conversations on a platform. For example, Tyson and Mindshare recently partnered with intelligence startup socialcontext.ai to create a tool called Impact Index. Using this index, the organization is measuring the social impact of its editorial content in the Black community.

The tool will mark content as positive, negative, neutral or toxic based on an algorithm made with tens of thousands of human annotations. Over time, the partnership wants to refine its editorial strategy and help shape Tyson’s media investments to be more diverse and impactful.

“Through our partnership, we’ve been able to uncover a variety of meaningful insights that are impacting our buying strategy moving forward,” said Courtney Ballantini, vp of marketing communications and design at Tyson.

“By considering the entire article and related metadata, we’re able to get down to the rhetorical level,” added Chris Vargo, CEO of socialcontext.ai. “By training algorithms to detect desired social outcomes, instead of just marketing cohorts and taxonomies, the Impact Index is more inclusive by design. Negative content is factual in nature. For instance, coverage of Black-on-Black crime is not toxic per say, but if it’s over-represented in news, it has a negative outcome on [the] perception of the Black community in society.”

Gonzalez of Scibids said AI content moderation is becoming necessary as the amount of digital information grows rapidly. This makes it difficult for even social giants like Facebook and Twitter to police user content and adapt their policies.

“With the world becoming more digital, it will be impossible for human moderators to oversee and anticipate everything,” Gonzalez said. “As with the AI we build for programmatic media buying, content generation and the data behind it will outpace and outperform human capacity.”

A broader strategy, from customers to C-suite

As machine learning improves, some agencies are also using these mechanisms as a way to help clients build a broader communications strategy at the leadership level. A new offering at Boathouse Group is focusing on behavior and engagement, aiming to help clients drive engagement and grow relationships in media, both internally and externally with customers.

Peter Prodromou, founder and president of Boathouse Palo Alto, sees AI as being useful to identify actionable tasks and solutions for executives. While working with Kaiser Permanente, for example, he said the focus was using AI tools to reduce the amount of data the CEO was getting so that he could make better decisions faster.

“Most of what is out there right now are either automated tools that do a fabulous job of telling you how to engage, or how to create volume and social and digital,” Prodromou told Digiday. “As opposed to thinking through what somebody who sits in my chair has to do to then put it in front of a client and make it work well and drive results.”

Additionally, Prodromou believes these AI-driven tools may spur greater changes in training and management at organizations. If a client has data on complaints about the company, for instance, the agency can recommend strategies on paid media and other customer service retraining for employees. They can use the data to help the client with solutions on the inside that impact perceptions on the outside.

“One of the things that we think has been undervalued is the opinion of employees, so the tools that we’re looking at focus on employee engagement as much as they do with external constituent groups,” Prodromou said. “There’s such a rich set of places that agencies can go based on the data outputs, and I think it’s going to create a lot of ‘aha’ moments between agencies and the executive suite.”

Sign up to get the day’s top stories at 6am eastern.