The 2024 State of Marketing AI Report reveals that only 36% of marketers are infusing AI into their daily workflows and only 19% have AI roadmaps. Marketers also shared concern about lack of AI training, tools and support.
As agencies and brands grapple with how to embark on their AI journeys, we approached a trusted client to come together to determine how AI could impact our results, partnership and culture. We needed a framework, the right team, and bravery to shake up the typical agency-client dynamic. With those ingredients, we navigated training, testing and learning, and ultimately transforming how we work while also executing one of the brand’s most important and successful campaigns of the year – a special delivery for expecting moms (which sold out in seconds and created a lifetime of loyalty).
There were initial fears. But we quickly came to understand that when agency and client partners both value ambition, innovation, and inspiration, we can go from trepidation to collaboration and fear to fact.
By integrating AI into every facet of our campaign, we learned that AI can deeply and positively impact our results, client relationships, and culture – if people remain central.
While we started with one specific campaign, our agency and client have a long-term commitment to becoming fully AI-integrated in our work together. This is the story of our beginning and initial impact.
Strategy
We wanted to go deep on AI with one client and one campaign. We applied nearly 20 AI opportunities within eight (8) weeks, documenting the human experience, key learnings on the approach, tools, and techniques, and future recommendations. We behaved like scientists, adding an A/B test component to effectively test and compare the AI-driven method (“A”), experience and results with the traditional (“B”).
We also aimed to test new ways of agency-client collaboration, team leadership, training and communication:
Traditionally, agencies aim to show up to clients as polished experts with all the answers. For this project, we prioritized discovery, partnership, and progress over perfection.
We approached this as a shared learning journey – sharing the wins and the misses, all with radical transparency. We had regular stand-ups with our account team and the clients together, we shared video diaries, and we didn't wait until we had all the answers.
We designed the right team to collaborate, experiment and solve together. The team included clients, core agency team members with a deep understanding of the business, and AI leaders across specialty areas: earned, digital, creative and production, and data and analytics. We empowered team members to be pilots rather than passengers. Driving transformation is as much about psychology as it is technology – we encouraged a growth mindset, learning and curiosity.
We examined each task across the communications spectrum to uncover opportunities for AI optimization and innovation. We co-developed our roadmap, aligning on guiding principles, use cases, approaches and tools. We evaluated risks alongside key stakeholders, including legal.
At the start, only 37.5% of our team had AI experience. However, after implementing AI as follows, 100% reported a solid foundation (+62.5%).
Creative and production
Video correction: To improve the quality, reduce background noise and increase resolution on UGC.
Audio: To create a voiceover when a traditional recording wasn’t scoped.
Content testing: To test video content with target consumers to understand sub-conscious (facial expressions) reactions.
Custom assets: To create custom pitch assets for top media. We used a mix of tools to superimpose a key campaign visual in front of target media buildings.
Influencer
Identification: To search, compare and identify influencer selects based on key criteria.
Vetting and background checks: To analyze potential issues across platforms, going back 10 years to see what influencers posted and what/how they responded to others’ posts.
Writing influencer briefs: To automate key briefing materials.
Impact: To study the impact of our influencer content on audiences.
Earned
Writing and editing: To create first drafts of materials for editing and conversely, to edit already-developed materials for grammar, brevity and more.
Media targeting: To uncover additional journalists to engage with and those to avoid.
Media personalization: To provide factoids allowing us to personalize outreach.
Basic research: To automate basic research tasks, such as developing media briefing sheets to prepare spokespeople for interviews.
Data and Analytics
Research analysis: To analyze the detailed results of a consumer survey and pull key themes and findings.
Earned landscape analysis: To provide summaries and analysis of the earned landscape to identify white spaces and opportunities.
Automated sentiment analysis: To automate sentiment analysis (positive, negative, neutral) on earned media coverage.
Consumer affinity: To determine emotional connection to brand deeper than basic sentiment coding.
Tracking and reporting: To automate basic results tracking and reporting.
Results
Beyond efficiency, AI creates efficacy.
Content Testing: We can pretest consumers’ sub-conscious reactions to creative content before publishing. Emotiv Content Testing uses AI to study facial responses to content, providing insights that are proven to predict content’s success.
Influencer Vetting: It’s more important than ever to ensure influencers are brand-safe. Our AI-driven process, FitCheck, reduces vetting time by 50% and is also more thorough than the manual approach. We can analyze potential issues across platforms, going back 10 years to see what influencers posted, responses to others’ posts, and red flags about their followers.
New Creative Possibilities: We used AI to create custom pitch assets for top. Now AI can help with personalization without much extra time or budget.
Media Vetting: We used AI to determine white space for media. Our campaign was live during a major global athletic event, and we identified when local athletes would be competing to avoid those pitch windows with media.
There were misses too, but with clear solves. Some AI tools didn't work well on the first try. The right prompts are critical to getting the best results. We also learned that we need to maintain a growth mindset and continue to experiment with multiple AI tools because each one has different strengths. Just like people!
The experience fostered a culture of innovation and learning, according to 94% of our team. 100% of our team said they will use AI moving forward, and 94% are now at least moderately confident in their AI capabilities.
