AI was a popular topic at the Martech Conference in Spring 2025 in late March. We did not search for transcripts for all sessions and panel discussions, but AI could have been mentioned in all sessions.
The three meeting sessions were completely dedicated to AI. Two of these were coffee talks where attendees could chat with the speakers.
The three sessions are as follows:
A real-world marketer who shares real-life AI success stories with Google’s global head Kendall Davis. Channan Sawhney, customer leader at global Amazon; Sarah Weiss, Vice President of Marketing at Qvest. I use Lisa Peyton, an immersive and strategic communications instructor at the University of Oregon, to use a generation AI tool for content creation. Trustinsights.ai and Constantine Von Hoffman, embrace agent AI with Christopher Penn, managing editor at Martech.
Combining the speakers and participants knowledge from these sessions led us to five conclusions on AI and marketing.
1. Small and start scale for actual impact
Instead of trying to overhaul your marketing right away, choose a specific issue (create targeted ad variations or automate repetitive tasks) – Test your AI-driven solution.
Why it matters: This allows teams to learn quickly, demonstrate tangible ROI, and systematically expand the most successful use cases into more important initiatives.
Example: Use AI to analyze the creatives of a single high priority campaign, see what resonates most in key segments, and then optimize from there.
2. Combining human expertise with machine efficiency
AI is excellent at processing huge amounts of data, highlighting trends, and automating labor-intensive tasks. However, “looping people” is essential. Marketers should continue to provide an understanding of strategy, creativity and audience psychology.
Why it matters: Marketers worry that AI will replace them, but AI works best when human intuition, brand knowledge and empathy lead to the machine’s output.
Example: Generate multiple versions of creative copies in AI, but rely on the marketing team to improve tone, voice and brand reliability.
3. Measure efficiency and performance lift
When assessing the effectiveness of AI, don’t look at just performance metrics (such as conversions and revenue). Track efficiency metrics such as saving time, improving costs per acquisition, or reducing creative production overhead.
Important reason: Stakeholders need business outcomes and operational ROI. AI can provide teams from manual work for free and free. This channels saved time towards higher value strategies and innovation.
Example: Compare the results of an AI-assisted campaign with a “regular business” campaign, quantifying click-through or revenue uplifts, saving production time and media spending.
4. We are leaning towards ultra-personalization and real-time adjustments
Use AI to generate or tailor messaging, visuals, and offers based on real-time signals such as browsing behavior, devices used, or time of day to make each consumer’s touchpoint feel uniquely relevant.
Why is it important: Consumers are increasingly expecting a coordinated experience. AI can combine many data signals to provide more timely and personalized offers than standard segmentation.
Example: If a consumer has repeatedly expressed an interest in a skincare product type, it uses an immediate promo code during peak purchase times to provide dynamic advertising featuring its exact product line.
5. Build partnerships and data strategies to maintain agile
AI-driven marketing works best when combining data (such as customers and past campaigns) with partner data (from retailers and digital platforms). However, success depends on having the right processes, including clean data, robust APIs, and organizational buy-in.
Why it matters: AI accuracy and relevance rests on fresh, high-quality data. Working with tech providers or retailer partners can enhance targeting and insight generation.
Example: Tap eCommerce Behavior Data (such as Purchase Frequency) for a retail partner to create an AI-driven re-engagement campaign where customers have lost due to curated product suggestions.
Learning for Marketing Leaders
It is important to integrate AI with business goals and measurable results. Advertising creative optimization, testing new products, or tailoring your complete customer journey. For strategy and brand voice, keep humans in the driver’s seat and let AI handle repetitive tasks and pattern exploration. By starting to build small areas of work and focusing on measurement, you will see faster and more reliable benefits that will allow you to scale your overall marketing operations.
Important insights from each AI session
Without registration, all three AI-centric sessions were available below. However, if time was running out, we chose to share important insights from each session.
Use Generation AI Tools for Content Creation
Important insight: Use “metaprompt” to help AI write prompts. In other words, one AI model (e.g. Claude) generates a high-quality prompt for another model (e.g. GPT-4). This approach saves time, generates more targeted instructions, and produces better AI-driven outputs.
Adopts Agent AI
Key insights: distinguish between mere automation (“done with you”) and true AI agents (“done for you”). True AI agents autonomously process tasks without overseeing each step, but automation requires direct participation. Knowing the difference prevents overpayment of “agents” that are merely automation, and full autonomy helps plan business meaning.
Real-world marketers who share real AI success stories
Key insights: Personalize at scale using AI to measure efficiency and performance lift. Marketers in this session successfully performed A/B tests by comparing the A-Optimized campaign to traditional campaigns and quantifying improvements in conversions, ROA, and time/cost savings. Ultra-personalization initiatives driven by AI can unlock better campaign outcomes and operational efficiency.
You can see the entire three sessions below.
