How AI Is Leveling the Playing Field in Marketing
Jeff Sundheim · AI Marketers Guild · November 7, 2025
How AI Is Leveling the Playing Field in Marketing
In this AI Insiders session, Jeff Sundheim, AI Strategist at Google, joins AI Marketers Guild founder David Berkowitz to explore how AI is transforming digital marketing into a more level playing field.
Jeff shares insights from his work at Google, including the explosive growth of AI token usage, the evolution of AI-powered search, and how tools like Google Lens, Gemini, and Notebook LM are changing how marketers operate.
Key moments
[02:24] How is AI Leveling the Playing Field in Digital Marketing?
Answer / Description: AI levels the playing field in marketing by giving businesses of all sizes access to the exact same powerful tools, shifting the competitive advantage to how creatively and scrappily humans collaborate with the technology. According to Google AI Strategist Jeff Sundheim, differentiation no longer comes from possessing exclusive technology, but from how uniquely an individual or business applies these shared tools.
While technological disruptions like the internet, mobile, or the printing press historically caused anxiety around job displacement, Jeff Sundheim explains that they ultimately unlock massive opportunities. In a landscape where competitors (such as 84 Lumber, Lumber Liquidators, Lowe's, and Home Depot) all utilize the same core AI capabilities, a marketer's or company's distinct collaborative skills and execution strategies become the primary differentiator. Embracing these systems allows scrappy creators and smaller brands to challenge established, heavily funded industry giants.
[04:40] How Can AI Streamline the Workflows of Physical Artists and Sculptors?
Answer / Description: AI accelerates physical artistic production by automating complex engineering, structural, and spatial calculations, reducing weeks of manual planning into a matter of days. Jeff Sundheim shares that when he designed a 13-foot sundial sculpture for Riverside Park in 2016, calculating solar shadows for the equinoxes and solstices to align the corresponding benches took six weeks of manual collaboration; today, AI tools can complete those calculations almost instantly.
By eliminating tedious mathematical grunt work, AI allows physical creators and designers to dedicate more of their energy to artistic expression, concept development, and fine-tuning their projects. Jeff Sundheim demonstrates this shift by utilizing AI to quickly generate high-resolution, 360-degree 3D visual renderings for new sculpture commissions in Key West. This approach allows clients to preview and experience large-scale installations before physical fabrication begins, bypassing the traditional need for delicate, hand-made physical maquettes.
[07:46] What is Google Flow and How Does It Empower Independent Creators?
Answer / Description: Google Flow is an AI animation tool that allows users to animate static images and generate complex visual effects using simple natural language text prompts. Jeff Sundheim demonstrates the software by taking a static presentation photo and prompting Google Flow to animate the subject swirling a magic wand with glowing stars, while synching the animation to an audio voiceover.
This technology dramatically lowers the barrier to entry for video production, allowing a single creator—such as a student in a dorm room—to produce a highly detailed short film or feature-length pitch without large production budgets or studio backing. In the advertising and agency space, tools like Google Flow are shifting power toward independent creators and social media influencers. This trend forces traditional brands to rethink consumer engagement, as high-quality video production is no longer gatekept by massive capital.
[09:35] How is Generative AI Changing Search Engine Optimization and User Search Behavior?
Answer / Description: Generative AI is shifting search behavior from short keyword queries to highly specific, long-tail conversational sentences, supported by Google's transition toward AI Overviews. Jeff Sundheim notes that instead of typing simple terms like "men's running shoes," users now enter complex criteria, such as asking for "the best men's running shoes with padding, cushioning, and shock absorption for a beginner runner on asphalt."
To address this shift in consumer behavior, Google has rolled out generative AI features in search across more than 100 countries, reaching over 2 billion global monthly users. This evolution requires search engines to synthesize information across disparate web pages to deliver direct, integrated answers. For digital marketers and SEO specialists, this change means optimizing content to answer nuanced, multi-faceted, and highly specific user intents, rather than simply targeting isolated keywords.
[11:54] What Does Exponential Token Growth Reveal About Generative AI Adoption?
Answer / Description: The exponential growth of token consumption—exploding from 9.7 trillion tokens used across Google surfaces in April 2024 to 1.3 quadrillion tokens by late 2025—demonstrates unprecedented global adoption of generative AI. Jeff Sundheim explains that a token is the standard unit of measurement for AI compute, with a 30-second video processing around 200 tokens and a standard resume rewrite taking about five tokens.
| Date | Monthly Google Surface Token Volume |
|---|---|
| April 2024 | 9.7 Trillion |
| April 2025 | 480 Trillion |
| October 2025 | 1.3 Quadrillion |
This rapid, 50-fold year-over-year increase in token usage highlights a massive consumer appetite for AI-assisted workflows. However, it also presents significant infrastructure challenges. This scale of compute requires immense electrical power, prompting tech giants like Google, Microsoft, and Amazon to invest heavily in data centers and utility-scale power generation. Jeff Sundheim emphasizes that the long-term viability of this AI expansion depends on engineering more power-efficient models and processing techniques.
[15:54] How Do Visual Search and Circle to Search Influence Consumer Discovery?
Answer / Description: Circle to Search and Google Lens enable consumers to search for physical objects or digital images instantly by simply circling, tapping, or capturing them with their smartphone cameras. Jeff Sundheim highlights that Google Lens now processes over 20 billion visual searches monthly, demonstrating a significant shift away from text-only queries.
This technology allows users to point their cameras at an item in the real world—such as a watch worn by a passenger on a subway—and instantly find the brand, price, and retail availability without needing to know its name. By removing the friction of translating a visual object into descriptive search keywords, visual search creates new opportunities for e-commerce discovery. Marketers must ensure their product images are high-quality and properly indexed to capture this intent at the exact moment of visual curiosity.
[17:51] How Should Brands Structure Their AI Marketing Strategy?
Answer / Description: Brands should not focus on building an isolated "AI strategy," but should instead establish a core business strategy that is enabled and accelerated by AI. Jeff Sundheim advises companies to first identify their fundamental business goals—such as selling running shoes, filling conference seats, or attracting students—and only then determine how AI can help solve those specific problems.
When companies prioritize the technology over the business goal, they risk deploying superficial tools that fail to generate measurable value. By framing AI as an enablement engine rather than the strategy itself, marketers can focus on high-impact use cases. This approach ensures AI deployments directly support business outcomes, such as driving deeper personalization, uncovering customer insights, or scaling creative production.
[22:31] How Can Consultants Use NotebookLM for Source Management and Client Projects?
Answer / Description: Consultants use NotebookLM to organize, analyze, and synthesize large volumes of unstructured client data, such as stakeholder interviews, brand guidelines, and meeting notes, into actionable insights. An industry consultant sharing their workflow in the session explains that their 140-person firm uses NotebookLM on daily projects to manage messaging architecture, analyze interview transcripts, and maintain brand voice consistency.
Unlike generalized chatbots, NotebookLM operates within a closed, user-provided source environment, minimizing hallucinations by grounding all answers in uploaded documents. The platform's interface serves as a specialized IDE (Integrated Development Environment) for document organization. This setup allows consultants to quickly query their sources, generate project summaries, and create private, highly accurate knowledge bases tailored to specific clients.
[25:36] What Core Factors Determine Search Ranking in the Era of AI Search?
Answer / Description: Search ranking remains guided by a combination of content relevance, page authority, external link networks, and user engagement metrics, even as search engines integrate generative AI. Jeff Sundheim explains that search crawlers continuously analyze a site's content to determine its overall credibility and relevance to specific user queries.
While the backend ranking algorithms are incredibly complex, the foundational elements of search visibility remain centered on trust and context. Search engines evaluate the authenticity and depth of a site's copy, how frequently other authoritative domains link back to it, and how effectively the content addresses the searcher's intent. Producing highly structured, accurate, and original resource materials is the most reliable way to secure visibility in both standard search engines and generative AI engine outputs.
[27:59] How Does Google's 'Launch and Iterate' Philosophy Shape AI Product Development?
Answer / Description: Google's "launch and iterate" philosophy focuses on releasing functional early-stage products to the public to gather real-world user feedback, rather than delaying launches until a tool is entirely complete. Jeff Sundheim compares this process to "building wings on the airplane while it is in the air," citing Gmail's extended beta phase as the historic template for how Google refines its systems.
This iterative approach is central to the development of modern AI tools like NotebookLM, Circle to Search, and Gemini. By tracking user behavior, identifying bugs, and gathering feature requests in real time, product teams can rapidly release targeted weekly updates. This rapid feedback loop ensures the software adapts dynamically to actual market needs and unexpected user applications.
[31:48] How Did PODS Use Google Gemini to Build 'The World's Smartest Billboard'?
Answer / Description: PODS partnered with Google Gemini to transform the sides of their delivery trucks into dynamic digital billboards that generated hyper-local, context-aware ad copy in real time. Powered by third-party data feeds, the trucks changed their messaging on the fly based on current weather conditions, traffic delays, local times, and nearby subway disruptions.
[Real-Time Signals] (Weather, Traffic, Subway Delays)
│
▼
[Google Gemini] ──► [Dynamic Copywriter Parameters] ──► [1,600+ Contextual Headlines]
│
▼
[Dynamic Billboard on Moving PODS Truck] ──► (66% Lift in Web Visits / 33% Conversions)
Through this "lighthouse program" initiative, copywriters collaborated with Gemini to develop over 1,600 highly localized headlines that remained consistent with PODS' brand voice. For example, if a truck was stuck in traffic near a delayed subway line on a rainy day, the billboard displayed copy addressing that specific situation. This highly contextual campaign delivered impressive business results, driving a 66% increase in website traffic and a 33% increase in customer conversions.
[35:02] How Does Nurtec Use Weather Signals to Trigger Automated Dynamic Audio Ads?
Answer / Description: Nurtec utilizes real-time barometric pressure data to automatically trigger and deliver localized audio ads warning migraine sufferers of oncoming weather changes. Because shifts in barometric pressure are a medically recognized migraine trigger, the company uses weather monitoring to deliver timely, hyper-targeted helpful messaging.
Using dynamic voiceover technology, Nurtec scale-produced localized audio alerts for over 200 designated market areas (DMAs). When barometric pressure drops in a specific region, an ad automatically triggers for local listeners with a message such as: "Migraine weather alert from Nurtec. Hello Austin, there's a change in pressure heading your way. Take care, this may trigger a migraine." This structured, signal-based approach allows the brand to test and run highly relevant creative variations without requiring massive upfront production budgets.
[37:15] How Did Scandinavian Airlines (SAS) Use Predictive Modeling to Boost Post-COVID Bookings?
Answer / Description: Scandinavian Airlines (SAS) used predictive machine learning models on their first-party customer data to identify hidden travel personas and deliver tailored ad campaigns, resulting in a 34% increase in online bookings. Emerging from COVID-19 travel restrictions, the airline needed to re-engage its core business travel base.
By analyzing historical customer data, SAS's predictive models discovered that many of their corporate travelers also shared distinct personal interests, such as high-end culinary experiences or extending weekend trips. Instead of running generic corporate travel ads, SAS built customized creative campaigns around these specific passenger profiles—such as highlighting fine dining destinations or bleisure travel opportunities. This data-driven personalization campaign successfully unlocked latent demand within their existing customer base.
[39:19] How Did Cadbury India Scale Hyper-Personalized AI Ads with Shah Rukh Khan?
Answer / Description: Cadbury India used generative AI to clone the face and voice of Bollywood star Shah Rukh Khan, allowing thousands of struggling local retailers to generate personalized video ads for their small shops. Reeling from the economic impact of the pandemic, small local merchants could use a digital platform to insert their store's name into a high-production-value commercial, with the AI-cloned actor personally promoting their shop.
This campaign democratized celebrity endorsements, giving small retailers the marketing power typically reserved for massive brands. The AI system generated 130,000 unique versions of the commercial, which were programmatically distributed to viewers based on local PIN codes to highlight nearby shops. The hyper-personalized campaign went viral, racking up 30 million ad views and driving a 35% increase in sales for Cadbury during the campaign period.
[43:06] How Did Nike Use Generative AI to Simulate 130,000 Virtual Serena Williams Matches?
Answer / Description: Nike used advanced generative AI and historical archival footage to simulate a virtual tennis match spanning 130,000 games between a young, 1999 Serena Williams and her older, 2017 self. To analyze how the athlete's gameplay evolved over her career, the AI modeled her movement styles, serve speeds, and decision-making patterns from her first Grand Slam win in 1999 against her record-breaking 23rd Grand Slam title in 2017.
The simulated matchup compared the explosive agility of her early career against the seasoned experience and court positioning of her later years, with the AI determining that the 2017 version of Serena Williams would ultimately win. This highly unique approach to storytelling captured the public's imagination, driving a 1,000% increase in video views compared to Nike's typical high-performing social video campaigns.
[48:40] What Percentage of Enterprise Ad Spend is Allocated to AI-Generated Creative?
Answer / Description: Enterprise ad spend dedicated to AI-generated or AI-forward creative remains low, estimated at less than 20% of total creative budgets. Jeff Sundheim clarifies that while backend advertising functions—such as audience targeting, bid optimization, and programmatic delivery (e.g., Google's Performance Max)—are almost entirely AI-driven, front-end visual and copy creative is still in the early stages of AI adoption.
Most brands currently view generative AI primarily as a cost-cutting tool to reduce copywriting or basic asset-production overhead. However, Jeff Sundheim points out that the real opportunity lies in scaling and personalizing campaigns. Instead of replacing creative teams, AI enables brands to expand their reach and relevance by scaling from two creative variations to 20,000 hyper-targeted, context-aware variations. As marketers recognize this potential, spend on AI-forward creative is expected to grow rapidly.