AI means marketers don’t need all the answers, but instead must ask the right questions and have the taste and experience to vet AI’s suggestions. Without the right tools or talent to fact-check responses, AI will “mansplain” to you, speaking confidently about things it doesn’t truly understand.

AI will change what makes customers feel special or seen as personalized marketing becomes cheap, ubiquitous, or even uncanny. Marketers have to thread the needle to come off as clever but not creepy.

But when applied with proper caution and creativity, AI is a massive force multiplier for small teams—it will allow them to refocus on strategy and empathy while AI handles the execution at scale. In turn, that could shift the makeup of teams, from including lots of junior employees implementing a top-down strategy to having fewer, more creative teammates experimenting and iterating with AI. Those who fail to adapt to using AI may find themselves obsolete: past shifts in technology were slower, but these new tools are proliferating too quickly to let employees ride out the rest of their careers relying on traditional skills.

Tofu and one of our investors, SignalFire, recently assembled a roundtable of CMOs from late-stage and public companies. Our goal was to define how AI is changing the pillars of marketing. Here are the top insights.

How AI changes marketing

• Junior content marketers are rightfully terrified. AI threatens their job security because it can produce content faster, cheaper, with depth across a broad range of topics, and without long back-and-forth editing cycles.
• CMOs believe that they or a senior content marketer equipped with AI can do the work of multiple teammates in less time.
• Heads of content and design with years of practice are becoming prompt engineers, using their taste and ability to recognize quality to curate what AI produces.
• Content still reigns, and it’s now easier to create large volumes of expert-level, bottom-of-funnel content, customized to different channels. This could replace haphazard, top-of-funnel content written by human generalists for a single channel.
• Marketers may be able to offset some AI content quality issues through quantity. For example, they can use AI to pull sound bytes and highlights out of long-form videos and turn them into content for every medium and platform, and turn a weekly newsletter into a daily one that has a slightly higher unsubscribe rate but a significantly higher growth rate.
• Chat engine ranking is the new SEO. Marketers are seeking ways to influence and measure how their products are recommended by AI interfaces like ChatGPT.

‎How AI will change sales

• Demand gen and biz dev employees are less worried about AI, which they think will enhance their efficiency and measurable results, rather than taking their job.
• However, simple transactional sales people are likely to be replaced with AI.
• CMOs are already using AI to summarize past customer conversations, company research, and contextual info to help sales reps get up to speed and personalize their pitches.
• Expect increasing “personalization blindness,” like ad-banner blindness. As every type of marketing becomes more tuned to individual customers, they’ll stop seeing personalized communications as novel or thoughtful.
• Enterprises are sitting on gold mines of data they don’t know how to excavate. The next big opportunity is training AI to proactively surface patterns or strategies from across large data sets, like what types of vocabulary are more or less effective on sales calls, or if more deals close when reps bring up certain features first.
• Salespeople currently spend up to 70% of their time updating their CRMs, but AI will take over that grunt work, freeing them up for more strategic or empathetic work.

Risks of using AI for marketing

• Constant changes by large-language-model providers can quietly degrade performance for certain use cases, so you can’t rely on this type of AI for consistent responses over time.
• Blindly cut-and-pasting AI outputs into customer-facing communications is negligent and can cause costly gaffes. Always edit and refine.
• Never put anything confidential into a generative AI tool, as it could turn into training data and cause leaks.
• Some regulated sectors like healthcare and finance are extremely cautious about how they use LLMs to avoid legal problems.
• AI won’t replace salespeople for large enterprise accounts anytime soon, as the complexity of aligning many stakeholders and relationship nuances make the risk of a deal-breaking mistake too high.
• AI-powered social engineering and deep fake scams are on the rise, so double-check outgoing payments and don’t believe someone just because they sound or write like a teammate.
• But there’s also a giant risk to not using AI and being leapfrogged by competitors. CMOs are encouraging bottoms-up experimentation by their teams to test AI tools and determine which to adopt more widely.

Popular AI marketing tools include:

• AdCreative.ai for making animated creative based on your landing pages
• ChatGPT for marketing and content copywriting, but with no confidential data entered
• CreativeX for managing adherence to style guidelines
• People.ai for prioritizing sales prospects and troubleshooting funnel problems
• Midjourney for making fine-tuned AI ad creative and content imagery
• Grammarly for editing AI and human writing for spelling, tone, and grammar
• Otter.ai and Fathom.ai for meeting transcription and summarization

Finally, one surprising secondary effect of AI is that it’s creating an “authenticity vacuum” that businesses can fill by doing marketing that AI can’t: events. As customers get more skeptical of AI-personalized marketing, atoms-based real-world marketing such as events carry more weight for building trust and deepening relationships. And with remote work leading fewer people to get their social needs met by office life—and long-distance business travel shifting towards Zoom—it’s become easier to get people to attend local events, where you can envelop a prospect in your brand.

‍