Back to Main
This Week in AI — 🚩 Charting the path forward for AI in marketing
AI

This Week in AI — 🚩 Charting the path forward for AI in marketing

October 28, 2025 | By: Dave Isaac

Each week we’ll gather headlines and tips to keep you current with how generative AI affects PR and the world at large. If you have ideas on how to improve the newsletter, let us know!

What You Should Know

 

The Numbers Say Most Marketers Still Struggle With AI. Here’s How to Change That

Marketers are storytellers, and a good, human anecdote is usually what makes readers feel connected. In the business world, though, it’s often the numbers that paint the clearest picture. While we know anecdotally that nearly everyone is using AI, new reports show where the industry really is in its AI journey.

Marketing Week’s latest State of B2B Marketing survey shows that 75.9% of B2B marketers say AI has become a bigger focus in their business this year, but 51.5% also note an AI skills gap — a bigger problem than shortages in analytics or martech expertise. While 83.4% are using ChatGPT, many apparently still don’t know how to translate that access into advantage. 

The Content Marketing Institute’s new B2B Content and Marketing Trends report shows a similar divide, but from a budget perspective. AI tools are the top area of planned investment for 2026 (45%), ahead of experiential marketing (33%) and owned media (32%), while human resources like training or team development sit at the bottom of the list at 9%. Nearly all say their organizations use AI-powered tools, but most classify their programs as “exploratory” or “developing.” 

The spending pattern confirms what many have suspected, that companies are racing to buy the technology before they’ve built the muscle to use it effectively. The tools are here, but the playbook isn’t. That’s not something an outside AI expert can solve, either. They won’t know the nuances of your business, and the technology itself changes so quickly that no universal playbook exists.

Each organization needs to build one for itself, and it starts with having the right culture. Here are five elements to get you started:

  • Create a safe space. Set clear rules and guardrails for data privacy, IP, and client content.
  • Experiment. With a secure environment or anonymized data, you can get creative with the kinds of stories you’re telling and figure out how much of your process AI can reasonably automate.
  • Document. Build a shared prompt library or process guide that evolves as the technology does, so you’re not treating today’s AI models like last year’s.
  • Learn from each other. Encourage your AI power users to share what they’re learning and help train others. Internal show-and-tells help everyone level up and understand how the technology applies to what your company does and how it operates.
  • Keep going. The models will likely change again next quarter. So should your approach.

Marketing runs on human connection, and when the industry learns to connect with AI as thoughtfully as it does with audiences, the results can be even more creative, more personal, and more powerful.

Elsewhere …

❓ Keep asking questions 

What’s happening: You may be hearing less and less about AI hallucinations as models improve, but they still exist. OpenAI recently published a paper that explains why models hallucinate and confirmed “Accuracy will never reach 100% because, regardless of model size, search and reasoning capabilities, some real-world questions are inherently unanswerable.”

Zooming in: OpenAI touts a lower hallucination rate for GPT-5, its latest model, but users report seeing the opposite, which underscores the importance of checking … well … everything.

Zooming out: Even if you ask AI to fact-check against material you provide, it can still get things wrong. We’ve seen AI claim things like “Yes, this point comes directly from the transcript you shared,” when that’s not the case, and if you ask it to quote the exact section of the transcript, it may finally fess up and say it wasn’t in there at all.

While AI can still make you more efficient, it’s imperative not to simply accept an output at face value. Asking specific questions (like “where did this come from?” for a specific portion of an output) will get you closer to a complete, factual piece of work. 

Quote of the Week

“We’d be doing a disservice if we weren’t pushing our employees to embrace AI, giving them the tools and training to do so, and really helping them grow their skill set. That’s going to be good for them, career-wise, whether it’s at eBay or somewhere else.”

— Jamie Iannone, President and CEO of eBay, during a fireside chat with Northeastern University President Joseph E. Aoun

How Was This Newsletter?

😀 😐 🙁