In a live poll of pet food professionals, almost 93% said they’ve personally used an AI tool like ChatGPT. However, when asked what holds them or their companies back in using AI or data more effectively, 28% said they don’t trust it with their formulas or data, while about 25% said their leadership hasn’t prioritized it and nearly 14% said they don’t know where to start. The remainder responded that they didn’t believe their data was clean enough, or it lived in silos.
Such is the state of AI usage in pet food today, at least in operations and formulation. The polls were posed by Carmen Sook, director of customer solutions at Bestmix Software, during a Petfood Essentials presentation on April 27, 2026. She urged attendees to try simple steps that required no budget approval but could jump-start at least a data collection process that could lead to savings in time and money.
A presentation during Petfood Forum on April 28 picked up on the “data is the thing” theme when it comes to AI. In this case, it was about pet food marketing and highlighted the fact that in this realm, adoption of AI isn’t the real gap in effectiveness; rather, it’s measurement.
3 barriers to AI success in pet food marketing
The presenter of the Petfood Forum marketing session, Jolanta Smulski, founder of Pet Pro Media, did a quick poll of her own: When she asked how many audience members’ teams had used AI in the previous 30 days, about half raised their hands. Yet to a follow-up question — if anyone could say what the AI actually delivered in terms of numbers like conversion rate, time saved or cost reduction — only a few hands remained raised.
“Most pet food brands are not behind on AI,” Smulski said. “They’re behind on measuring whether AI actually works. And those are two very different problems.”
She identified three barriers to effective AI usage in pet food marketing:
- The measurement gap — Citing a 2025 PetfoodIndustry.com poll, 67% of pet food marketers using AI have never measured the results, Smulski said, and across all industries, it’s not much better: 51% of marketers cannot prove ROI on their AI investments. “If you cannot measure it, you cannot scale it. And AI will optimize the wrong thing, beautifully and efficiently. It’ll produce content that gets clicks but tanks conversion. It’ll generate ads that attract attention but push them to the wrong audience. The output will look like progress while the marketing budget quietly erodes. Measurement is what keeps AI honest.” And it must happen before you deploy a marketing campaign, not six months later.
- Planning paralysis — Smulski called this the most expensive barrier, sharing that 56% of pet food marketing get stuck in the planning and research phase, with 63% saying they’re satisfied with existing tools. “It’s not caution — it’s complacency dressed up as caution,” she said. “Here’s the data point that should make every leader in this room sit up: 42% of companies abandoned their AI initiatives in 2025, up from just 17% the year before. The reason? Most of them skipped measurement, couldn’t prove value and gave up. The cost of waiting is not zero, and it’s rapidly accelerating. Every month you spend evaluating is a month a competitor spends iterating.”
- The scaling challenge — “This is what quietly kills most AI programs in pet food,” Smulski said, adding that 91% of AI use in pet food companies is limited to individuals or small teams and never reaches the full organization. Using an iceberg analogy, she said the 70% underwater represents people, organization and process. “So the next time someone pitches you a new AI tool, remember: you’re buying 10 to 20% of the equation. The other 70% is on your team,” she explained. “This is not a technology problem. It's a people and process problem.”
The good news: Solving the problem
The scaling problem is actually good news, Smulski said, because it can be solved with another list of three. The key is following the steps in exact sequence:
- Step one — Measure first. Define your core marketing metrics before you adopt a single AI tool.
- Step two — Test and learn. “Run structured pilots on key use cases, with pre-committed expansion and kill criteria,” she explained.
- Step three — Scale capability. Take the data from those pilots and turn them into training and repeatable processes.
Smulski provided a 90-day roadmap for implementing the program, with 30 days each for establishing measurements (“define three to five outcome metrics — not vanity metrics”), running two pilots (content at scale and GEO [generative engine optimization] authority content) and scaling capability (taking the pilot results and turning them into a playbook).
“The reason this framework works is that it reframes the skills barrier as a sequencing problem rather than a technology problem,” Smulski explained. And it applies no matter the size of the company: Brief case studies from two pet food companies, large (Hill’s Pet Nutrition) and small (Spot and Tango), both illustrated success. “AI visibility does not correlate with budget,” she concluded.
