This month, we’re bringing you another round of posts highlighting AI’s understanding of some interrelated product categories. Our new focus is Cleaning Products space, and we’ll start with a look at Non-Toxic Cleaning Products. Each of these quick reads will offer a little data and a lot of insight into how Novi identifies category-specific strategies for upping AI visibility and consumer awareness.
Before we begin our review of Non-Toxic Cleaning Products, let’s refresh ourselves on some key metrics that inform these briefs.
AI is pretty literate when it comes to Non-Toxic Cleaning Products. Our models find category alignment of 70% along with a category fit score of .708. The sorting captured by those measurements rests on two of the most fundamental signals in AEO: Ingredient Transparency and Certifications. The strongest anchors are terms like EWG Verified, fragrance-free, EPA Safer Choice Certified, Made Safe Certified, free and clear, and they drive the 70% success rate that we see today.
Some of the misalignment we see in Non-Toxic Cleaning is occurring at the product type level within the category itself. If a dish soap, a glass cleaner, and an upholstery cleaner are all Non-Toxic and communicating that using the same signals, AI has a hard time knowing where to sort any of the three.
Across categories, we see some misalignment linked to Antiseptics and Cleaning Supplies, Liquid Hand Soap, Hand Sanitizers and Wipes, and Cosmetic Tool Cleansers. But the most significant confusion happens in connection with Pet-Safe Cleaning Products (where overlap measures .929) and Sensitive-Skin Cleaning Products (overlap of .879). The mix-ups make sense, since Non-Toxic, Pet-Safe, and Sensitive-Skin designations are all driven by a similar lexicon communicating safety and gentleness.
Language that specifies an intended use case like dishwashing or tile cleaning can also lead to misalignment. Consider AspenClean. Their Dish Soap, Bathroom Cleaner, and Glass Cleaner are all Top 10-ranked in terms of AI visibility, but they’re sorted incorrectly. Safety certification language like the terms noted above move AI in the right direction to start, but the format and application language cause confusion. Dish Soap and Bathroom Cleaner get sorted as Cosmetic Tool Cleaners, and Glass Cleaner is deemed an Eyewear Lens Cleaning Solution.
Non-Toxic Cleaning Products will be best served by descriptions that prioritize their non-toxic, certified-safe attributes. Application-specific language matters, but giving it second position will avoid pulling focus from the most important distinguishing feature. In practice, this means better visibility where it counts for a glass cleaner marked EWG Verified cleaner safe for use on glass and mirrors instead of glass cleaner, EWG Verified. It’s a subtle difference, but it’s meaningful.
Given AI’s perception of heavy overlap among Non-Toxic, Sensitive-Skin, and Pet-Safe Cleaning Products, achieving alignment with a precise balance of safety and category signaling will be a common theme throughout this series.
“First things first” is good advice. But what do you do when the correct first step isn’t clear? Novi’s product-level research and LLM modeling help brands understand all the pieces of the AI decision-making process. When every step is clear, you can trust that each move you’re making is the right one. Reach out to learn more about how to earn your products the visibility they deserve.
