The IntelliBowl Nutrition Engine
IntelliBowl is not a search filter or a brand-sponsored ranking list. It is a dedicated nutritional intelligence system: a proprietary AI model purpose-built to reason about dog food the way a trained nutritionist would. Every recommendation it generates is the output of a multi-layered scoring engine that evaluates more than 4,000 commercially available dog food products against a single dog's complete biological and lifestyle profile. No brand can influence where it lands.
What powers the engine
The IntelliBowl model was built from the ground up as a canine nutrition specialist. It was fine-tuned on the current AAFCO Dog Food Nutrient Profiles (the federal reference standards defining minimum and maximum levels of 37 essential nutrients across every life stage) and on WSAVA feeding guidelines, which assess manufacturer credibility based on research investment, nutritional expertise, and feeding trial evidence. Beyond these foundational standards, the model was trained on a curated library of peer-reviewed research in canine dietetics, clinical nutrition, and veterinary internal medicine, keeping its knowledge aligned with the most current science in the field.
Critically, the engine's decision logic was not designed in isolation. IntelliBowl was built in direct consultation with certified veterinarians, who helped translate clinical knowledge, including breed-specific disease risks, life-stage nutritional transitions, ingredient-sensitivity patterns, and conditions requiring dietary management, into the structured rules the engine applies during scoring. The result is a system that carries the reasoning of veterinary expertise at scale, applied consistently to every recommendation it produces.
Core principles
- · Recommendations are nutrition-first and brand-agnostic. The engine has no commercial allegiance.
- · No brand pays for placement. No brand is excluded from consideration. Every product is evaluated against identical criteria.
- · The model's scoring logic was designed with input from certified veterinarians and trained on peer-reviewed research.
- · Affiliate links are appended after ranking. They are attached in a completely separate step and have zero influence on scoring.
- · Medical and therapeutic diet decisions should always go through a licensed veterinarian.
Knowledge sources
The engine's nutritional intelligence draws from five authoritative knowledge sources, each serving a distinct function in the recommendation pipeline:
- AAFCO Nutrient Profiles. The Association of American Feed Control Officials defines the nutritional floor and ceiling for commercial dog foods in the United States. IntelliBowl uses the current AAFCO profiles for adult maintenance, growth, and all life stages as a hard filter: any food that fails to meet the minimum nutrient targets for a dog's life stage is removed from consideration before scoring begins. This ensures that every food that appears in a recommendation has already cleared the baseline standard for nutritional completeness.
- WSAVA Feeding Guidelines. The World Small Animal Veterinary Association guidelines provide the framework for evaluating manufacturer quality beyond the label. The engine uses WSAVA criteria to assess whether a manufacturer employs board-certified veterinary nutritionists, conducts AAFCO feeding trials, publishes nutritional research, and maintains rigorous quality control. These factors contribute to a manufacturer quality signal in the engine's scoring model.
- Peer-reviewed research in canine nutrition. The model was trained on a curated library of current scientific literature in veterinary dietetics, clinical nutrition, and canine health. This includes research on breed-specific metabolic profiles, life-stage energy requirements, common dietary sensitivities, and the nutritional implications of ingredient processing methods. This research layer allows the engine to go beyond label compliance and reason about food quality in a clinically informed way.
- Veterinarian-designed clinical decision rules. A set of structured rules governing how health conditions, breed risks, and ingredient sensitivities are weighted during scoring was developed in direct consultation with certified veterinarians. These rules encode clinical judgment (the kind that would otherwise require a vet appointment) into the engine's logic, applied automatically to every dog profile submitted.
- 4,000+ product catalog. Nutritional data sourced from product labels, manufacturer-published guaranteed analyses, and ingredient declarations across a broad cross-section of the commercial dog food market. Not a single brand's lineup. Not a curated shortlist. The full market, evaluated on a level playing field.
How the scoring engine works
When a dog's profile is submitted, the engine does not retrieve a pre-built list. It computes a fit score for every eligible product in the 4,000+ catalog in real time, applying four weighted signals in sequence. The output is a ranked list of foods ordered by how well each product fits that specific dog, not by brand popularity, not by margin, not by what's trending.
- 1. Nutritional adequacy gate. Before scoring begins, the engine applies AAFCO nutrient minimums as a hard filter. A food that does not meet the minimum protein, fat, essential amino acid, or micronutrient targets for a dog's life stage is excluded entirely. This gate ensures the recommendation pool contains only nutritionally complete and balanced products, as defined by the industry's governing standard. Foods that fail this gate do not receive a score. They are not ranked low; they are removed.
- 2. Individual profile fit. Breed size, age, reproductive status, current weight, and activity level are used to compute life-stage energy requirements and adjust macronutrient targets to the individual. A highly active large-breed adult dog has materially different nutritional needs than a sedentary small-breed senior, and the engine treats them differently. Health history, including prior diagnoses, known sensitivities, and conditions flagged in the quiz, narrows the eligible field further using the veterinarian-designed clinical decision rules.
- 3. Ingredient constraint enforcement. Declared ingredient avoidances and known sensitivities are applied as strict hard exclusions. A food containing a flagged ingredient does not appear in results regardless of how high its nutritional score would otherwise be. This signal cannot be overridden by any other part of the scoring model. Safety constraints take absolute precedence.
- 4. Practical household fit. Budget range, preferred food form (dry kibble, wet, freeze-dried, raw, or mixed), and feeding logistics are used to surface options that are realistic and sustainable for the owner. The engine does not recommend a food that is nutritionally ideal but financially or logistically inaccessible to the household submitting the profile.
Brand neutrality, by design
No brand has paid for a position in IntelliBowl results. The scoring engine evaluates every product in the catalog against identical criteria. A premium brand and a budget brand are processed through the same four-signal model, with the same weightings, every time. There is no sponsored tier, no featured placement, and no preferential treatment based on manufacturer relationships. Affiliate commissions are earned only when a user clicks through to purchase a product from their results page. The existence of an affiliate relationship is established after ranking is complete and has no influence whatsoever on which foods are scored, how they are scored, or in what order they appear.
This is not a policy commitment. It is a structural feature. The ranking engine and the affiliate-link system are two separate, sequentially ordered processes. The engine produces its output first. Affiliate links are appended to that output second. The second step cannot modify the first.
What IntelliBowl is not
IntelliBowl provides nutritional guidance, not veterinary advice. The engine is designed to help owners navigate the commercial dog food market and identify products that are nutritionally appropriate for their dog's specific profile. It is not a diagnostic tool. It does not evaluate or treat medical conditions. It does not prescribe therapeutic, hydrolyzed, or veterinary-exclusive diets. Dogs with active medical conditions, complex health histories, or conditions requiring dietary management, including kidney disease, pancreatitis, diabetes, or severe food allergies, should always have their food selection reviewed and approved by a licensed veterinarian. IntelliBowl is a starting point, not a substitute for veterinary care.