TL;DR
September polling shows limited but pragmatic adoption in pet food marketing: teams prioritize analytics over creative automation and struggle with measurement. Here’s how to pilot, measure, and scale responsibly.
Where marketers are actually using AI (and where they aren’t)
A September follow‑up poll (n=119) finds 56% aren’t using AI in campaigns yet. Among adopters, usage concentrates in customer insights & data analysis (21.21%), with content creation (11.36%), email personalization (6.06%), and visual/design (5.30%) trailing. The most‑cited measurable benefit is faster content production (15.38%), but two‑thirds (66.67%) haven’t quantified results—evidence that many proofs are still “toy pilots.” Fieldwork: Sep 4–18, 2025; published Sep 22, 2025. PetfoodIndustry
Adoption pattern: 56% of teams are still in planning/research; 25% limit experimentation to one or two people; 9% run department‑wide training; 3% partner with external specialists. And 63.33% say they’re satisfied with current tools—so AI must displace incumbents with clear incremental value. PetfoodIndustry
A 90‑day plan to move from exploration to evidence
Days 1–15: Frame the problem and the risk
- One business question per pilot. Examples: “Lift repeat among sensitive‑stomach dog owners by 8–10%?” or “Cut CPA for indoor‑cat hairball control by 12%?” Decisions must be pre‑tied to thresholds.
- Data hygiene sprint. Standardize pet type, lifestage, benefit claims, retailer channel, and basket attributes across CRM/PIM/ecomm; stale or inconsistent taxonomies undermine modeling and claims safety. The August poll flagged data quality and expertise as recurring hurdles—treat cleanup and training as part of the AI budget, not overhead. PetfoodIndustry
- Guardrails. Spin up a lightweight human‑in‑the‑loop review for nutrition claims and imagery, with an immutable approval log.
Days 16–45: Pilot two use cases
- Insight accelerator: cluster and score high‑value cohorts (e.g., “wet‑toppers + digestive care + monthly cadence”). Artifacts: segment definitions with size, reach, and expected LTV.
- Content velocity with controls: generate on‑label copy/visual variants mapped to 2–3 cohorts. Keep personalization rules‑based at first (e.g., benefit emphasis by species/lifestage) before model‑driven decisioning.
Days 46–90: Test, measure, decide
- Design clean experiments. Geo or audience holdouts; pre‑register lift thresholds (e.g., +8% CTR, +5% conversion, +10% 60‑day repeat).
- Instrument production ops, not just media. Track time‑to‑first‑draft, review rejection rate, cost per approved asset alongside CTR/CPA/ROAS. The September poll’s 66.67% “haven’t measured” is the gap to close. PetfoodIndustry
- Scale or shelve. Upgrade wins to playbooks (briefing checklist, prompt library, segment taxonomy). Archive learnings for failed tests and move on—no sunk‑cost drift.
What to prioritize (grounded in the data)
- Analytics over automation. Early adopters lean on insight and analysis far more than generative creative. Follow the signal: audience discovery, churn/risk scoring, trial‑to‑repeat prediction. PetfoodIndustry
- Content speed—with brand safety. Since content velocity is the most measurable gain, formalize templates, approved claim language, and asset metadata to reduce back‑and‑forth. PetfoodIndustry
- Sequenced personalization. Start with deterministic rules; progress to model‑scored recommendations only after you can prove uplift and explain eligibility logic.
Common pitfalls (and fixes)
- Jumping to hard problems first. Dynamic pricing and real‑time offer orchestration are Phase 2. Prove value with analytics + controlled creative variants first.
- Under‑equipping talent. The August poll names expertise the #1 barrier. Budget for hybrid talent (data‑savvy marketers; marketing‑savvy analysts) before you add tools. PetfoodIndustry
- Skipping taxonomy governance. If product benefits and lifestage attributes aren’t consistent, models can’t target or message correctly—tighten the taxonomy now to avoid later rework.
A pragmatic expectation set
With 63.33% of marketers satisfied with their current stack, AI has to earn its seat. You don’t need moonshots: a sustained few‑point lift in repeat for a strategic segment or 20–30% cut in content cycle time can justify continued funding—if you measure it and keep oversight tight. PetfoodIndustry
Source
Lisa Cleaver, Pet food industry takes cautious approach to AI marketing (Sep 22, 2025; poll fielded Sep 4–18, 2025; n=119). PetfoodIndustry
