• by Andrew Gall , Op-Ed Contributor, 5 hours ago

The marketing world is racing to operationalize generative AI, but most brands are still experimenting in silos, struggling or failing miserably. 

Earlier this year, Northwestern’s Spiegel Research Center offered up a structured model designed to help brands use Gen AI more effectively. The framework, called ITMS (Introduce, Test, Measure, Scale), was published in the Journal of Applied Marketing Analytics.

Here’s how to use this framework to help launch a product for your CPG brand:

Introduce: Set guardrails and choose focus areas. First things first: Brands need to decide where AI belongs and where it doesn’t. Create guidelines to define acceptable AI use in packaging copy, product claims, and consumer data privacy. Start with content-heavy, low-risk areas like product descriptions, social captions, or shopper marketing assets. Marketing, legal, and R&D should all be setting the boundaries for AI-generated claims.

Let’s say you’re a yogurt brand usingGen AI to brainstorm new flavor names and create social captions inspired by trending cultural phrases. That’s great, but you should have some checks, including having legal review all outputs before testing with consumers.

Test: Run pilots with specific goals in mind. Instead of vague “let’s see what happens” experiments, tie your trials to specific objectives that link to content, creative, and customer engagement.

• Content generation: Use AI to assist with variations of ad copy for a single product line to test engagement.
• Creative testing: Use AI to help develop packaging design mockups, retail display concepts, and so on.
• Customer engagement: Deploy AI-powered chatbots for things like recipe ideas or product pairing suggestions.

If you’re a frozen-meal brand looking to run a 10-day pilot, you could compare human-written and AI-generated Instagram captions to see which style drives stronger sentiment and better engagement.

Measure: Quantify what works with hard metrics. Link AI initiatives to specific, quantifiable outcomes. For CPG brands, that means tracking both efficiency and effectiveness. Efficiency can be defined as time saved in content or asset production, like a 40% faster shopper marketing turnaround.

Effectiveness is more complicated. You’ll want to look at performance, ROI, and trust. Performance can be measured through a lift in click-through or sales conversion from AI-personalized messaging. ROI can be defined by cost savings from AI-assisted creative development. Brand-safety adherence and voice consistency across all AI outputs should be your metric for trust.

A coffee brand could use AI to dynamically adjust messaging by region, for example emphasizing “energy” in commuter-heavy markets and “flavor ritual” in suburban ones. It could measure in-store sales lift against control markets using traditional non-AI-enhanced creative.

Scale: Embed and expand accordingly.  Embrace new best practices, like embedding AI workflows into brand playbooks. Don’t forget to keep testing and learning. Shift to a mindset where AI augments existing human creativity.

If your beverage brand finds out its AI flavor-naming assistant reduces development time by 60% after your pilot, it’s time to make that a formal part of the process. 

The Bottom Line
Gen AI may be the key to locking machine-learning insights, but it still takes human-led creative power to bring products fully to market.