In partnership with
TGIF!
After a week of proof gaps and abandoned restructures, here is an AI story with a number attached to a shelf. Also, TGIF, so it’s fun!
Modern Retail reported on 29 August that The Hershey Company has started using a proprietary AI tool, built in-house, that uses computer vision and augmented reality to work out where s'mores displays should go in stores, and then measures how those displays perform against consumer data. The company also started using AI in its marketing-mix model this year to read consumer behavior closer to real time.
Hershey estimates s'mores drive $200 million to $250 million in sales across peak season, roughly Easter through Labor Day, with its own s'mores sales up double digits year over year.
Notice what is absent. No chatbot. No agent writing copy. No announcement about becoming an AI-first company. A merchandising problem as old as supermarkets, handed to a model that can see. The least fashionable AI story of the week, and the only one where somebody can point at the money.
Today’s sponsor: Whispr Flow
Reply to everything. Edit nothing.
Your inbox is full. Slack is piling up. Client messages need a response yesterday. Typing thoughtful replies to all of it takes hours you don't have.
Wispr Flow turns your voice into clean, professional text you can send the moment you stop talking. Speak like you would to a colleague — tangents and all — and get polished output. Emails, Slack, LinkedIn, WhatsApp, whatever's open.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Works on Mac, Windows, and iPhone.
TALK ON THE MARKET
Stephanie Berman, vice president of retail at The Hershey Company, in the same piece:
“The big thing for us as we look at AI: It's been about real-time data, it's been about getting closer to the consumer in more real time, and it's about partnering with our retailers on how we see the consumer behaving, so that we can iterate faster and more meaningfully with the consumer.”
She adds that it is all about real-time data, faster decisions, then optimization, in-store and online.
Read it again for what it does not contain. No content, no models, no agents, no transformation. An operator describing AI as an instrument for seeing sooner, which is a much older ambition than this year's vocabulary.
NUMBERS MATTER
45%. The share of marketers already using agentic AI that can act without human review.
Nearly half have handed a system the ability to send, score, and reallocate before anyone checks. Hershey pointed its AI at a shelf. Plenty of others pointed theirs at the customer.
Today’s partner highlight:
DAILY PROMPT
Create a detailed customer journey map for [PRODUCT/SERVICE NAME] targeting [TARGET AUDIENCE DESCRIPTION]. I need you to identify and analyze each stage of the customer journey.
## Journey Stages
Map the following stages:
- Awareness: How do customers first discover [PRODUCT/SERVICE]?
- Consideration: What information do they seek? What are their pain points?
- Decision: What factors influence their purchase decision?
- Retention: How do we keep them engaged post-purchase?
- Advocacy: How do satisfied customers become promoters?
## For Each Stage, Include:
1. **Customer Goals**: What does the customer want to accomplish?
2. **Touchpoints**: Which channels/platforms are they using? (e.g., social media, email, website, sales calls)
3. **Pain Points**: What frustrates or blocks them?
4. **Emotions**: How do they feel at this stage?
5. **Marketing Actions**: What specific marketing activities should occur here?
6. **Success Metrics**: How do we measure progress through this stage?
## Additional Context
- Current conversion rate: [X%]
- Average customer lifetime value: [AMOUNT]
- Primary competitors: [LIST 2-3 COMPETITORS]
- Key differentiators: [WHAT MAKES YOUR OFFERING UNIQUE]
- Budget constraints or channel preferences: [ANY LIMITATIONS]
Provide the journey map in a structured format that can be shared with cross-functional teams. Include specific, actionable recommendations for optimizing each stage.Use this when your AI investment keeps landing in the same two places, content and email, while the expensive parts of the journey go untouched. Hershey found value at the shelf because somebody mapped where the decision gets made. Yours may not be where your tooling is.
CMO CORNER
Your job is to create clarity and then disappear without chaos.
The Hershey work is a good test of whether you can. Deciding AI goes to merchandising rather than content is one call, made once, and everything downstream follows without you. Data plumbing, retailer conversations, display measurement, none of it needs a CMO in the room.
It needs somebody who said where the constraint was and did not change their mind every quarter. If your last four AI decisions all required your presence, your team doesn’t lack initiative. You never made the call clearly enough to act on alone.
We provide consulting services for CMOs on AI strategy for marketing teams through our Teamless offering.
TOOL UPGRADE
Getting closer to the customer in real time, without a retail media budget:
Manychat — best for turning Instagram, WhatsApp and TikTok conversations into a captured contact rather than a nice interaction. Try Manychat here →
Watermelon — best for resolving service questions across channels with a clean handoff the moment judgment is needed. Try Watermelon here →
BotPenguin — best for small teams who need WhatsApp, Instagram and web covered without an engineer on the project. Try BotPenguin here →
→ We have 300+ curated, filtered, and categorized AI tools in our directory. Browse to find the best ones that help your marketing team.
LEARN THIS
If this week has left you wanting somewhere concrete to start rather than another framework, 3 Ways to Level Up Your AI Marketing is deliberately practical.
It picks three moves a team can run without new budget or new headcount, which is the situation most marketing departments are actually in. For the leader who has heard enough about strategy and wants something to do on Monday.
→ We publish workshops, courses, guides, manuals, and more on AI marketing for free. Browse our learning resources for free.
BRAND SLOP
Two camera companies spent last weekend selling cameras with pictures no camera took.
PetaPixel documented both on 22 August. Xtra published an AI-generated video to Instagram, using animals, to sell its Muse gimbal camera. The generated clips sat in the same folder as more than twenty clips of real footage shot on the Muse 2 Pro. The company had the real thing and used the fake one.
Ricoh GR Europe appears to have generated a person rather than hiring one. The tell was the thumbnail: a man holding a Ricoh GR to his eye. The GR has no viewfinder. The images were quietly deleted, and the company did not respond when asked.
A camera exists to capture something that happened. Generating the advertisement admits you could not be bothered to point your own product at the world.