Mall of America Deploys AI to Decode Real Customer Behavior
PLUS: AI Prompts for Scaling Support Excellence and Building Trust Through Transparency
Start every workday smarter. Spot AI opportunities faster. Become the go-to person on your team for what’s next.
🗓️ August 13, 2025 ⏱️ Read Time: ~5 minutes
👋 Welcome
Today, we've got everything from Mall of America using AI to watch how people shop, to Cresta finally making email support intelligent, to actual performance data showing AI agents can handle most customer issues without making people angry. It's like watching the industry grow up in real time.
📡 Signal in the Noise
The common thread across this week's stories? Companies are using AI to understand customers better, not just automate them away.
🧠 Executive Lens
What's interesting is how these AI implementations focus on reducing uncertainty for customers rather than showing off how smart the technology is. That's probably why they're working.
📰 Stories That Matter
🎯 Mall of America figures out what customers are actually doing
Mall of America deployed AI-powered video analytics to count cars in their parking areas. Their old system couldn't handle Minnesota weather (snow, rain, cold) and gave unreliable data. The new AI system from Axis Communications accurately tracks parking usage across dozens of locations throughout the mall. They use this data to make decisions about mall hours and event planning based on actual customer traffic patterns.
Why This Matters: Accurate parking data helps them understand real customer demand patterns, which drives better operational decisions about when to stay open and how to plan events.
Try This: Think about operational metrics you rely on that might be inaccurate due to weather, location, or other factors. Better data could improve your customer experience decisions.
Source: Retail Customer Experience
⚡ Cresta tackles the email support problem everyone forgot about
While everyone's been building chatbots, Cresta launched AI-augmented email support that brings real-time intelligence to what's still a huge part of customer service. Think about it: your chatbot might be smart, but then customers get generic, slow email responses that feel like they're from a different company. Cresta's trying to make all your support channels feel equally intelligent.
Why This Matters: Customers notice when one channel is way smarter than another. It creates this weird experience where your company seems inconsistent, which kills trust pretty quickly.
Try This: Look at your different support channels and honestly assess which ones feel "dumb" compared to others. Those gaps are probably confusing your customers.
Source: PR Newswire
📊 Legal & General connects CX improvements to actual revenue
Legal & General announced they're using Tealium and Snowflake to tie their AI-powered customer experience work directly to business growth. What's smart about their approach is they're not just trying to make customers happier, they're using AI to predict customer behavior that impacts revenue. Instead of measuring satisfaction scores, they're measuring business outcomes.
Why This Matters: Finally, someone's treating customer experience like a business function instead of just a "nice to have." When you can predict which customers are likely to buy more or leave, CX becomes strategic.
Try This: Pick one customer behavior that, if you could predict it a day or two early, would let you do something that impacts revenue or retention.
Source: Globe Newswire
📈 Salesforce shares real numbers on AI agent performance
Salesforce hit one million AI-powered conversations on their support site and shared some interesting data. Their AI agents are resolving 85% of issues without human help, with only 2% needing to escalate to complex cases. What's really telling is that human handoff rates dropped from 26% to 5% over time. That suggests customers are getting more comfortable trusting AI when it consistently works.
Why This Matters: These aren't demo numbers or pilot project results. This is what happens when AI agents handle real customer support at scale, and the numbers are pretty convincing.
Try This: Think about what 85% autonomous resolution would mean for your team's capacity and what you could do with all that freed-up time.
Source: SalesforceDevops.net
🔬 AI industry gets serious about autonomous agents
The big AI conference in Vegas this week (Ai4 2025) has over 8,000 people talking about "agentic AI," which is basically AI that can make decisions and take actions on its own. The focus wasn't on making AI more human-like, but on making it reliable enough to handle specific jobs without constant supervision. It's less about artificial general intelligence and more about artificial specific competence.
Why This Matters: The AI industry is maturing from "look what this can do" to "here's what this should own." That's a much more practical approach for business applications.
Try This: Look at your customer experience processes and identify where consistent execution matters more than creative problem-solving. Those are good candidates for autonomous AI.
Source: Ai4 2025
✍️ Prompt of the Day
Title: CX Quality Audit for AI Readiness
You are a CX quality auditor. Analyze our current customer service process and identify:
1. Interaction volume and complexity patterns
2. Common resolution bottlenecks
3. Tasks suitable for AI automation vs human expertise
4. Quality metrics that matter most
5. Change management considerations for AI adoption
For each area, provide specific recommendations with expected impact timelines. Focus on maintaining service quality while scaling efficiency.
Current process details: [INSERT YOUR PROCESS DESCRIPTION]
What this uncovers: Workflow inefficiencies and automation opportunities you probably know exist but haven't mapped out
How to apply it: Build a realistic AI adoption plan that addresses people concerns, not just technical ones
Where to test: Start with high-volume, routine interactions where consistency is more important than creativity
🛠️ Try This Prompt
Create a comprehensive AI agent training guide for customer service that includes:
1. Core conversation principles and brand voice guidelines
2. Escalation triggers and handoff protocols
3. Knowledge base integration best practices
4. Performance measurement frameworks
5. Continuous improvement feedback loops
Structure this as a practical playbook that both technical and non-technical team members can implement. Include specific examples for our industry: [YOUR INDUSTRY]
Focus on building trust, maintaining quality, and scaling human-centered support.
Immediate use case: Make your AI agents sound like your brand instead of generic robots
Tactical benefit: Reduce the confusion that happens when AI behavior is unpredictable
How to incorporate quickly: Use this to standardize how your AI handles different situations
📎 CX Note to Self
"Good AI feels like the company finally figured out how to help customers the way they always wanted to be helped."
👋 See You Tomorrow
That's Wednesday. Looking at today’s stories, there's a clear pattern: the companies getting AI right are using it to understand customers better, not just process them faster.
The interesting thing is none of these stories are about replacing humans. They're about using AI to make customer interactions more predictable and less frustrating.
What's the biggest source of unpredictability in your customer experience that AI could help fix? Hit reply and let me know.
Forward this to someone who's trying to figure out AI's role in customer experience.
Have an AI‑mazing day!
—Mark
💡 P.S. Want more practical prompts? Grab the FREE 32 Power Prompts That Will Change Your CX Strategy – Forever and start improving your customer experience approach today. 👉 FREE 32 Power Prompts That Will Change Your CX Strategy – Forever
Special offer for DCX Readers:
The Complete AI Bundle from God of Prompt
Get 10% off your first purchase with Discount code: DI6W6FCD