🎧 Case Study #1 – “The Best AI Conversation in Years”
Live Voice Chat with Ava, Store Manager’s AI Assistant
Topic: Product Inquiry – Howler Head Banana Bourbon
When Juan called ARC to ask about bourbons—specifically Howler Head Banana Bourbon—he was routed to Ava, the store manager’s virtual AI assistant. What followed was a seamless, human-like voice interaction that didn’t just answer his question but left him genuinely impressed.
Ava confirmed availability, answered follow-up questions, and offered options to continue the conversation or reserve the bottle. By the end of the call, Juan described it as “probably the best AI conversation I’ve experienced in years.”
Why This Matters:
- Proves voice AI can handle niche product inquiries accurately
- Reinforces trust by providing consistent, reliable information
- Showcases the emotional payoff when the tech feels human
- Reduces staff interruption during busy periods
- This kind of interaction is exactly why we test and refine relentlessly. The tech is only getting better—and so are we.
That same evening, Juan visited the store in person to purchase the 'Howler Head Bourbon'—and shared his praise directly, reinforcing what we’re building: not just automation, but a smarter, friendlier customer experience that delivers value in real time.
Press play to listen
💬 Case Study #2 – “Does Emily Know What a 8-Pack Is?”
Text AI Chat with Emily, ARC on Harvey’s Digital Assistant
Topic: Product Availability – Busch Beer (8-Pack)
A customer messaged Emily asking if we carried an eight-pack of Busch Beer. Emily politely replied that we didn’t—but that wasn’t quite true. The product was in stock, just listed in the inventory CSV as “8C of Busch Beer.”
This moment revealed an important learning opportunity. Emily didn’t misfire because of a lack of logic—but because of a language gap. Like most AI assistants, she was being overly literal, interpreting “8-pack” as a hard keyword rather than connecting it with equivalent formats like *“8C,” “pk,” or “cans.”
Rather than just fixing the response, we taught Emily how to interpret product shorthand and cross-reference phrasing like:
“8-pack” → 8C, 8pk, 8 cans
“case” → C, Case, Pack of __
Brand-specific formatting exceptions (e.g., “Bud 15pk”)
Why This Matters:
- Demonstrates how edge-case conversations drive smarter AI logic
- Shows our ability to train AI through improvisation and real-world feedback
- Highlights the importance of natural language interpretation in retail
- Avoids missed sales by improving product matching
- Today, Emily confidently translates everyday customer language into accurate product lookups—proving that even small misunderstandings can be powerful moments of growth.
The correct resonse!
🥃 Case Study #3 – “Irish Whiskey & the Art of Recovery”
Text Chat with Emily, ARC on Harvey’s Digital Assistant
Topic: Product Inquiry – Middleton & Irish Whiskey Selection
In this case, a customer asked Emily for a list of Irish whiskeys in stock. She responded with a detailed list… but missed one key item: Middleton Green Spot. When the customer followed up, asking specifically about products from the Middleton Distillery, Emily quickly realized the oversight, corrected it, and offered both Green Spot and Yellow Spot 12 Year Old with accurate pricing and availability.
Despite the initial fumble, Emily’s adjusted response impressed. She maintained the flow of conversation, answered nuanced questions (like the difference between Yellow and Blue Spot), and appropriately offered human assistance when asked about ordering out-of-stock items.
Why This Matters:
- The customer called out the gap directly, creating a pressure moment.
- Emily’s polite recovery showed a level of professionalism and adaptability that would pass for a human staff member.
- She successfully de-escalated the situation, guided the conversation, and concluded with a positive customer experience.
Key Takeaways:
- Demonstrates how AI can self-correct mid-conversation without losing trust
- Confirms Emily can handle multi-step customer logic while staying on-brand
- Highlights the importance of catching not just the first question—but the context behind the next few
- This exchange shows the value of conversational resilience—Emily didn’t get it perfect at first, but she got it right in the end.
What These Case Studies Show Us
Each of these interactions reflects a key theme in our AI journey at ARC: progress through real-world testing. These bots aren’t just a novelty — they’re becoming an essential part of how we serve customers with speed, accuracy, and personality.
From Ava’s natural-sounding voice chat to Emily’s text-based product lookups, these examples prove that even when things don’t go perfectly, our systems can adapt, improve, and ultimately deliver value.
Looking Ahead
As we continue to refine prompts, enhance logic, and expand to other ARC locations, one thing remains clear: the technology is working, and it’s working for our customers.
The next phase is simple — roll out these proven improvements across the entire network, keeping store-specific inventory and hours in sync, and maintaining the same high standard of service. With every test, we grow smarter. And with every conversation, we build loyalty.
Here’s to more great interactions, smarter systems, and a connected ARC experience like never before.