FreshAi Transforms Drive-Throughs With AI-Driven Accessibility And Efficiency - Episode Hero Image

FreshAi Transforms Drive-Throughs With AI-Driven Accessibility And Efficiency

Original Title:

TL;DR

  • FreshAi's AI-driven drive-through system handles 95% of orders, demonstrating enterprise-scale AI impact by processing 150,000 daily orders with 60% reduction in customer-facing "sorries."
  • AI integration in drive-throughs enhances accessibility by enabling orders in multiple languages and accommodating atypical speech, overcoming communication barriers and improving customer experience.
  • The FreshAi platform reduces order-taking turns from 12 to 3, significantly improving efficiency and customer satisfaction by minimizing friction and wait times.
  • Wendy's leverages AI to identify non-obvious customer friction points, such as the agent's use of "sorry," leading to targeted improvements and a better overall experience.
  • The development of FreshAi highlights the iterative nature of AI product creation, emphasizing continuous learning and adaptation based on real-world customer interactions and feedback.
  • AI in customer service requires a shift in metrics, moving beyond traditional KPIs to track subtle indicators of friction and customer sentiment, like the frequency of apologies.
  • Innovative AI applications, inspired by experiences at Disney World, aim to transform the drive-through from a functional necessity into a delightful and engaging customer interaction.

Deep Dive

Wendy's FreshAi initiative demonstrates that artificial intelligence can deliver tangible enterprise-scale benefits by transforming the customer drive-through experience. This AI-driven approach moves beyond simple automation to create enhanced accessibility and novel customer interactions, fundamentally altering how the fast-food giant serves its diverse customer base and generating new insights from transactional data.

The core of FreshAi's success lies in its ability to handle the inherent complexity and variability of drive-through orders, which the company estimates can be as high as 200 billion ways to order a single menu item in English. By deploying an AI agent capable of understanding and processing orders in multiple languages, including Spanish, Wendy's is removing significant barriers for customers. This is not just about convenience; it's about inclusivity. For instance, the system's ability to display the order on a digital screen provides crucial confirmation and builds trust, particularly for customers with hearing impairments or those who are not fluent in English. The initiative has also prioritized accommodating atypical speech patterns, such as stutters, by integrating feedback from speech therapists and conducting rigorous real-world testing. This focus on accessibility, exemplified by the reduction in order "turns" from an average of 12 to 3, directly translates to a more efficient and less frustrating customer experience, preventing potential lost business and enhancing brand loyalty.

Beyond direct customer interaction, FreshAi generates valuable second-order insights by analyzing speech data. By tracking metrics like instances of "sorry" from the AI agent, Wendy's identifies friction points in the ordering process that might otherwise go unnoticed. This data-driven approach allows for continuous improvement, enabling the company to refine the AI's performance and enhance the overall customer journey. The initiative also highlights a broader shift in how businesses should approach AI implementation. Instead of simply applying AI for its own sake, Wendy's emphasizes understanding customer needs and leveraging AI to serve fundamental business goals, such as delivering fresh food with a great experience. This perspective suggests that successful AI adoption requires a deep understanding of core business principles, with AI serving as a powerful enabler rather than a replacement for them.

The implications of FreshAi extend to the workforce and the future of retail. By automating routine order-taking, Wendy's frees up human employees to focus on more complex, "human" tasks, potentially leading to a more engaging and rewarding work environment. This hybrid model, where AI handles transactional efficiency and humans manage nuanced interactions, offers a blueprint for other industries looking to integrate AI. The success of FreshAi underscores that the "normal" customer order is a myth, and that AI's ability to adapt to diverse needs and preferences is key to unlocking significant business and societal value, paving the way for more personalized and accessible services across various sectors.

Action Items

  • Audit FreshAi ordering system: Identify 3-5 areas where atypical speech or regional dialects cause order errors, and develop targeted training data.
  • Create a framework for measuring customer delight: Track new metrics like "number of 'sorries'" and "order turn count" to identify and reduce friction points.
  • Design accessibility features: Implement 2-3 new language options for FreshAi, prioritizing those with the highest demand based on customer data.
  • Develop a "what if" scenario playbook: Document 5-10 innovative customer experience ideas inspired by non-competitor industries (e.g., Disney, museums) for potential integration.
  • Measure AI value add: For 3-5 core drive-through functions, quantify the correlation between AI implementation and improvements in speed, accuracy, or customer satisfaction.

Key Quotes

"I was working for a company in Nashville I'm just by nature a pretty curious human and I remember the first time that I typed into ChatGPT it was a really simple prompt and it was like magic I put in the sentence I think it came back with like two sentences or something this was back in early 2023 and I just remember thinking wow this is incredible I'd never experienced anything like this."

Will Croushorn describes his initial encounter with generative AI, highlighting the "magic" and unprecedented nature of the experience. This quote illustrates the profound impact of early exposure to advanced AI, sparking his curiosity and motivation to explore its potential.


"What also stood out to me was there are very few moments in human history that are this democratized I was sitting around a table with the CEO and engineers and all our data scientists and I was a product person and we were all figuring out together literally every week we'd show up and put oh I learned this or oh I tried this and there's very few moments like that where nobody knows what to do with this technology and I had the thought of I want to help lead that."

Croushorn emphasizes the democratizing effect of emerging AI technology, where diverse roles and expertise converge in a shared learning process. He expresses a desire to lead in this nascent field, driven by the unique opportunity to shape its development collaboratively.


"For us that looks like creating accessibility it looks like creating experiences but it all stems back to wow I've never experienced anything like this before what are the cool ways that we can use that."

Will Croushorn explains that the core driver behind FreshAi's development is not merely automation, but the creation of novel customer experiences and enhanced accessibility. This quote underscores his belief that AI's true value lies in enabling capabilities previously unimaginable.


"The agent's going to be able to handle it and then on screen you're going to see on a digital menu board you're going to see your order we've done a lot of research to understand how do we help guests you know looking at the cognitive load order in an easier process and so one of the ways that we do that is we actually will show you the transcript so you can see what the agent is saying you can see what you're saying and it does a couple of things it gives you a little bit of trust to know oh that it heard me correctly and then also for customers or guests that might have hearing challenges it also just gives that reassurance of I know what's happening on the screen."

Croushorn details how FreshAi enhances the drive-through experience by displaying the order on a digital screen and providing a transcript of the interaction. He explains that this feature builds trust by confirming accuracy and aids customers with hearing impairments by offering visual reassurance.


"Nobody should feel that if you pull up and you don't speak fluent English that shouldn't be a barrier so how do we widen the road and so we've done that in a couple of ways we've built FreshAi in tandem with Google Cloud and so we've worked really closely with them to train out our Spanish model."

Will Croushorn highlights FreshAi's commitment to accessibility by enabling customers to order in languages other than English, specifically mentioning the collaboration with Google Cloud to develop a Spanish language model. He articulates the goal of removing language as a barrier in the drive-through experience.


"One of the pieces of feedback was the agent is interrupting too much and because of that it feels rude and so what we did was we said all right well what's the edge case like what's the most extreme that we probably should account for well one of those is atypical speech we have something like 10 of the us population has some sort of stutter and so we said all right well can the agent handle that."

Croushorn discusses the refinement of the AI agent based on customer feedback regarding interruptions, leading to a focus on handling atypical speech patterns. He explains the proactive approach to address edge cases like stutters, aiming to create a more inclusive and polite interaction.

Resources

External Resources

Books

  • "The First Day is the Worst Day of AI" - Mentioned in relation to the concept that AI systems improve over time.

Articles & Papers

  • "Neuroquery" (MIT Museum) - Mentioned as an inspiration for analyzing speech patterns to understand user behavior and efficiency.

People

  • Will Croushorn - Product manager at Wendy's, cocreator of FreshAi.
  • Sam Ransbotham - Host of Me, Myself, and AI, professor of analytics at Boston College.
  • Shayan Mohanty - Thoughtworks' Chief Data and AI Officer.
  • Kara Miller - Host of the podcast "It Turns Out."

Organizations & Institutions

  • Wendy's - Fast-food restaurant utilizing FreshAi for drive-through ordering.
  • Google Cloud - Partner in training Wendy's Spanish language model for FreshAi.
  • Thoughtworks - Company providing insights on responsible AI and data systems.
  • MIT Sloan Management Review - Producer of the Me, Myself, and AI podcast.
  • Massachusetts Institute of Technology (MIT) - Producer of the Me, Myself, and AI podcast, owner of the ME, MYSELF, AND AI® trademark.
  • MIT Federal Credit Union (MITFCU) - Offers financial products and solutions.
  • Disney World / Disney Imagineering - Cited as an example of creating engaging guest experiences.
  • DHL - Mentioned in relation to an early episode about AI.
  • Air Force - Mentioned in relation to a study about average pilot measurements.
  • National Football League (NFL) - Mentioned in relation to a study about average pilot measurements.

Websites & Online Resources

  • area.com - Website to see the Area enterprise AI platform in action.
  • mitfcu.org - Website for MIT Federal Credit Union.
  • thoughtworks.com - Website where insights on responsible AI and data systems are posted.
  • itturnsoutshow.com - Website to learn more about the "It Turns Out" podcast.

Podcasts & Audio

  • Me, Myself, and AI - Podcast exploring the future of artificial intelligence.
  • It Turns Out - Podcast featuring academic research and compelling minds.

Other Resources

  • FreshAi - Wendy's voice-based AI ordering system for drive-throughs.
  • ChatGPT - AI model mentioned as an example of experiencing AI's capabilities.
  • Area - Enterprise AI platform for speed and security.
  • ME, MYSELF, AND AI® - Federally registered trademark of Massachusetts Institute of Technology.
  • Tokens - Concept discussed in relation to responsible AI.
  • Mechanistic Interpretability - Concept related to understanding AI model neuron activation.
  • Agentic Enterprise - Concept related to the intersection of structured and unstructured data systems.
  • Easter Eggs - Hidden features or content within the AI ordering system.
  • Sorry Metric - A metric used to identify friction and issues in customer interactions.
  • Autopilot - Feature of a Tesla vehicle, used as an analogy for trusting AI.
  • iPhone Keyboard - Example of innovation through the merging of different fields.

---
Handpicked links, AI-assisted summaries. Human judgment, machine efficiency.
This content is a personally curated review and synopsis derived from the original podcast episode.