Key Takeaways:
Amy Errett, the CEO and founder at Madison Reed, said that her phone frequently dings with requests from friends asking for help getting an appointment at the company’s hair color bar in the Flatiron neighborhood in Manhattan.
“Good luck, that’s all I can tell you,” Errett told BeautyMatter. “Flatiron is our biggest store and our flagship, and you can’t get appointments, period.”
Excessive demand can result in lost revenue for a hair color bar operator like Madison Reed, which is why the company has publicly unveiled a new artificial intelligence agent called Madi this week. Madi was designed to make online bookings more seamless, using natural language to have an online conversation with guests.
Rather than a customer maneuvering through the app on their own to look for a free appointment at every individual Madison Reed salon in the region, Madi’s agentic capabilities allow it to offer alternative bookings for the Manhattan guest who may be willing to visit a nearby location in the Upper East Side or Tribeca.
The service was tested for over nine months, and Errett said that Madi has proven to double the likelihood that a guest would book an appointment at the company’s 97 hair color bars. The agentic system is also rolling out for two other priority use cases: online product discovery and customer support.
“The thing about AI is that it isn’t effective as a general strategy,” Errett told BeautyMatter. She believes that fewer use cases would be more impactful for Madison Reed, boosting both revenue and improving the company’s net promoter score, a metric that brands focus on to gauge customer loyalty.
Consultancy McKinsey has estimated that the beauty sector can broadly add $9 billion to $10 billion to the global economy by leveraging generative AI tools, although many of the most impactful potential use cases, such as virtual try-on and online chatbots, are still far from perfect.
Founded in 2013, Madison Reed was conceptualized to sell more prestige, salon-quality hair color formulations for the 50% of women who would dye their hair at home. The intent was to convert those women to Madison Reed’s direct-to-consumer business, though the business has since expanded and now sells in Walmart, Ulta Beauty, and other national retailers, as well as the company’s own hair color bars that employs around 1,100 stylists.
Since its inception, Errett said Madison Reed has utilized AI algorithms to better understand what women were looking for when coloring their hair. The company would ask a series of 18 questions, such as “What are your hair goals?” and “How often do you color your hair?” to make a recommendation for colors ranging from Pescara Black to Amalfi Blonde. Errett lauds the brand’s loyal customer base, as 70% of revenue is derived from membership subscriptions.
“We’re very focused on getting the color right,” Errett said.
The original questionnaire was developed to ask a series of 18 questions, and over the course of Madison Reed’s history, it has compiled data from 20 million unique hair profiles.
But with Madi, Errett said the company discovered that there are really only five questions that matter to accurately predict the right color. Madi also offers tips about adjacent products, like shampoos and conditioners. Around 20% of Madi’s interactions today are for product discovery.
The third use case is for questions around customer service, including changes to a home address or when a customer may want to adjust how often their hair color kits are refilled. Madison Reed said human staff is closely monitoring every question that’s being asked of Madi, and over time, answers have gotten sharper.
“We still have a very large customer service department,” said Errett. Her intention with AI is that the technology can take on some of the more simplistic requests, allowing for the customer service-focused stylists and cosmetologists to help with more complex questions.
AI is also working in the background to more efficiently plan staffing levels at the company’s hair color bars. Algorithms are now used to predict demand patterns based on a mix of internal, backward-looking data, as well as external factors that affect when women come to the salon, including weather and even what day of the week a holiday may fall on in any given year. AI makes automated recommendations to help general managers plan staff scheduling with greater certainty.
Errett said that before AI, utilization, which is the measure of how many hair color bar chairs are in use each day, was in the low 70%. That figure has increased to the mid-80% after the application of AI. Conversely, for times when demand may lull, AI-written automated messages are blasted out to members to offer discounted rates for those who may want to grab a last-minute open appointment.
“The key is getting somebody in,” Errett said.
