Artificial Intelligence, or AI, isn’t new. AI-driven applications have been around for years, from Gmail’s spam filter and Netflix movie recommendations to investment management by virtual advisors.
But AI took the world by storm when OpenAI released ChatGPT in November 2022, with Google’s Bard and Microsoft’s Bing Chat quick to follow in Q1 2023.
The intelligence in AI comes from computers parsing large data sets to “know” and predict all sorts of conditions such as which emails are likely to be spammy (Gmail), which movies you might like (Netflix), or which investment mix is the best fit for you. All are examples of machines learning from the vast amounts of data they parse.
Anyone can start playing with free versions of the latest available AI tools. “AI is evolving so quickly, that if you do not start learning how to use it, by the time it evolves and becomes really, really good, you’ll be left behind,” says Anand Arvind, Chief Technology Officer at Zenoti.
Generative AI is most often used to create text and images. Arvind names Dall-E from OpenAI, Firefly from Adobe, Midjourney, and Stable Diffusion for image creation.
As part of Innergize 2023, the premier beauty and wellness event, Arvind shared just a taste of what’s possible with AI. Here, we recap his insights on using AI to become more efficient and provide a better customer experience.
1) Save time managing customer feedback
Responding to customer reviews quickly is crucial for any brand. It’s not enough to track reviews and respond fast. Showing empathy and trying to make a connection is just as important, especially when dealing with negative feedback.
Arvind asked ChatGPT to generate an email response to a positive online review for a Zenoti customer. ChatGPT’s 75-word response hit all the right notes, with phrases like “we’ll be sure to pass along your kind words” and “your loyalty means the world to us.” Arvind was also able to prompt the tool for a shorter version, more appropriate for use as a text message.
Next, Arvind gave an example of having ChatGPT generate a response to a negative review. The well-worded response included an apology, appreciation for the feedback, and an offer of a free service to make things right.
A free service to win back the unhappy customer wasn’t part of the request – and, in fact, you can specify that condition. “You could say in the prompt,” says Arvind, “‘I want to respond to this negative feedback. I want to be empathetic, but I do not want to give them a complimentary service.’”
ChatGPT can also be seen as a source of ideas for work emails and everyday situations, especially complex ones. “You’ll be surprised at the kind of outcomes you get,” says Arvind, “and how you can improve how you respond to things.”
Key takeaway: AI saves your staff time when responding to customer reviews. It quickly generates drafts they can revise and personalize.
2) Identify what’s important to your customers
AI can spot the main themes in customer feedback. Simply upload publicly available reviews to ChatGPT and have it summarize why customers like or dislike your salon or spa. If the AI notices customers raving about your talented stylists but struggling with scheduling issues, you can act on that knowledge. You can celebrate your stylists to keep them motivated and improve appointment booking to address issues with scheduling.
Giving the AI free rein is one way to do it. Just feed it reviews and ask it to identify what customers think about your business. Or take it a step further: refine your prompts to “ask” AI to focus on a handful of categories you want to track.
Start with simple prompts to see what you get. Then, move on to more nuanced or complex prompts for information that can help you improve the guest experience. You can even refine prompts to reorder items from most to least important.
This exercise is just a glimpse into some of the most basic power of AI: ChatGPT summarizes review feedback in seconds, a fraction of the time it would take a staff member to read all the reviews, look for patterns, and summarize the findings.
When acting on feedback, however, be sure to add the human touch: high-five Amy for the glowing review her service earned the salon. Check into why appointment booking is a source of frustration for your customers.
Key takeaway: AI quickly analyzes feedback from customers to identify what they like and don’t like about your business. It helps you prioritize areas for improvement.
3) Improve operations by having AI generate summaries of large documents
As a business owner or manager, you may not always have time to comb through every report that comes your way. To be more efficient, use an AI tool to process large documents. You could have AI summarize a 30-page document to get up to speed in 2 minutes and move on to your other to-do’s.
As an example, Arvind used ChatGPT to cull takeaways from a 45-minute pre-event meeting transcript. He quickly learned the purpose of the meeting, the event structure, and event details like team assignments, the dress code, and even an appointment to keep at the event.
Important: When using an AI tool, be careful not to upload business data that is confidential, proprietary, or contains personally identifiable information (PII). Professionals in more heavily regulated industries, like healthcare and finance, should be extra careful with using AI for content generation and data analysis.
What about PDF documents? ChatGPT can summarize those that are available online. Simply give it the path to the PDF and ask any question about the PDF. Again, as an example, Arvind used this method to make short work of a lengthy, text-heavy PDF on Covid regulations for salons.
“It’s an easy way to look at a large PDF, ask a question, and get a summary,” Arvind explains.
Key takeaway: AI summaries of large documents free up time to focus on your other priorities, helping you become more efficient.
4) Analyze customer behavior in minutes instead of hours
If you run or manage a salon or spa brand and want to identify the most popular services or your busiest service days and times, you could use GPT-4, OpenAI’s most advanced tool, to analyze your data set.
Let’s say you have data with columns for appointment date, service name, center name, start time, and end time. You could upload the spreadsheet – with customer names removed – and ask GPT-4 questions like “What are the 10 most common services?” It can generate a table with the answer. If you like visuals, you can even have it create a word cloud with the more popular services shown in a larger font size.
You won’t have to do any parsing, categorizing, or data-sorting. The AI does it for you.
To illustrate ChatGPT’s way with data, Arvind had it generate a bar chart showing appointment volume for a salon/spa business between 10:00 a.m. and 8:00 p.m. In less than a minute, the tool returned a color-coded histogram for a data set with a few hundred rows.
Remember to be as specific as you can when asking questions or giving directions. When given vague prompts, ChatGPT makes incorrect assumptions.
“Because ChatGPT predicts using neural networks, it can make up facts – always double-check historical, factual information,” says Arvind. “Refrain from sharing personal, financial, or sensitive data. Use the model as a supplementary tool, not as a primary source of information.”
Key takeaway: AI analyzes guest behavior data in minutes instead of hours, giving you insights to help you staff up or down, create marketing promotions, and more.
Using AI to improve customer support
As the software platform of choice for more than 25,000 businesses across the globe, Zenoti has a support team that stays busy. Zenoti uses AI to address 25% of support tickets, says CTO Anand Arvind. The AI draws from the help document library to resolve those tickets, condensing large files into a short list of bullet points for each support issue along with a link to a detailed help article.
“We are trying to get to the point where every response can be almost instantaneous and contextual to the data,” says Arvind.
The efficiency frees up the support team to solve more complicated issues or requests. In addition, the automated transcription of support calls can help gauge whether customer requirements are being met.
The AI tool helps the support agents, too: they automatically receive personalized reports on their work. The tool can spot active listening skills, even from transcripts, helping agents better understand customer questions and concerns. “It’s an extremely good 1-1 coaching tool,” says Arvind.
Feature image: Zenoti CTO Anand Arvind at Innergize 2023