A new crop of spice blends with playful names like Arizona Dreaming, Grand Lake Shake and Hell’s Kitchen Vampire Killer are breathing new life into grocery aisles, but while the memorable wordplay might succeed in capturing shoppers’ attention, the non-descriptive names create a problem: How do consumers use these products?
Spiceology, a rapidly growing spice company in Spokane, Wash., has a novel solution – let artificial intelligence help sort it out.
In August, the spice manufacturer launched AI chatbot REMI (Recipe Experimentation and Meal Inspiration) on the Spiceology website to connect users to its database of some 900 recipes and its product catalogue of more than 300 spices and blends.
“AI, when you boil it down, is big-data lookup meets context,” said Eric Anderson, chief marketing officer for Spiceology. “For us, the big-data lookup is this massive recipe database and the big product catalog, and the context is, ‘This is what I like to eat. This is what’s in my fridge. This is what I’m trying to prepare for dinner tonight.’ So, it’s wedding those two ideas together.”
Anderson said the goal is not only to help its retail shoppers, but also to educate customers in Spiceology’s foodservice business.
Spiceology’s creative partner
Spiceology says REMI doesn’t replace human intuition, but enhances it by providing “a creative partner that helps chefs and home cooks alike explore new flavor frontiers with confidence.”
Inquiries typically return recipe options that include Spiceology products, Anderson explained.
“If REMI doesn’t have that, he is allowed to give you ideas, like, ‘That would be great on a salad,’” Anderson said.
For example, customers can provide a simple prompt such as: “I have some leftover chicken that I want to use. I need a simple recipe idea for a dish with a little kick.”
REMI’s response: “If you’re looking to add a little kick to your leftover chicken, why not give it a twist with our Smoky Honey Habanero seasoning? You could make a quick chicken salad by shredding the chicken and mixing it with a touch of mayo, diced celery, and a sprinkle of the smoky, sweet and spicy seasoning. Serve it on a toasted bun or over a bed of greens for a fiery, flavorful meal.”
REMI also offered a second suggestion for spicy quesadillas: “Toss your chicken with the seasoning, add some cheese, and grill it in a tortilla. Serve with a side of sour cream and guacamole, and you’ve got a simple yet satisfying dish with a kick. Enjoy!”
The chatbot is receiving about 30 inquiries a day and providing valuable shopper data for the company, Anderson said.
Spiceology does not sell data or connect it to individual users, he noted. The chatbot never asks for any personal information and is disconnected from customer accounts, which renders all the data anonymous to Spiceology.
“We can see the interactions that the customer has with REMI, so we know what they’ve asked, but we don’t know who the customer is,” he said.
REMI use cases
Anderson says consumers are using the chatbot in two main ways – finding Spiceology products to enhance their meal idea and finding recipes to go with a Spiceology product they’ve already purchased.
Shoppers can ask about how to use Spiceology’s Spanish-inspired Mediterranean blend Tomatador, and REMI will offer multiple recipe options. Users can also iterate on their queries to narrow their focus to diet-specific dishes or those that exclude user allergies, Anderson explained.
Those additional layers of context help the chatbot zero in on what consumers are seeking, he said. “As the thread continues, REMI is understanding your preferences and tailoring them to you,” Anderson said.
The chatbot is built on OpenAI’s popular large language model, ChatGPT, and Pickaxe, a no-code platform for monetizing custom AI tools.
REMI’s responses are limited to Spiceology’s product and recipe database.
“We’re putting something forward that doesn’t scrape the web and steal anyone else’s recipes,” Anderson said.
AI helpers and foodservice education
REMI’s name is a crafty acronym-inspired nod to the chef’s rat assistant, Remy, in the Disney classic Ratatouille.
The chatbot follows the lead of other AI-powered recipe and shopping assistants, such as BuzzFeed Tasty’s Botatouille cooking app, the Meal Reveal app by Hellmann’s and Instacart’s OpenAI-powered chatbot, Ask Instacart.
REMI is similarly a consumer-facing AI assistant, but it also helps the company educate its foodservice customers, according to Anderson.
“As it pertains to REMI, I think we see the most opportunity in foodservice,” he said.
Spiceology sales reps face a steep learning curve when representing the company that offers hundreds of spices and blends, each of which can have dozens of applications, Anderson said.
Accompanying their knowledge with an AI tool like REMI enables Spiceology to work contextually with the foodservice customer’s needs, Anderson said.
While the company’s foodservice business took a hit during the first few years of the COVID pandemic, it has since bounced back due to Spiceology’s focus on the sector, according to Anderson.
“We’ve continued to maintain our base in direct-to-consumer, but have really grown the foodservice channel very consciously and continue to have a strong relationship with the chef community,” Anderson said. “We’ve expanded in distribution, and then also ventured quite a bit more into multi-unit regional chains as our customer base in foodservice.”
He noted that Spiceology’s fastest-growing channel is in retail, with warehouse chain Costco as its biggest customer.
“But we also are in about 600 Targets today and about 2,000 Walmarts,” he said, adding that Spiceology’s presence is growing at regional grocery chains.