Powerful
Powerful
Powerful
Here's how Savantic drives engagement and retention for businesses using AI agents
Here's how Savantic drives engagement and retention for businesses using AI agents
Here's how Savantic drives engagement and retention for businesses using AI agents
Hodos, a flight booking platform like Skyscanner and Kayak, was struggling to meet their sales expectations and retain users. At the time, they would send out a weekly newsletter to their 500,000 email subscribers, simply with a selection of 2 or 3 budget flights that were selected from a vanilla recommendation algorithm, without much supporting content. It's not a surprise their click-through and conversion rate was so low. Their main problem: they put their products first, instead of their users.
They tried Savantic, and within minutes, had a personal AI agent set up for all of their customers. The agent first indexed Hodos' flight database; then, without needing to be prompted, searched and indexed all mainstream and niche travel content on the internet: not just by reading written travel blogs, but also by processing and understanding images, videos, and metadata from social media. Hodos' agent realized the domain and automatically built a proprietary travel knowledge base to make it on par with or better than any existing local guide or travel agent.
With access to its users information, Hodos' AI agent automatically began to introduce itself and communicate with Hodos' users via email to learn about their travel preferences, conversing with them like a travel agent would. No internal staff were needed to assist it. Once it had built a compelling trip idea for a user, it would share it in a beautifully-designed email. This recommendation was personalized to the user's taste and written in Hodos' brand voice; it included rich images and even videos that the agent would produce itself. Every such email came with a relevant Hodos' flight attached, too, which was now more compelling to book. To make this process easy, the AI agent would create a one-click booking link for the flight where the user could purchase it directly in the email itself.
Each interaction didn't stop after just one purchase. Having built a rapport with its users, Hodos' AI agent continues to interact and recommend new trips to them today, and has dramatically increased Hodos' overall sales and retention rate.
Hodos, a flight booking platform like Skyscanner and Kayak, was struggling to meet their sales expectations and retain users. At the time, they would send out a weekly newsletter to their 500,000 email subscribers, simply with a selection of 2 or 3 budget flights that were selected from a vanilla recommendation algorithm, without much supporting content. It's not a surprise their click-through and conversion rate was so low. Their main problem: they put their products first, instead of their users.
They tried Savantic, and within minutes, had a personal AI agent set up for all of their customers. The agent first indexed Hodos' flight database; then, without needing to be prompted, searched and indexed all mainstream and niche travel content on the internet: not just by reading written travel blogs, but also by processing and understanding images, videos, and metadata from social media. Hodos' agent realized the domain and automatically built a proprietary travel knowledge base to make it on par with or better than any existing local guide or travel agent.
With access to its users information, Hodos' AI agent automatically began to introduce itself and communicate with Hodos' users via email to learn about their travel preferences, conversing with them like a travel agent would. No internal staff were needed to assist it. Once it had built a compelling trip idea for a user, it would share it in a beautifully-designed email. This recommendation was personalized to the user's taste and written in Hodos' brand voice; it included rich images and even videos that the agent would produce itself. Every such email came with a relevant Hodos' flight attached, too, which was now more compelling to book. To make this process easy, the AI agent would create a one-click booking link for the flight where the user could purchase it directly in the email itself.
Each interaction didn't stop after just one purchase. Having built a rapport with its users, Hodos' AI agent continues to interact and recommend new trips to them today, and has dramatically increased Hodos' overall sales and retention rate.
From large e-commerce retailers like ASOS or COS to smaller boutique firms, most fashion businesses struggle with traditional e-mail marketing and recommendations. That’s because their depth of understanding on customer preferences and latest trends, and personalization and intrigue of each recommendation is limited.
Savantic allows these merchants to create their own AI agents for their customers, which take the persona of personal stylists and can even interact with their users over SMS! Once setting up Savantic, a merchant defines the persona of the agent (in this case, naming it 'Charlie'), and provides it with examples of its brand voice. Then, just like before, its agent reads and indexes its entire product catalog, able to process both written descriptions and the raw images of the clothing items. It supplements this by scouring the internet, from fashion blogs to Instagram posts, to build a strong thesis on fashion. This process takes minutes.
The agent then communicates to its users via SMS, learning about their fashion sense and consumer desires (through sharing pictures and asking questions). Here, 'Charlie' is helping a user find the perfect outfit for a bachelor party in Las Vegas, just like a personal stylist would. It puts together multiple items from the merchant's product catalog into one outfit, and responds to the user's taste and adjustments to reach a final set of items. It presents this with curated images of models (that it thinks has the same look and physique of the user) wearing all components, and uses many sources of information (including the predicted weather of the user's destination) to create the ideal outfit.
By connecting to the merchants’ product information management software, the agent automatically has an updated overview of what products it should prioritize to sell out inventory. Not only that, but through these customer interactions, it builds a valuable repository of user insights that it shares with the merchant to give them deeper, omniscient signals on customer habits, which they can use to optimize their supply, too.
*images taken from Asos' product catalog
From large e-commerce retailers like ASOS or COS to smaller boutique firms, most fashion businesses struggle with traditional e-mail marketing and recommendations. That’s because their depth of understanding on customer preferences and latest trends, and personalization and intrigue of each recommendation is limited.
Savantic allows these merchants to create their own AI agents for their customers, which take the persona of personal stylists and can even interact with their users over SMS! Once setting up Savantic, a merchant defines the persona of the agent (in this case, naming it 'Charlie'), and provides it with examples of its brand voice. Then, just like before, its agent reads and indexes its entire product catalog, able to process both written descriptions and the raw images of the clothing items. It supplements this by scouring the internet, from fashion blogs to Instagram posts, to build a strong thesis on fashion. This process takes minutes.
The agent then communicates to its users via SMS, learning about their fashion sense and consumer desires (through sharing pictures and asking questions). Here, 'Charlie' is helping a user find the perfect outfit for a bachelor party in Las Vegas, just like a personal stylist would. It puts together multiple items from the merchant's product catalog into one outfit, and responds to the user's taste and adjustments to reach a final set of items. It presents this with curated images of models (that it thinks has the same look and physique of the user) wearing all components, and uses many sources of information (including the predicted weather of the user's destination) to create the ideal outfit.
By connecting to the merchants’ product information management software, the agent automatically has an updated overview of what products it should prioritize to sell out inventory. Not only that, but through these customer interactions, it builds a valuable repository of user insights that it shares with the merchant to give them deeper, omniscient signals on customer habits, which they can use to optimize their supply, too.
*images taken from Asos' product catalog
Hodos, a flight booking platform like Skyscanner and Kayak, was struggling to meet their sales expectations and retain users. At the time, they would send out a weekly newsletter to their 500,000 email subscribers, simply with a selection of 2 or 3 budget flights that were selected from a vanilla recommendation algorithm, without much supporting content. It's not a surprise their click-through and conversion rate was so low. Their main problem: they put their products first, instead of their users.
They tried Savantic, and within minutes, had a personal AI agent set up for all of their customers. The agent first indexed Hodos' flight database; then, without needing to be prompted, searched and indexed all mainstream and niche travel content on the internet: not just by reading written travel blogs, but also by processing and understanding images, videos, and metadata from social media. Hodos' agent realized the domain and automatically built a proprietary travel knowledge base to make it on par with or better than any existing local guide or travel agent.
With access to its users information, Hodos' AI agent automatically began to introduce itself and communicate with Hodos' users via email to learn about their travel preferences, conversing with them like a travel agent would. No internal staff were needed to assist it. Once it had built a compelling trip idea for a user, it would share it in a beautifully-designed email. This recommendation was personalized to the user's taste and written in Hodos' brand voice; it included rich images and even videos that the agent would produce itself. Every such email came with a relevant Hodos' flight attached, too, which was now more compelling to book. To make this process easy, the AI agent would create a one-click booking link for the flight where the user could purchase it directly in the email itself.
Each interaction didn't stop after just one purchase. Having built a rapport with its users, Hodos' AI agent continues to interact and recommend new trips to them today, and has dramatically increased Hodos' overall sales and retention rate.
Hodos, a flight booking platform like Skyscanner and Kayak, was struggling to meet their sales expectations and retain users. At the time, they would send out a weekly newsletter to their 500,000 email subscribers, simply with a selection of 2 or 3 budget flights that were selected from a vanilla recommendation algorithm, without much supporting content. It's not a surprise their click-through and conversion rate was so low. Their main problem: they put their products first, instead of their users.
They tried Savantic, and within minutes, had a personal AI agent set up for all of their customers. The agent first indexed Hodos' flight database; then, without needing to be prompted, searched and indexed all mainstream and niche travel content on the internet: not just by reading written travel blogs, but also by processing and understanding images, videos, and metadata from social media. Hodos' agent realized the domain and automatically built a proprietary travel knowledge base to make it on par with or better than any existing local guide or travel agent.
With access to its users information, Hodos' AI agent automatically began to introduce itself and communicate with Hodos' users via email to learn about their travel preferences, conversing with them like a travel agent would. No internal staff were needed to assist it. Once it had built a compelling trip idea for a user, it would share it in a beautifully-designed email. This recommendation was personalized to the user's taste and written in Hodos' brand voice; it included rich images and even videos that the agent would produce itself. Every such email came with a relevant Hodos' flight attached, too, which was now more compelling to book. To make this process easy, the AI agent would create a one-click booking link for the flight where the user could purchase it directly in the email itself.
Each interaction didn't stop after just one purchase. Having built a rapport with its users, Hodos' AI agent continues to interact and recommend new trips to them today, and has dramatically increased Hodos' overall sales and retention rate.
From large e-commerce retailers like ASOS or COS to smaller boutique firms, most fashion businesses struggle with traditional e-mail marketing and recommendations. That’s because their depth of understanding on customer preferences and latest trends, and personalization and intrigue of each recommendation is limited.
Savantic allows these merchants to create their own AI agents for their customers, which take the persona of personal stylists and can even interact with their users over SMS! Once setting up Savantic, a merchant defines the persona of the agent (in this case, naming it 'Charlie'), and provides it with examples of its brand voice. Then, just like before, its agent reads and indexes its entire product catalog, able to process both written descriptions and the raw images of the clothing items. It supplements this by scouring the internet, from fashion blogs to Instagram posts, to build a strong thesis on fashion. This process takes minutes.
The agent then communicates to its users via SMS, learning about their fashion sense and consumer desires (through sharing pictures and asking questions). Here, 'Charlie' is helping a user find the perfect outfit for a bachelor party in Las Vegas, just like a personal stylist would. It puts together multiple items from the merchant's product catalog into one outfit, and responds to the user's taste and adjustments to reach a final set of items. It presents this with curated images of models (that it thinks has the same look and physique of the user) wearing all components, and uses many sources of information (including the predicted weather of the user's destination) to create the ideal outfit.
By connecting to the merchants’ product information management software, the agent automatically has an updated overview of what products it should prioritize to sell out inventory. Not only that, but through these customer interactions, it builds a valuable repository of user insights that it shares with the merchant to give them deeper, omniscient signals on customer habits, which they can use to optimize their supply, too.
*images taken from Asos' product catalog
From large e-commerce retailers like ASOS or COS to smaller boutique firms, most fashion businesses struggle with traditional e-mail marketing and recommendations. That’s because their depth of understanding on customer preferences and latest trends, and personalization and intrigue of each recommendation is limited.
Savantic allows these merchants to create their own AI agents for their customers, which take the persona of personal stylists and can even interact with their users over SMS! Once setting up Savantic, a merchant defines the persona of the agent (in this case, naming it 'Charlie'), and provides it with examples of its brand voice. Then, just like before, its agent reads and indexes its entire product catalog, able to process both written descriptions and the raw images of the clothing items. It supplements this by scouring the internet, from fashion blogs to Instagram posts, to build a strong thesis on fashion. This process takes minutes.
The agent then communicates to its users via SMS, learning about their fashion sense and consumer desires (through sharing pictures and asking questions). Here, 'Charlie' is helping a user find the perfect outfit for a bachelor party in Las Vegas, just like a personal stylist would. It puts together multiple items from the merchant's product catalog into one outfit, and responds to the user's taste and adjustments to reach a final set of items. It presents this with curated images of models (that it thinks has the same look and physique of the user) wearing all components, and uses many sources of information (including the predicted weather of the user's destination) to create the ideal outfit.
By connecting to the merchants’ product information management software, the agent automatically has an updated overview of what products it should prioritize to sell out inventory. Not only that, but through these customer interactions, it builds a valuable repository of user insights that it shares with the merchant to give them deeper, omniscient signals on customer habits, which they can use to optimize their supply, too.
*images taken from Asos' product catalog
Hodos, a flight booking platform like Skyscanner and Kayak, was struggling to meet their sales expectations and retain users. At the time, they would send out a weekly newsletter to their 500,000 email subscribers, simply with a selection of 2 or 3 budget flights that were selected from a vanilla recommendation algorithm, without much supporting content. It's not a surprise their click-through and conversion rate was so low. Their main problem: they put their products first, instead of their users.
They tried Savantic, and within minutes, had a personal AI agent set up for all of their customers. The agent first indexed Hodos' flight database; then, without needing to be prompted, searched and indexed all mainstream and niche travel content on the internet: not just by reading written travel blogs, but also by processing and understanding images, videos, and metadata from social media. Hodos' agent realized the domain and automatically built a proprietary travel knowledge base to make it on par with or better than any existing local guide or travel agent.
With access to its users information, Hodos' AI agent automatically began to introduce itself and communicate with Hodos' users via email to learn about their travel preferences, conversing with them like a travel agent would. No internal staff were needed to assist it. Once it had built a compelling trip idea for a user, it would share it in a beautifully-designed email. This recommendation was personalized to the user's taste and written in Hodos' brand voice; it included rich images and even videos that the agent would produce itself. Every such email came with a relevant Hodos' flight attached, too, which was now more compelling to book. To make this process easy, the AI agent would create a one-click booking link for the flight where the user could purchase it directly in the email itself.
Each interaction didn't stop after just one purchase. Having built a rapport with its users, Hodos' AI agent continues to interact and recommend new trips to them today, and has dramatically increased Hodos' overall sales and retention rate.
Hodos, a flight booking platform like Skyscanner and Kayak, was struggling to meet their sales expectations and retain users. At the time, they would send out a weekly newsletter to their 500,000 email subscribers, simply with a selection of 2 or 3 budget flights that were selected from a vanilla recommendation algorithm, without much supporting content. It's not a surprise their click-through and conversion rate was so low. Their main problem: they put their products first, instead of their users.
They tried Savantic, and within minutes, had a personal AI agent set up for all of their customers. The agent first indexed Hodos' flight database; then, without needing to be prompted, searched and indexed all mainstream and niche travel content on the internet: not just by reading written travel blogs, but also by processing and understanding images, videos, and metadata from social media. Hodos' agent realized the domain and automatically built a proprietary travel knowledge base to make it on par with or better than any existing local guide or travel agent.
With access to its users information, Hodos' AI agent automatically began to introduce itself and communicate with Hodos' users via email to learn about their travel preferences, conversing with them like a travel agent would. No internal staff were needed to assist it. Once it had built a compelling trip idea for a user, it would share it in a beautifully-designed email. This recommendation was personalized to the user's taste and written in Hodos' brand voice; it included rich images and even videos that the agent would produce itself. Every such email came with a relevant Hodos' flight attached, too, which was now more compelling to book. To make this process easy, the AI agent would create a one-click booking link for the flight where the user could purchase it directly in the email itself.
Each interaction didn't stop after just one purchase. Having built a rapport with its users, Hodos' AI agent continues to interact and recommend new trips to them today, and has dramatically increased Hodos' overall sales and retention rate.
From large e-commerce retailers like ASOS or COS to smaller boutique firms, most fashion businesses struggle with traditional e-mail marketing and recommendations. That’s because their depth of understanding on customer preferences and latest trends, and personalization and intrigue of each recommendation is limited.
Savantic allows these merchants to create their own AI agents for their customers, which take the persona of personal stylists and can even interact with their users over SMS! Once setting up Savantic, a merchant defines the persona of the agent (in this case, naming it 'Charlie'), and provides it with examples of its brand voice. Then, just like before, its agent reads and indexes its entire product catalog, able to process both written descriptions and the raw images of the clothing items. It supplements this by scouring the internet, from fashion blogs to Instagram posts, to build a strong thesis on fashion. This process takes minutes.
The agent then communicates to its users via SMS, learning about their fashion sense and consumer desires (through sharing pictures and asking questions). Here, 'Charlie' is helping a user find the perfect outfit for a bachelor party in Las Vegas, just like a personal stylist would. It puts together multiple items from the merchant's product catalog into one outfit, and responds to the user's taste and adjustments to reach a final set of items. It presents this with curated images of models (that it thinks has the same look and physique of the user) wearing all components, and uses many sources of information (including the predicted weather of the user's destination) to create the ideal outfit.
By connecting to the merchants’ product information management software, the agent automatically has an updated overview of what products it should prioritize to sell out inventory. Not only that, but through these customer interactions, it builds a valuable repository of user insights that it shares with the merchant to give them deeper, omniscient signals on customer habits, which they can use to optimize their supply, too.
*images taken from Asos' product catalog
From large e-commerce retailers like ASOS or COS to smaller boutique firms, most fashion businesses struggle with traditional e-mail marketing and recommendations. That’s because their depth of understanding on customer preferences and latest trends, and personalization and intrigue of each recommendation is limited.
Savantic allows these merchants to create their own AI agents for their customers, which take the persona of personal stylists and can even interact with their users over SMS! Once setting up Savantic, a merchant defines the persona of the agent (in this case, naming it 'Charlie'), and provides it with examples of its brand voice. Then, just like before, its agent reads and indexes its entire product catalog, able to process both written descriptions and the raw images of the clothing items. It supplements this by scouring the internet, from fashion blogs to Instagram posts, to build a strong thesis on fashion. This process takes minutes.
The agent then communicates to its users via SMS, learning about their fashion sense and consumer desires (through sharing pictures and asking questions). Here, 'Charlie' is helping a user find the perfect outfit for a bachelor party in Las Vegas, just like a personal stylist would. It puts together multiple items from the merchant's product catalog into one outfit, and responds to the user's taste and adjustments to reach a final set of items. It presents this with curated images of models (that it thinks has the same look and physique of the user) wearing all components, and uses many sources of information (including the predicted weather of the user's destination) to create the ideal outfit.
By connecting to the merchants’ product information management software, the agent automatically has an updated overview of what products it should prioritize to sell out inventory. Not only that, but through these customer interactions, it builds a valuable repository of user insights that it shares with the merchant to give them deeper, omniscient signals on customer habits, which they can use to optimize their supply, too.
*images taken from Asos' product catalog
© savantic 2023