July 2023 Product News
Train your AI with URLs + A better Product Finder
Welcome! It’s July and hot! Canada is having a record wildfire season, so I hope all who are affected are staying safe. Temps are rising so try to stay cool out there.
We’ve got some great updates this month, so let’s jump right in.
Training your AI with URLs
Since we’ve integrated ChatGPT, we trained the AI to read and scan your Home, Product, and FAQs pages.
Now you can add to the AI’s reading library with additional URLs so it can better understand and answer questions about your products, policies, and anything else you can imagine.
Want to train it on your About page or Shipping Policy? Blog? Go for it.
Have any other content on external URLs like subdomains, community forums, etc.? Feel free to add those, too.
Improved Product Finder
What if finding products was as easy as having a conversation? That’s where we’re going with this. With GPT, we’ve unlocked more cool stuff like AI generated filtering questions.
What does this mean? Well, if you have a big shoe catalog, instead of showing a customer 100 pair of shoes, now the AI will ask questions to filter it down to a few of their preferences—you know, like an in-store sales rep.
It can ask for:
What size shoe do you wear?
What color do you prefer?
Pretty cool, huh?
No more of this
Oh, one more thing. When the AI shows the carousel of products in-chat, shoppers will see the bestselling items first.
Thanks for the feedback
A user requested this, so thanks for that feedback.
More Contextual Recommendations
Another fix courtesy of merchant feedback: recommendations.
If you don’t set up recommendations, then Rep will generate some on your behalf. These include:
Complementary product recommendations
Similar item recommendations
Again you can manually set these up. You know your store better than anyone, but if you need a little assistance, please contact: [email protected]
Let us know how it goes.
Preview Before Launch
Want to see how your concierge responds before you officially launch it? Now you can. This will be the final step for onboarding, right after the “design the look” step.
After you click “preview” you will be able to test 3 different scenarios:
Just click the “Test on your website” button to play around with it. It’s a great way to see how Rep responds before, especially if you’ve made any major changes to your store.
Try before you commit to major changes
ChatGPT 50% Faster
There’s always something to optimize, right? We made GPT’s calls to get info and response times 50% faster.
Because nobody likes to wait, we’re going to keep improving this.
Coming this month, early August
Some huge updates are on their way:
AI Tagging will enable more natural conversational search. What do I mean? Imagine a customer saying the following:
I want a black pair of shoes that are size 10.5, comfortable, and suitable to wear for a dinner party.
This kind of request begins with great tags. Now, if you have a huge catalog, maybe you set up tags already. But maybe you added more products over time, and, well, all the product tags kinda fell by the wayside.
With AI, we can help you fix that. Imagine being able to tag things according to:
Customer rating (from reviews)
Size: Particularly important for fashion, home decor, or any items where size or dimensions are significant.
Material: If certain customers prefer or want to avoid certain materials (like cotton, leather, metal, etc.)
Season: Useful for seasonal items (e.g., summer clothing, Christmas decor).
Occasion: For items associated with specific events or occasions (e.g., wedding gifts, birthday presents, party decorations).
Feature or Benefit: Tagging items based on specific features or benefits can help highlight unique selling points (e.g., eco-friendly, hand-made, energy-efficient).
Trends or Themes: If products are part of a specific trend or theme (e.g., boho-chic decor, vintage fashion).
Functionality or Use Case: How the product is used or its primary function (e.g., indoor, outdoor, travel, workout).
It’s a unique challenge, but it’s a nut we want to crack. Stay tuned.
Remember that black pair of shoes for that special occasion? Here’s 3-4 options that match that criteria.
When shoppers make a very specific request like that, the response should be able to match the level of specificity from the shopper.
With a more natural language exchange takes place between the shopper and your AI, you’ll finally have a reliable, always-on sales concierge attending your store—something you can depend on.
With proper tags in place, Conversational Search will make it faster for shoppers to sift through huge catalogs to find something just for them.
The way it should be.
Train your AI with PDFs
Same as URLs, but with this feature, you’ll be able to upload PDF documents.
How would you use this? Well, if you've spent time fine-tuning your support macros in response to customers’ most frequent questions and concerns, you already understand the value of quick, consistent, and accurate answers.
Now imagine if you could replicate the efficiency and reliability of these macros with AI’s 24/7 availability.
By uploading a PDF of your customer service macros, you're training your AI to learn from all your hard work and offer assistance that perfectly aligns with your support policies and tone of voice.
Coming this month.
To increase the flow of feedback, we’re opening up a Slack group in Rep Universe. Think of it as a direct line to our top-notch development team.
This group is specifically for:
This group is NOT for individual support. Your plan comes with its own support channel so please continue to use this so it’s more manageable on our end. Thank you!
You can access the Slack group with this link:
Hope to see you there.
Here’s to a prosperous July. Stay cool.