What is new?
Last updated May, 2024.
May 24
Twise is now quite a bit faster
Response times are down to around 3 seconds most of the time
That is much faster than most text based website chatbots even
The website widget now has a welcome message
You can configure the message as well as if the browser title should change or not to alert the user that they can talk to the twise
Customise the welcome card
You can change the colors and text for the welcome card (the popup that appears when the user starts the conversation)
That can be under Design > Welcome card
Further design customisation
We've added more design tweaks, including setting header text weight and link fonts and colours.
We've also added font preview in the dropdown boxes.
Go to DESIGN > Theme to change this
Custom voice management
You can now view and add custom voices
You can find this feature under your account (the avatar in the top right corner in the admin)
May 9
Knowledge card creation is better and faster
The creation of knowledge cards from documents, scrape a website, audio or video files is now faster
We've deployed new AI models to deal with this. Please let us know if you see any issues.
Leads are now captured on a separate page & email alert
If you turn on the leads function your twise will ask your visitors if they want to leave their email or phone number
That information is now shown on a separate lead list
Each time a lead comes in it will also be sent to you, as the owner of the twise, via email together with the conversation history
The Call To Action setting has moved to the Lead page
Activating the CTA will show a button at the top of your twise
Clicking the button will either open a form, or an external link
You can activate this setting on the Lead page
Design customisation
You can now change the font sizes and the text alignment. There is a new tutorial for this here
You can also toggle on/off all the other elements on the twise canvas, including the avatar and the hamburger menu. This is helpful when you are using the embed feature or the widget to not have one menu on your site and then another one inside the twise.
The input box, where the user types their message, can be switched to a dark version, which might match better your website design
All these features are found on the DESIGN tab
Open the widget from a custom button
If you use the website widget, where the twise opens in a layer above the page, you can hide the avatar and instead use any button or element on your page to open it
Any link or element on your page can be given a class which will open the widget
Learn more here
Custom domains
We have enabled the option to host your domain which means that your twise can have its own domain
Please reach out to us if you're interested in activating this: hello@twise.ai
Videos and audio can now auto-play
A cool way to make your twise come alive is to have a video as your welcome visual
You can have the video auto-play as well as have it loop
These settings are avilable when you upload a new video
April 30
Design customisation
You can now change the visual aspect of a twise, including changing the fonts and the colours.
This can be found in the admin under the Design tab
Widget that can be embedded on your site
You have seen the widgets on websites everywhere where you click a little avatar or icon down in the right corner to open a chat bot
We have added a similar feature so your users can speak with your twise without leaving the page
Just copy a piece of code and add it to the webpage where you want it to show
Lead capture
You can now have your twise ask your customers for their contact information if/when they show that they are interested. Their contact information will be added to a separate tab.
We will soon build out further functions for how you can get the email sent to you via email as well as have it pushed into other systems, like a CRM.
Content import via XML or CSV
You can bulk upload content into your twise by uploading a list of links, or documents, using CSV and XML formats.
Speed and latency improvements
We have made many smaller changes to reduce the speed and latency.
We have been running promising trials with new language models that will drive down the latency even further although at the moment none of them are good enough to deliver the quality we need.
Last updated