What do I need to get started?
- Synthesia account: With Synthesia you generate pre-recorded AI videos to be used in the chatbot flow.
- VideoAsk account: VideoAsk is a platform for hosting chatbot-style video widgets. You upload Synthesia generated videos to VideoAsk.
- Alternative - Tolstoy account: Tolstoy is another high-quality platform for chatbots and interactive videos.
Why should I use this type of a chatbot?
Think about this type of chatbot as a modern website guide. It is very useful, especially on mobile, to guide your website visitors to your main conversion points.
This way you can quickly:
- explain what your product/service does through a video
- drive more demos / product signups
- drive people that are not yet interested in a demo towards secondary objectives like webinar registrations, email list subscriptions, etc.
What can't this technology do yet?
This is not a real-time chatbot. It will not, for example, replace the text-based chat on your website (eg. Intercom or Hubspot chat). Instead, think about this more of a website guide. It helps you condense what you do in a short videos and guide your visitors through a flow that makes sense for them.
How do I set it up?
Here are the basic steps. Let's start with the Synthesia-related steps:
- Create your video chatbot structure. You can do it on paper / or in a tool like Miro. Basically just sketch out: "Ok, when people land on my website I want to present them this video. Then I want to give them these 3 options. Then if they click on the first option they should see this video. And so on."
- Prepare the video scripts. Usually there would be 3-6 separate videos for a quick website chatbot guide. Make the scripts as short as possible. They should definitely not be longer than 30 seconds to not bore your visitors.
- Generate videos in Synthesia. Keep in mind that the videos are going to be cut on the sides, because VideoAsk widget shows videos in portrait mode. That's why we recommend the following options in Synthesia: pick an avatar, use the full body option, place it in the centre of the video screen, make it about 75% size (so it's not too big), pick a desired background and that's it. Then just work on your script until you're satisfied with the final result.
- Finish all videos that you need to do in Synthesia. It's a good practice to first finish all the videos in Synthesia and then go work with the flow in VideoAsk.
And then let's switch to VideoAsk. Here are the steps to follow there:
- Go to VideoAsk and create a new videoask. Then choose "Upload a video" option. Upload the video. Optionally add some overlay text. And then decide how people should "answer" after the particular step. We usually use either a Multiple Choice option or a Calendar option (for steps that include booking a meeting). This then creates your first "step" of the chatbot.
- Create more steps in VideoAsk. Each video that you created in Synthesia in step 3 is a new step in VideoAsk. Most steps will end with a Multiple Choice option to navigate between different steps. And the steps at the end of journeys might end with a Button (to redirect people to a website), Calendar (to offer a way to book a meeting) or an Open Ended type (limit it to text input to capture emails for example).
- Embed the video chatbot on your website. In VideoAsk go to Sharing options and select Widget. Type in your CTA text, choose the style of the widget and placement. We usually choose a rectangle style and place it on the left site of the website (as we have chat on the right side of the website). Then click "Get the code" and place the code in your website.
Where can I see some examples?
What about when something changes?
This is the power of combining Synthesia with VideoAsk. It's extremely simple to change a video when you want or potentially add another step. You can literally refresh a video chatbot in minutes with new content.
Simply find the original video in Synthesia, duplicate it and change the text or anything else that you want changed. Re-upload the video to the specific step in VideoAsk and that's it.