Join 50,000+ readers learning how to use AI in just 5 minutes daily.
Completely free, unsubscribe at any time.
Summary
In the latest Synthesia CEO Livestream, the groundbreaking AI video assistant feature is unveiled. This innovative tool simplifies video creation by allowing users to transform text, PDFs, URLs, or simple ideas into videos without using cameras or professional editing software. The AI assistant efficiently processes content to create engaging videos with options for various inputs, promoting ease of use and creativity. With engaging features and friendly usability, it marks a new era in video production, making it accessible and fun for all.
Highlights
Introducing Synthesia's groundbreaking AI video assistant that transforms text into video magic! 🎥✨
Explore the fun and effortless video creation process with AI technology! 🧙♂️🤖
Learn how AI can generate 80% of your video content, leaving you with just the fun parts! 🌟📝
Meet the three modes: Link, Document, and Idea. Perfect for transforming any content into engaging videos! 💡📄📹
Exciting examples include business expense hacks, dog nutrition insights, and stress management tips! 🐾💼😊
Key Takeaways
Synthesia’s AI assistant revolutionizes video creation by transforming text and ideas into engaging content. 🚀
The AI tool supports multiple input formats, making video production easy for various content types. 🎬
Easily accessible for non-experts, boosting productivity and creativity in video creation. 📈
Overview
Welcome to the exciting world of Synthesia's AI video assistant! Imagine turning your blog posts, PDFs, or even just ideas into videos without lifting a camera. Synthesia’s innovative tool allows you to create engaging videos effortlessly, using clever AI to do the heavy lifting. It can tackle the pesky first draft for you, leaving only the fun part of tweaking the final product.
Ever thought of transforming your company’s blog into lively video content? With accessible tools like the link and document options, you can easily convert existing content. Picture Bob, a marketer turning blog posts into video gems for social media, or a pet shop owner impressively converting care tips into pet-friendly videos. The possibilities are endless!
Lastly, dive into the potential of video prompts for general topics like stress management. The AI doesn’t just convert; it innovates! With user-friendly features for editing inside the video, Synthesia makes sure that creating professional-grade videos is not only for experts. Enjoy the magic of AI with a sprinkle of humor. 🍿
🪄 AI Video Assistant | Synthesia CEO Livestream Transcription
00:00 - 00:30 Welcome back, everyone! Have you ever
sat in front of a blank screen for hours, wishing there'd be a magic wand that
would create the video for you? Well, you are in luck! It's not real magic, but
a clever AI feature from Synthesia we're introducing today. So put on your magic hats and
let's get going, Victor. The stage is all yours. Welcome, everyone! Great to see so many people
here. Super excited to welcome you to the first
00:30 - 01:00 webinar of the year. Today, we're going to
be talking about our AI video assistant, which is something I've been looking forward to
launching for a long time. We've been having this feature internally for the last couple
of months, and it's just truly magical. Before we talk about the feature, though,
for those of you who are new to Synthesia, let me quickly remind everyone
what we do and why we're doing it, and why this feature is important. Our mission
is pretty simple. We want to make video easy for everyone. We want to eliminate the
need for using cameras and studios and Adobe products that are hard to use. We want
to make it into something that everyone can do.
01:00 - 01:30 There are, of course, a lot of different
components around making video easy to do. A few of them are here on the
screen. The avatars, of course, are what we're very well known for. It eliminates
the need for you to film with a camera. You can just generate videos of these actors with AI.
We have lots of templates on our platform, essentially pre-made videos you can
take and edit them, make them your own. Once you've made your video, you can add in
different languages, or you can translate your video into different languages with just one
click. And today, what we're going to be talking
01:30 - 02:00 about is going from that kind of idea to a first
draft, what we also think of as text-to-video. The reason we think that it's important
to make video easy is because when you look at the world today and the digital economy, it's very obvious that people want to watch and
listen to their content. They don't want to read that much anymore. That's very obvious
when you look at people's private lives, where most of us spend more time on podcasts
and YouTube and TikTok than we do on reading. Even though I love reading, I have to
say when I want to learn something new,
02:00 - 02:30 I actually usually go to YouTube and TikTok.
And it makes perfect sense because biologically, we're just hardwired to better understand and
remember visual content. It stimulates more of our senses. It feels more like the experience
we have when we consume information in real life. And that kind of results in us being more
engaged and remembering more of the content. There's some research that shows you remember
around 70 to 80% of what you watch in a video versus 10% of what you read. That's a pretty
big differential. And that's why it's super
02:30 - 03:00 important for anyone who's making content
today, which a lot of you out there are, to also work in video. Writing
blog posts is no longer enough to engage your customers or employees,
whatever you're making the content for. With Synthesia, we're really trying to
increase that productivity. We want this to be a tool that you don't have to be an
expert to use. We want to enable anyone in our company to be able to go in, make
an AI video in just a few minutes, to get to that vision, and to make
it so easy that anyone can do it. We've been working on our AI video assistant
feature for quite a while now. The AI video
03:00 - 03:30 assistant is all about taking you to
that first draft, getting you 70, 80, maybe sometimes even 90% of the way, so that
your job as a content creator is to just edit the draft you get in front of you, as opposed to
having to start from an entirely blank screen. The AI video system works with three
different modalities. You can start from a prompt or an idea that's been live for
a few months in the product. Now you can also use a link to a blog post, for example,
or you can upload a document, such as a PDF,
03:30 - 04:00 to get you started on making videos. And
especially those last two are quite important. We launched this prompt feature a while
back, and we spoke to many of you awesome people using our product. A lot of you told
us that the feature is great, but actually, most of the time when you're making videos,
you're starting from something else. You're starting from a different piece of content
that you want to convert into a video. Again, this could be a blog post, could be a PowerPoint
slide, it could be literally anything else. So with the prompt version of the AI video
assistant, it's kind of hard to take existing
04:00 - 04:30 content and use it. What we've built now is the
link and the URL way to use our video assistant. The first one we're going to go through is
how to use the video assistant with a link. So let's say you're Bob. He's a content
marketer at a software company in Austin, and he wants to take his startup's public-facing
text content, a lot of blog posts, and turn them into videos for use on social media and their
general custom education and training. Louise,
04:30 - 05:00 the CMO, gave him a brief
pitch. Let's watch the pitch, and I'll show you how we solve that
problem with the AI video assistant. "We need more explainer videos for our social
media. Can you take our best-performing blog posts and turn them into short, engaging videos?
Also, add some humor if possible. Thanks." All right, that sounds great. But of course,
it's a pretty daunting task to turn 10 or even hundreds of blog articles into video
content. Luckily, we have the video assistant, and I'm going to show you exactly how we're
going to go about this in just a second.
05:00 - 05:30 So here's Synthesia. Let's click "new video" here.
Let's click the AI video assistant. Let's input the link, which I have open over here. So this is
just a boilerplate blog post-style website. I'll copy the link, put it in here, and then there's
a few things here that I can edit. For example, in the objective here, we can tell the system
what the objective of this video is. In this case,
05:30 - 06:00 it is to educate the viewers on why it's important
to track your business expenses. This is a financial product. We want the tone of voice to be
professional but with a touch of humor. Like this, and the audience here, we'll put in as owners
of small to medium-sized businesses who are not
06:00 - 06:30 financial experts. This will guide it to know that
it shouldn't use a lot of lingo that you only will understand if you're an accountant, for example.
You can also choose the length. I'll stick with that five scenes. Now let's hit "generate
video" and see what the assistant comes up with. Now, as you can see, it's writing the
script. If you read what it's writing here, you'll see that it definitely took
the touch of humor seriously. It's actually writing this in the style of speech
language, not in written language. Of course,
06:30 - 07:00 this is very different. There's also some other
smart things going on under the hood outside of just writing the actual script. What it also
does is it looks at the content for each scene and tries to figure out which of the scenes
in our template matches that content the best. So here it picked a small quote. Of course,
the first one is the intro slide. Here it picked just the avatar talking. Here we got
some bullet points. We got a "thank you" for your presentation, which seems to make sense. Of
course, this probably isn't perfect. So we can go
07:00 - 07:30 in here, continue the editor, and now you can edit
the video like with any other video. In my case, maybe I want to change the avatar. Let's put
Jazz in here instead. Maybe I want to change up the colors. I want to do some script. You
can, of course, do whatever you want here, but it gets you again to that 80% video done. Let's
hit "generate," and we've cheated a little bit. So we'll show you a video with a little bit more
love put into it and what that could look like. "Many of us daydream about our business growing
into this world-conquering powerhouse. But sometimes we neglect the minor steps that
can transform our fantasy into reality. Yes,
07:30 - 08:00 I'm talking about tracking business expenses.
Think of the meticulous tracking of your business expenses as having x-ray vision
for your financial health. It provides a clear picture of funds coming in and going out,
helping you avoid falling into a financial abyss. Now, to the scary stuff – taxes. Navigating
the labyrinth of tax regulations can give
08:00 - 08:30 even the bravest soul palpitations. But
accurate expense tracking is your suit of armor. It eases claiming deductions while
also ensuring you stay within the realms of tax laws. The bottom line – pain-free
compliance and more saved money. Thank you for your time today. If you're looking for
well-managed and transparent expense tracking, please don't hesitate to contact me.
Don't be fooled by its mundane facade,
08:30 - 09:00 for in its depth lies the key to
your business growth and prosperity." Okay, so that's the first modality. The
second modality you can use here is PDFs or documents. Now, the link is great
for content that's already public, a lot of the times that could be things
like marketing materials. But this feature is great for internal-facing content that
may not be available on the public internet. This could be some of your internal processes,
sales trainings, it could be a million different things. The example I have here is, she's a pet
shop owner from Colorado. Every time they send
09:00 - 09:30 out pet food to their customers, they put in a
PDF with some tips and tricks on how to think about nutrition for that particular dog breed.
This is working well. The customers love it. But Nick from customer support had an idea
to make this into something that's easier to digest for their customers, which, of course,
is video. Let's hear what Nick has to say, and then I'll show you how we go
about working from a document. "Those tips for dog owners are
great. Is there a way we can
09:30 - 10:00 turn them into videos and upload to
our website to educate our clients? I get tons of these questions daily,
and this would be really helpful." All right, so Nick has a good point. Let's
turn those PDF documents into videos. Again, we'll go up here, new video, AI video assistant.
I go to the document this time, and I will find my file. So I have a Labrador Retriever file. This is
what it looks like. We'll add that in. Objective here will be to educate dog owners on how to
think about their dog's nutrition. Here, we would
10:00 - 10:30 like the speaker to be a veterinarian. So that's,
again, the tone of voice and the style of someone. In terms of the tone here, let's say informal
and a bit fun. Let's change the template. So for this video, I want something a little bit
different. Let's maybe take this one. I'll pick
10:30 - 11:00 that template, and maybe just to show you here,
when you set the template, you can, of course, see all the different slides that it'll choose
between, right? So we'll use this template, and it will generate the video. And now what
the video is doing is it's reading through that document. It's passing all the information
in it, and then, with the help of our prompt, it's turning that into something that is a
video that matches what we want it to do.
11:00 - 11:30 Again, here, we're writing the scripts.
I selected five scenes this time. Maybe I should have selected a bit more because this
could probably afford a bit longer content. And as you can see here, it's picking all the
different scenes inside of the template, depending on the actual content.
So I'm sure we'll see some bullet points somewhere, a few different bullet
points for some of the text in here. And, of course, the last one is the goodbye
slide. So let's open up the editor again. As with before, you can change anything you
want here. This is essentially the same thing, but again, it gets you to that kind of starting
point that you can then work from. And we're, of
11:30 - 12:00 course, working on also being able to take in all
the images on the website or the document so that this is even better in the first draft version
of it. So let's see what that video looks like. "Let's turn our attention to our lovable
Labradors today, especially their unique dietary needs. These energetic furballs require a
diet rich in proteins, such as chicken, beef, or fish. It's not just about proteins. Our Labs need
a buffet of nutrients – a mix of proteins, fats,
12:00 - 12:30 carbs – to ensure they can outlast a toddler on
a sugar rush without gaining those extra pounds. And let's not forget, Labs come with that
beautiful dense coat. Omega-3 fatty acids are like a spa treatment for their fur. Also, look
for controlled levels of calcium and phosphorus in their food – not too much, not too little,
just right. You might even want to consider dog foods like Hill's Science Diet or Merrick's
Grain-Free Chicken and Sweet Potato Recipe.
12:30 - 13:00 Let's keep those tails wagging, ears flopping,
and bellies full of yummy, nutritious food." The last example is from just a prompt
or from an idea. So this functionality is really good if you're making content
about something that's general by nature. Let's say something like soft skills – how to
handle conflict, how to do stress management. Those kinds of topics are fairly general by
nature. They wouldn't vary too much between different companies, and they're not
personalized to your specific company.
13:00 - 13:30 This is where this feature can come in really
handy. So here we have Rachel. She's an HR lead at a law firm in New York City, and she
wants to generate training videos for these general soft skills. Ben, her manager, gave
her a brief. Let's hear what he has to say. "Hey, Rachel. As discussed yesterday, we need
that stress management training in a video format. We want to share it with the employees
next week. We'll jump back into the studio, and then we'll use the prompt. The prompt here
is the one we call 'idea.' What we want to do
13:30 - 14:00 here is now give it some context about what we
want to create a video about. So let's say here we want to do an internal training video about
stress management for employees of our law firm in the USA. All right, this probably needs to
be a bit longer than five scenes. Maybe we'll give it eight. The tone of voice here should
be professional because this is a law firm.
14:00 - 14:30 The audience would be lawyers who are working
high-stress jobs. We can leave it at that. There's another cool functionality for all
of these, which I haven't shown before, which is the language. You can actually also
give us a prompt or a document in English, and then you can get it out in a different
language if that's what you'd like. But we'll leave that out for now. Again, selecting
a different template here, this one a bit more professional. Generate the video. Now, what it's
doing right is it's not reading through a specific
14:30 - 15:00 URL or a blog post you gave it. It's not reading
through a document that you fed it here. It's actually the general intelligence technology,
GPT-3, that you've probably heard of before, which from this prompt kind of comes up with some
tips and tricks on how to do stress management. That's where this feature really excels if you're
doing these more general types of content. Then this is really good at coming up with kind
of a boilerplate that you can then work from.
15:00 - 15:30 When you see here what it's doing is essentially
coming up with a bunch of different ways you can mitigate stress in a corporate environment. It's
again selecting all the different scenes. And as with all the other modalities, of course, you
can edit all the details of it after it's done generating here. I think this template, for
example, works well for this example. If it's a law firm, then you probably want to have an
avatar that has a suit on. We want to look a little bit more professional. We're, of course,
going to be adding many more templates. And also, one other thing I'll talk about in a second is
how we're going to enable you to do this but
15:30 - 16:00 in your own template with your own styling and
colors. It's actually quite cool. I like this tagline to put under here – 'Stress Management for
Lawyers: Balancing High Stakes and Daily Strain.' That's pretty cool. Let's see what that video
looks like with a little bit more extra love. "Welcome to the Stress Management guide.
It's designed for diligent lawyers like you, managing high-pressure cases alongside
daily legal profession strains. Let's start with identifying the sources of our
stress, which may include heavy case loads,
16:00 - 16:30 demanding clients, or tight deadlines. The
first step to managing stress is acknowledging its presence. Once identifying the sources
of your stress, let's talk about the impact of stress on physical and mental health. When
stress becomes overwhelming, it increases the risk of anxiety and depression. Additionally,
it can lead to sleep issues and muscle tension. Finally, it's time to learn how to
help yourself with stress. First,
16:30 - 17:00 eat healthy. There is a growing amount of evidence
showing how food affects our mood. Second, exercise. Physical exercise can be very effective
in relieving stress. Even going out to get some fresh air and walking to the shops can help.
Third, take time to relax. Striking a balance between responsibility to others and to yourself
is important in reducing stress. And fourth, be mindful. Mindfulness meditation can be practiced
anywhere at any time. It can alleviate the effects of stress and other related issues, such as
insomnia, poor concentration, and low moods."
17:00 - 17:30 All right, so that's the kind of three main
modalities of the feature. Can't wait for you guys to start using it. Tell us what
you think, what can we do better? And we really believe that this is a very important
part of the product, and there's, of course, still lots of things to do. The next step for
us really is learning from all of you what you'd like to see. Some tips and tricks from us: the
first one is the context. So you've seen me do this with all of the videos. Make sure to add in
enough context for the AI system to understand
17:30 - 18:00 what you want. Who's the audience? What's the
objective? It can be tempting to just put in a few lines and let it go from that. And sometimes
that works, but the more context you give it, the better result you're going to get
out of it. It takes it to the next level. Number two, be specific with your prompt. So the
more action language, easy to understand it is, the better it will work. So try and keep it very
down-to-earth and very concrete when you're asking it to do specific things. The last one, which is
actually quite important, is the length. So we saw me generate a few different lengths here, but if
you give us a 15-page or 20-page PDF document and
18:00 - 18:30 you ask us to only make three scenes of that,
then you're going to get a very short summary of what was in those 20 scenes. So in that case,
unless you want a very short summary, of course, you probably want to go to 15 scenes, maybe, to
make sure you capture all that content. And also, the other way around, if you give us half a page
of text and you ask us to make 15 or 18 scenes, the system is never going to build a lot upon
the content that you fed. So trying to find the right balance there is really, really helpful
in terms of getting the result that you want.
18:30 - 19:00 The last thing I want to show you is a related
feature. We also think of this as the video assistant, but this is all about doing it inside
the editor. So not as a step to generate the draft but to actually work with you while you're
making your video. So if we go back into the video we did before here, this is the one for stress
management and lawyers. If I mark my text here, I can preview the audio like I've always been able
to do. But I can also click here to edit with AI. "Edit with AI" is a free-form prompter that can
just help you do whatever you wanted to do. You just type out what you want there. Suggestions
are under here, as you can see – make it longer,
19:00 - 19:30 shorter, change tone, summarize. In this case
here, I feel like this is a bit short for this video. So let's ask it to make this particular
paragraph longer like this. Again, you can see what it's doing here. It's typing with you in
real time, and it should just take this paragraph, put a bit of meat on it, and just make it a little
bit longer, right? So let's say we like this. Yes, that's pretty good. Down here, let's say for
this one, we want to make it a bit longer and tell a joke. So we got a bit of a joke here –
'Why don't scientists ever trust ATMs? Because
19:30 - 20:00 they make up everything.' Let's go with that.
And it can get as wacky as you want it to become, really. Like, let's take this one here, and
let's do something even more crazy. Let's say, write this paragraph in the style of Homer
Simpson. So this is probably not something you're going to be using much, but I think it's a really
fun feature. So let's say "go" and see what this comes up with. "D'oh! When you're dealing with
stress, it's like having a monkey on your back,
20:00 - 20:30 man. Stress can come from all sorts of places –
work, home, maybe even that one time you tried to parallel park and ended up on the sidewalk.
But don't fret, buddy. There's a way out of the stress tunnel, and it's not with a donut,
even though that would be delicious right now."