ESS AI 2.0 - Dictation Demo

What's up, y'?

All?

Jim here.

You have to pardon my voice.

I am sick.

as a dog right now.

But, since the unlimited plan for ESS AI is expiring on Monday, March 2nd at midnight, and we've only got a little bit of time left, I wanted to shoot you a quick video to walk you through exactly how I'm using it.

specifically the dictation feature, and just how easy it makes to, you know, word dump, a very kind of raw idea, and get a polished email out.

All right, so, as you can see, I've got a conversation open here.

I've been cranking, for the last, I don't know, 40 minutes or so and pumped, out a couple of emails.

And so I am now ready to move on to the next one.

And so what I'm going to do is I'm going to click on this little, microphone icon and I'm just going to word dump the idea that I have for the email, and then I'm going to let the bot take it from there.

Okay, here is the, story for the next email.

One thing that I have always struggled with as a writer is dictation.

So I just find, or at least have found in the past, that the way my mind works, I just, I can't.

I really struggle to communicate in a verbal form, and I'm just much more comfortable writing things out on a page.

And, you know, it's one of the reasons why I've struggled a lot with public speaking and have been working, you know, a lot on that to.

To get more comfortable doing that, because I feel like it's just two totally different modes of operation for my brain.

And so anyway, yeah, you know, I've known about dictation, tools for a long time and heard lots of people talk about them, but I've always been very resistant to even doing, you know, the slightest bit of experimentation with them.

I've just, I've been like that kind of grouchy old man, you know, sticking to my guns and just, you know, insisting upon typing everything in from scratch.

however, you know, in creating my own software product, ESS AI, one of the sort of unintended, realities that has come with that is that I have to test every single feature, just like, if for no other reason than to make sure it works and to make sure that I know how it works.

And so this has led me to, doing a lot more work with the dictation feature and you know, just as a quick aside, when I initially set it up, you know, it was wired up to a different AI model and the, the speech to text was so bad and it just infuriated me, like it was so crappy and it drove me up the wall and so I had to kind of dig into it and figure out, okay, what are the best models and, you know, what's the best way to kind of get this set up.

And so anyway, now I have it set up properly and it works great.

And sort of unintended byproduct of this is that I've begun to use the dictation feature a lot more and actually I'm really starting to come around on it because I find that once I get going, like, oftentimes that makes it a lot easier.

and the ideas just kind of start to flow and, and I'm able to articulate myself in my own words and you know, that kind of personality and those catchphrases and all the distinctive elements of my voice really, start to shine through.

And so anyway, I'm really coming to enjoy it.

And so, you know, if you're like me and you, also have kind of insisted on being a typer and like, you know, just kind of typing things out, I would strongly encourage you to you know, to play around with the dictation feature specifically because all you have to do is dump the raw, idea in and the bot will take care of the rest.

Right?

It will extract, all of the key details from that word dump into a detailed step by step outline for your approval.

And then after which, you know, once you approve it, if you have to make any changes, you can, but once you approve it, then it will take that outline and turn it into a finished email, which is, which is pretty freaking cool.

And you know, has really, you know, cut down on my email writing time even more and made it possible for me to crank, you know, five, six, seven emails in the span of 60 to 90 minutes.

Which is, which is pretty freaking cool.

Okay, so, as you can see, that was quite a, quite a word dump.

But I find that's often what happens is, you know, initially you don't really think you have a lot to say or you're like, okay, I just have this raw idea.

And then, you know, once you get going, then you just get going, right?

You can see it, you know, it transcribes all this stuff really, really accurately.

So I'm just going to hit enter and again, you can see this is just like me word dumping for whatever, three or four minutes there.

And then what it's going to do is it's going to turn this into a strategic outline for me to approve.

And then once I have approved it, then I just give it the go ahead and it'll turn that into a finished email.

So basically it's going to take the, you know, the bulk of what I gave it and break it out into bullet points.

and then we've got the lesson.

Sometimes the tools you resist the most end up being the ones that, unlock the biggest breakthroughs.

Don't let old habits keep you from experimenting with new ways of working.

Okay, sweet.

Yeah, I love that.

Dumber.

ID and bot.

Yeah.

Yeah, all this looks good.

Okay, this looks good.

Let's go ahead and write the email.

All right, so now the bot's going to take that outline and it's, it's gonna, turn it into a finished email.

So, I wanted to do this just to give you a glance at what, what a RAW output looks like, and then also show you how I like to work through, you know, making edits and like, what my actual workflow looks like.

So, subject line.

I was wrong about dictation.

I've been a typer my whole life.

Lol.

Full disclosure, I've always struggled with dictation as a writer.

The way my mind works, I just can't communicate in verbal form.

I'm much more comfortable writing things that are on a page.

It's one of the reasons I've struggled with public speaking.

I've been working on it a lot because it feels like two totally different modes of operation for my brain.

Anyway.

Known about dictation tools for a long time.

Heard lots of people rave about them, but I've always been very resistant to even the slightest bit of experimentation.

I've been like a grouchy old man sticking to my guns, insisting on typing everything out from scratch.

However, in creating my own software product, Essai, one of the unintended realities.

I have to test every single feature just to make sure it works and to make sure I know how it works.

This has led me to doing a lot more work with the dictation feature.

Quick aside.

When I initially set it up, it was wired to a different AI model and the speech text was so bad it infuriated me.

Lol.

like it was so crappy, it drove me up the wall.

So I had to dig in and figure out the best models and the best way to set it properly.

Now it works great.

And here's the unintended byproduct.

Okay, all right, I'm going to go ahead and take this and paste this into my document.

Actually, I'm going to do the body just at once because what I typically like to do here now you can iterate with the bot.

And, and you know, I think like, that's the way people are really used to doing things with Chat GPT or with Claude is like, hey, no, change this line or change that line.

That can definitely work.

but the way that I've set ESSAI up is that.

So if you prefer.

And again, this is, I built it this way because this is the way I like to work.

I prefer to take the RAW output, like the first draft from the bot.

I take it, I make my own edits in a document and then it's ready to go.

Then I paste it back in and say, look, here are the changes that I made.

And then what the bot will do is identify the differences and update my voice profile accordingly.

All right, so what I'm going to do, I'm just there, there are a couple edits I want to make here.

I just want to kind of polish things up a little bit.

okay, Okay, I'm going to change this.

I'm going to say, in fact, so I had to dig in.

okay.

And wouldn't you know it, all of a sudden I've begun using the dictation feature a lot more and I'm really starting to come around on it because once I get going, the ideas just start to flow.

I'm able to articulate myself, my own words and all those catchphrases and distinctive elements of my voice really shine through.

I'm actually really enjoying it now.

Okay, I have already kind of said that.

I'm going to say moral of the story is, and then we'll bold this.

So if you're like me and you've insisted on being a typer all this time, I strongly encourage you to play around with the duplication feature inside Essai.

All you have to do is press, record and then, and then start talking.

And the bot takes care of the rest.

It extracts all the key details from your word dump into a, and turns it into details sub base bell line for your approval.

so you can make changes if you need to.

And then once you, and then once it's approved, you get to sit back and Watch as the bot turns that in that outline into a finished email that sells, which is pretty freaking cool.

this has cut down my email writing time more, making it possible to crank up five, six, seven emails, seven plus emails in the span of 69 minutes.

okay, not bad, right?

now a friendly reminder.

USSCI unlimited plan expires on Monday.

Which means if you want, I'm gonna all caps this unlimited, avatars, voice profiles and generations.

We'll say instead of E usage, To play, play around with dictation and every other, feature I've built in, you got less than 72 hours to sign up.

And actually this is going out on on Sunday.

So this is in fact less than 48 hours to sign up.

Tap below to sign up now.

Okay, sweet.

So, that is good to go.

Now what's the subject line that it gave us?

I was wrong about dictation.

Okay.

Yeah, I'm, I'm.

Yeah, okay.

Actually I quite like the subject line in this, pre header.

So I'm going to, I'm going to now copy and paste this and the whole thing.

And this is for the purpose of instructing the system on the changes that I've made.

So I'm going to say, great, here is my finished version with edits.

And then I'm going to paste it in.

Now just a quick side note, the system is designed to take the formatting from your document.

If you're doing this in Google Docs like I am, for example, it's designed to take this formatting and convert it into markdown.

So, so that's what these asterisks are.

For example, the two asterisks indicate, bolding.

And then likewise there's one for links as well.

but one thing is sometimes it bugs out a little bit when you paste it in here.

And so there's one missing at the beginning and the end.

So anyway, I'm just going to go ahead and add that back in and then you can see what the bot is doing is updating my voice profile, based on the changes that I made.

And so once it's done that, it will outline the patterns that it's noticed.

And the way that this works under the hood is once, you've created your voice profile initially with the voice cloning bot, it extracts a whole bunch of stuff based on the samples, you know, whatever you give it.

but then there is an additional section in the voice Profile for ongoing learnings.

And these ongoing learnings are what gets updated over time, right?

So as the bot continues to either, passively observe things.

So if you're just iterating with it and going back and forth, even if you don't tell it to update, it will over time update and add to the voice learnings.

however, if you do it like I've just done it right here, where you take your finished version, you paste it back in and you know, say, hey, look, here's my finished version with edits, then it will go ahead and trigger an update that identifies all the different patterns.

And one final note on that is that you can actually see all of the patterns that have been associated and observed with your voice profile, depending on, you know, which, which ones you have, by going to this personalization section in the settings menu and you choose your voice profile, you hit Manage Learnings and then you'll be able to see all the different learnings that the system has picked up and some of these will be pending, some of these can be confirmed.

You can manually confirm some of these learnings if you like.

I wouldn't recommend, you know, hard coding everything, because the way the system is designed to work is it's, is it's based on a confidence score, right?

And so you'll see some of these are 25, 50, and these are based on how many times the patterns have been observed.

And so again, I would encourage you to let the system kind of do its thing in terms of you know, the, the the significant majority of, of your voice patterns will they will get coded in over time just passively without you really having to kind of like come in here and you know, and forcefully approve things.

But again, if there are things you really want to hard code in asap, then you can come in and for example, click this confirm button.

And then what it will do is it will go confirm and then it will shift this confidence score to 60% and, and that will ensure that it gets injected, every time going forward with your voice profile.

So anyway, pretty cool stuff, obviously.

I'm super excited about it.

And you know, as you can see, the end result, especially as you continue to work with it, is that it's able to emulate your voice really, really well.

so anyway, encourage you to check it out and sign up for a $1 trial before the unlimited plan expires on Monday at midnight.