Speaker 1 (00:00)
Can you hear me okay?

Speaker 2 (00:01)
Yeah, I can hear you. How's it going?

Speaker 1 (00:04)
Good, good. Down goes the window on a rainy day. All right.

Ilan (00:09)
Hey, welcome to Prompt and Circumstance. I'm Ilan

David Vuong (00:13)
And I'm David.

Ilan (00:14)
And today we're talking about AI note takers.

On today's episode, we're going to be talking about 10 AI note takers, and we have a little bonus for you. We'll be putting these all through a suboptimal situation where we need them to take notes for us. We'll play you a couple of audio clips from that,

Then we're going to evaluate the recorders and we're going to give you a recommendation on which ones we think are the best.

Ilan (00:49)
The tools we tested are Grain, Circleback, Sybill, Sembly, Fireflies, tl;dv, Zoom, Notta Granola, and Loom.

David Vuong (01:08)
That's right. Now let's talk a little bit about our method.

Ilan (01:11)
Now is a great time to talk about our first sponsor for this week, Querio.ai

Query is a powerful tool for getting insights into your products. allows you to ask natural language questions and then it will find the answers in your data. It saves hours a week for me, the product team I work with and our data And just for our listeners, they'll give you a two month free trial if you sign up and let them know that David and Ilan sent you. Try it out.

David Vuong (01:41)
Okay, so what we did was I was in a car with the windows rolled down and Ilan was at a busy commuter train station.

lots of background noise. And we recorded a 16 minute conversation talking about OKRs for the podcast. we dropped a bunch of terms into here and we were very curious as to how well the AI tools were picking up on not only what we talked about, some of the terms, but also how they structured the summary.

So we did this in just one attempt. So knowing how AI models behave, that could be a bit of a limitation they sometimes do things, even with the same prompt. we also did it over Zoom using Gmail logins. And so what that means is that if you use Microsoft Teams or Google Meet,

didn't include those in our evaluation, but I think this is fairly representative of what the transcription is able to do. Now, one thing to ⁓ note is that Zoom itself has noise canceling capabilities. And so that certainly helped, I would with the quality of the audio that came through. But with that, let's have a look at the evaluation rubric that we are using today.

All right, so the most important thing, of course, is the summary itself. I mean, this is the main thing that people will be looking at.

So we are giving that 10 points. Next is going to be accuracy. So this is where you look at the transcript because sometimes you need those specific details. So we are looking at whether or not that transcript is accurate or whether there's any issues or challenges with that.

And for accuracy, we are going to give five points.

Next is the free trial and how generous the product is with the free trial. So we want to make sure that if you try out these tools that you have a good shot and we want to include that in the evaluation of these tools. It's not going to be weighed very heavily, just two points, but still meaningful.

Ilan (03:51)
In terms of overall scoring, one of the things that I focused on was I preferred for free trials a count of meetings rather than a number of days. So I tended to give those higher scores

David Vuong (03:51)
Edit.

Lastly, of course, is security and privacy, namely SOC 2 compliance, GDPR compliance, and HIPAA compliance. So each of those giving one point for a total of three points in security and privacy. if you add all of that up, we are going to be at 20 points total. That's up for grabs for each of these tools.

Ilan (04:23)
so David, you shared your notes beforehand. I didn't share mine. I didn't check yours either. So we're both going into this blind to what each other thought about these 10 tools.

David Vuong (04:33)
Yeah, I'm curious whether you came to the same conclusion that I did about how surprisingly one of these performers are.

Ilan (04:41)
We'll get started with the tools. The order is determined by how quickly they gave us the notes for the meeting.

David Vuong (04:48)
That's right. So the first tool, Grain, was the fastest.

Ilan (04:52)
So I gave Grain a 5 for the summary. How about you?

David Vuong (04:58)
I gave it a five out of 10 as well. There were some discussions that we had about the OKRs that didn't get included into the, relevant section for the objective. It also failed to ⁓ group the KRs under the O's.

Speaker 1 (05:12)
So I'm thinking an objective, it would be great for us go through with a full-on

maybe the objective could be actually launch some

Speaker 2 (05:26)
I'm thinking objective number two is tremendous to

prompt and Circumstance listeners

Last time I checked, 15 subscribers, I think we should be able to 5x in the next quarter.

Ilan (05:42)
Yeah, I also mentioned here that the summary was terribly formatted.

David Vuong (05:47)
we also had ⁓ an action to reconnect in two days and it didn't even include that.

Speaker 1 (05:52)
Sounds good. All right, and we'll reconnect in two days.

Ilan (05:57)
So much for getting your action items from the recorder on accuracy. I gave it a two. How about you?

David Vuong (06:00)
Yeah.

I gave it a four out of five. I thought that it had some inaccuracies where, for example, we said DAU as in daily active user and it mistranscribed that. And also the transcript itself was in very small font when you exported it.

Ilan (06:18)
How about on the free tier?

David Vuong (06:20)
On the free tier, I thought it was perfectly fine. I gave it a two out of two. The free tier gives you 20 meetings, which is plenty generous to evaluate.

Ilan (06:30)
same here. And on compliance it checked two of the three boxes that we were looking for. So my total here was 11 points. How about you?

David Vuong (06:38)
And,

and mine was 13.

Ilan (06:41)
All right. The next fastest tool was Circleback. So how did you do there?

David Vuong (06:46)
I think Circleback gave one of the best summaries out of all these tools we evaluated. It did a great job of grouping the KRs under the O's. And in fact, even labeled the KRs properly with the numbers and everything with key figures bolded, action items upfront, excellent work. ⁓ The only thing that kept it from full marks was it would have been nice to have timestamped links to show, okay, this is where these topics were decided or discussed.

Ilan (07:15)
Yeah, I also gave it a pretty high score. I gave it a seven, which was the second highest score of all the tools that we looked at. On accuracy, I gave it a four out of five. Pretty happy with the accuracy of the notes. How about you?

David Vuong (07:29)
⁓ I also give it a four out of five. And, ⁓ what I found actually was that it misattributed some speech to the wrong speaker. So sometimes when we were talking and we overlapped a little bit, it didn't identify that it was you who was talking and not me.

Ilan (07:44)
The thing I did like there was that

grouped together, parts of the conversation into a single point as opposed to breaking it up by timestamp, which I found difficult to go through with some of the other tools.

David Vuong (07:58)
That makes sense. In terms of the free trial and the generosity there, I found seven days to be very limited and also that a credit card was required in order to start the trial. I didn't like that very much. So I only give it one out of two.

Ilan (08:14)
Yeah, same here. And on compliance, it also checked two of the three boxes. My total for Circleback was 14. How about you?

David Vuong (08:21)
I gave it a total of 16. What's interesting about Circleback is it has this neat little action items section where it'll pick up on the action items that you've got from your conversations. this way you don't need to review all those summaries. It's a summary of the summaries, I

Ilan (08:38)
Moving on to Sibyll, which was the next fastest

Sibyll, I gave them a 5 for summary. I thought it was lacking. My notes are also lacking here, so I'm curious to hear what you thought.

David Vuong (08:55)
I gave it an 8 out of 10. I thought that it nailed the OKRs with grouping. mean, the formatting could be a little bit better. It would have been nice to see also a list of attendees. You'll notice that some of these might or might not have that. The problem I had with it was that the export of the summary was text only. You could only export a TXT file. There was no rich text export available. What about the accuracy?

Ilan (09:23)
I gave it a five out of five on accuracy. I thought it was the most accurate of all of the tools. How about you?

David Vuong (09:29)
I actually gave it a three out of five because instead of saying that your speech was associated with Ilan, it was associated with Ilan's Loom note taker. I'm pretty sure your note taker was not the one that was talking.

Ilan (09:43)
Fair enough though that was weird the Actual text that it produced and the the flow to be the most Easy to follow out of all of them easy to find the sections of the

In terms of the free trial,

they give 14 days. also gave it a one out of two.

David Vuong (10:05)
I felt 14 days was okay. I gave it two out of two. I think two weeks is enough. It's a sprint. Come on.

Ilan (10:12)
You

David Vuong (10:12)
Now in terms of security and privacy, it met two out of the three. So that's two points there. And for me, that totals 15 out of 20 points.

Ilan (10:21)
and I got 13 out of 20 points.

David Vuong (10:24)
What's interesting about Sybill is that they have this ⁓ email follow-up feature, right? So you hit a button that's related to your meeting and you can just generate a follow-up email. It's a sales oriented tool, which some of these are and some of them aren't. But I thought that was a really neat feature.

Ilan (10:41)
Next up in terms of speed was Sembly And on the summary, I gave Sembly a 6. How about you?

David Vuong (10:47)
I gave them a seven out of 10. They captured the objectives well, ⁓ but they didn't quite group them together with the KRs too nicely.

It would have been nice to see action items split out in the summary. They have this tasks function, which is separate from the summary and you get tasks assigned to you. But it would have been nice to see it in the actual summary

Ilan (11:11)
The format of the summary was very much project focused. I felt like this would be the best tool for a project manager

about on accuracy?

David Vuong (11:19)
I give it a 4 out of 5. It overall got it okay. There were some parts of speech where again with the overlap it misattributed the speech to the wrong person. How about you?

Ilan (11:31)
I gave it a three, I noticed the same thing so I just docked it harder for those.

David Vuong (11:35)
You are a tough grader.

Ilan (11:40)
I think we're gonna end up seeing that across the board. I was pretty tired when I reviewed these, so maybe I was grumpy. ⁓

David Vuong (11:45)
Yeah, the angry evaluation. Well, in terms of

the free trial, I thought that they were plenty generous. So this is one where it's not time limited and they just simply have limits that reset each month. I thought that that was super generous. I would have loved to give them more than two out of two. But of course they maxed out for me. How about you?

Ilan (12:05)
Yeah, I felt the And on compliance, they're the first tool out of these that checks all three boxes.

a total of 14 points for my scores. You?

David Vuong (12:15)
and it's a 16 for me. What's interesting, if you look at the video recording, and this is not just for Sembly but for some of these other tools, the video recording is actually a screen recording of Zoom. So I thought that that was really interesting.

Alright, so next is Fireflies. And for the summary, I gave it an 8 out of 10.

I felt that the OKR grouping, ⁓ was done very well, ⁓ where it numbered the, objectives and also the, the key results. So that was quite nice. ⁓ in fact, in the summary, it even spelled out DAU properly. ⁓ so that was, ⁓ pleasant to see. It did miss the action to regroup in two days. So therefore it did not get full marks. What was your take on it?

Ilan (13:04)
I also gave it a pretty high score for me. I gave it a six out of 10. I liked as well how it structured the objectives and key results in the summary

David Vuong (13:14)
I found that the, ⁓ in terms of accuracy, the transcript had some misspellings, but overall it was okay for me, so I gave it 4 out of 5.

Ilan (13:24)
Yep, felt the same, gave it the same score.

David Vuong (13:26)
Now, in terms of the free trial, it's a little bit brutal here. I mean, with a seven-day limit and requiring a credit card, that was ⁓ not too generous. And so I gave it only one out of two.

Ilan (13:39)
Yeah, I gave it the same score.

However, on compliance, it did meet all three of our compliance requirements. So it got a three out of three there for a total of 14 points on my scorecard. How about you?

David Vuong (13:50)
And on mine, Fireflies got 16 out of 20. Now what's interesting about Fireflies is that they have a button that allows it to join a live meeting. So if you have a meeting that's going on, maybe that was impromptu, you could have a join. Hey, here's the, here's the link and it'll, it'll join. What I found really frustrating about using Fireflies though, is that I have multiple Gmail accounts and ⁓ it would not let me choose which Gmail accounts to use. It would insist on using just the first one that

I've logged into. And so I had to open a private browser tab in order to log into using the account that I wanted to. So that was kind of frustrating.

Ilan (14:30)
Next up is tl;dv and very curious to hear what you thought of this one. For the summary, this got the second lowest score out of all of them for me. It was a three for me. How about you?

David Vuong (14:45)
I gave it a three as well. Three out of 10. This got the worst score. It tied for worst for me and it missed several key results. It was a very brief summary. The action assignment was incorrect. I mean, that's brutal. So that's a, that's a really low score.

Ilan (15:03)
Then on accuracy, I gave it a one. I did not see much accuracy out of the tl;dv. Summary, how about you?

David Vuong (15:12)
It misattributed the speakers and it did not give any timestamps. Now I was a bit more generous. gave it a three out of five.

Ilan (15:21)
On the free tier, it's free forever, so I got a 2. ⁓ I know there's an upgrade to pro option there, but ⁓ that's where it got top marks for me. How about you?

David Vuong (15:32)
Yeah, I mean it's not time limited, but it is a summary limited ⁓ trial. So it maxes out at 10 summaries, but that's okay. I think that's fair for a free trial. So two out of two.

Ilan (15:45)
On compliance, I actually gave it a 1.5 because on SOC 2 compliance, it only has type 1 compliance. So I gave it a 0.5 for that and that gave me a total score of 7.5.

David Vuong (15:57)
And for me, I gave it a one. It's either you have type two or you so in total for me, it was a nine out of 20. Now this is a sales oriented tool. What's, also frustrating about it is that there was no export capability for the transcript. ⁓ and, even if you want to copy the text, it has this paywall where they tried to upsell you the ability to copy more than five minutes worth of text. That was silly. ⁓ what is interesting about tl;dv

is that it has this multi-meeting capability that can summarize across multiple meetings. I didn't see that in any of these other tools, but I thought that was really interesting because it's possible to have conversations that extend across multiple meetings and you want to summarize across all of them.

Ilan (16:42)
For me what's interesting is every time I see tl;dv all I can see is too long David Vuong.

David Vuong (16:48)
What are you trying to say?

David Vuong (16:52)
All right. Next is Zoom which is interestingly the, you know, the video conferencing tool that we used. And I was very surprised to see not only how late it came with the notes, also, poor the summary was. I gave it a three out of 10 for the quality of the summary. It was not able to mark objectives as sections.

It didn't put the key results as children of the objectives and it just overall was not a very good quality summary.

Ilan (17:26)
Now it's interesting you say that I gave Zoom a five, so not the greatest score, but of the things I downgraded some of these tools on was wall of text

What about accuracy?

David Vuong (17:42)
The, the transcript accuracy was also poor. And I think this ⁓ led to, for me, at least from my perspective, a poor summary. I gave it a two and some examples of bad transcriptions were when you said 10 DAUs it reinterpreted that as probably 10 thou. So it set our goal as instead of 10 daily active users to be 10,000.

Ilan (18:07)
the accuracy, I gave it a three. One of the things I noted on Zoom and a couple of other tools breaking out into very think that sometimes led to lower accuracy on the summaries as well.

it couldn't sort of group together conceptual ideas and then that led to a poor summary.

David Vuong (18:35)
In terms of the free trial, get days, ⁓ and that's with the whole transcription, ⁓ add on, suppose the transcription and AI summary add on. So 14 days. It's all right. ⁓ but you need a credit card for that. So I docked it, ⁓ marks. can't get full marks if you need a credit card. So I gave it a one.

Ilan (18:56)
I also gave it a one, just being consistent. I gave all of these kind of time bound seven or 14 days a one. So Zoom got that as

David Vuong (19:07)
trying to cancel that premium subscription, by the way, on Zoom was, it was extra difficult ⁓ because the payment is through Google play. And it, didn't know which, which account I had used and that would create all sorts of frustrations, but on the security and privacy side of things, it got full marks. I mean, it's Zoom it's very well established. So, ⁓ three out of three.

Ilan (19:13)
Ha

Right on, so how did that total out for you?

David Vuong (19:36)
total that comes to nine out of 20. So that's a failing grade for me.

Ilan (19:42)
It came out to 12 out of 20 for me, so still not top marks.

David Vuong (19:48)
Something to note for Zoom is that the transcript, when you download it, it comes out as a VTT file, which is meant for machines to read, as opposed to somebody to open it up in, say, a Word document. The other thing is that if you want to use the summary, if you want to export that, too bad, because they are forcing you to use Zoom docs.

Ilan (20:10)
You

David Vuong (20:13)
I don't know if you know what that is, but they're trying to take over office software, I suppose. And, ⁓ it has some really dark UX patterns, ⁓ that prevent you from escaping the Zoom ecosystem.

Ilan (20:27)
doesn't sound like the best experience trying to understand why did we decide in this meeting and who said that?

David Vuong (20:33)
Exactly.

Ilan (20:34)
Next up we're on Notta Notta I thought the summary was also kind of I gave it a five on the summary.

David Vuong (20:45)
I gave it a six and what I noticed here is that it was a lengthy summary. It gave a four page summary for a 16 minute conversation, which I think is not really much of a summary at all. It, it missed some things like the reconnection in two days. It didn't put the KRs succinctly as children for the objectives. So not, not top marks here.

Ilan (21:15)
On the accuracy side, well you shared the transcript of this and it said can't download, can't copy, so...

David Vuong (21:23)
Yes,

that's right. Yeah. So, ⁓ you cannot download nor copy the transcript. I gave it a one overall. not only out of frustration for that, but also when you do look at that on the website, it, it misattributed some things to the wrong speaker. It, incorrectly transcribed some things, ⁓ in one funny example, ⁓ you told me.

Ilan (21:29)
He

David Vuong (21:52)
because I was driving, you said, don't get into an accident. And the transcription missed the don't.

Ilan (21:59)
I had, Notta had some evil intents for me.

David Vuong (22:01)
So according to a note that you told me to get into an accident. Yeah.

Yeah. There you go. Yeah. See, I have it on record that, that you were jinx- jinxing me. Yeah.

Ilan (22:11)
That's right.

Speaker 2 (22:12)
Okay, I'm gonna have to, like, put it off-screen.

Speaker 1 (22:14)
Don't get into an accident.

Speaker 2 (22:15)
I'm like so focused.

Speaker 1 (22:16)
right, let me make sure that you end your destination

David Vuong (22:20)
in terms of the free tier, it's quite generous because the balances will reset month over month. Uh, however, I, uh, still gave it a one out of two because a credit card is required.

Ilan (22:33)
Yeah, you were consistent on that. The credit card requirement didn't bother me, but I did like the resetting balances each month, so I gave it a 2.

David Vuong (22:43)
The, ⁓ security and privacy side of things three out of three. ⁓ so that's quite nice. They have this mind mapping feature, ⁓ which is, ⁓ novel. We didn't see that in any of the other tools, but, ⁓ it leaves a lot to be desired. The mind map for our conversation was, ⁓ pretty much unreadable. It was a, it was only one level deep and it was very tall, just not really usable at all. So good idea, but, ⁓ execution.

leave some things to be desired.

Ilan (23:14)
Agreed the mind map, I found it unreadable. It's just a reminder to product teams. If you want to differentiate your product, make sure that you're thinking through what are the features that are good to differentiate you from your competitors.

David Vuong (23:28)
Yeah. They also have a summary templates, uh, which some of these other tools do have, but when I, uh, what that does is it'll, uh, tell the AI that summarizes the transcript, what kind of structure to give it. However, I couldn't find any templates that matched our conversation. So anyway, overall, it was 11 out of 20 for me.

Ilan (23:51)
Overall I gave it a 10. That's the difference on accuracy I think between our scores

David Vuong (23:57)
Makes sense. Also of note is that, Notta, when you do the onboarding, ⁓ they ask you what kind of role you're in and they grouped project manager with product manager. So I'll let the audience decide whether that's, ⁓ that's okay.

Ilan (24:12)
Up next we had Granola Granola had one of the best summaries in my opinion I really liked how succinct it was how well grouped all of the information was I gave Granola an 8 on the summary and that was top marks for me on the the summary

David Vuong (24:31)
So top marks is eight out of 10. What happens with the last two points?

Ilan (24:36)
I didn't

think anyone got a 10 out of 10.

David Vuong (24:39)
Okay.

Well, I gave it nine out of 10, which is also my near maximum. So it really nailed the structure. The objective headers were bang on key results nested as children. It nailed all of the action items. The only reason why I didn't give it a 10 is because it would have been nice to see an attendees list.

Ilan (25:02)
Hmm actually brings us to the accuracy which I gave it a one for the transcription accuracy thing with Granola is that it runs on your desktop It doesn't join the meeting and so it doesn't actually know who's in the meeting probably why it doesn't give the attendee list and so that the transcript just says you and them and

David Vuong (25:20)
see.

It's us versus them.

Ilan (25:29)
That's right, and it just gave it also gave just a Replay on the transcript as opposed to grouping together or common sentences that the person was was giving the copy of the transcript was pretty bad But even in their in the UI where it's more readable. It's still it wasn't

David Vuong (25:53)
Yeah, I gave it a two and for the same reasons, it was a bit of a jumbled mess. ⁓ there were no timestamps, from what I saw. So not too good.

Ilan (26:03)
Yeah, on the free tier, Granola has a pretty generous free tier. Also one that resets and has a pretty broad range of 25 meetings recorded. I gave it a two on the free tier. How about you?

David Vuong (26:17)
Yeah,

same thing. ⁓ know, 25 meetings is plenty generous for a free tier. mean, you know, think about what, what a free tier is for. It's for you to try it out before you buy. And if somebody can't decide after 25 meetings, then maybe they're not ⁓ an actual buyer.

Ilan (26:34)
That's right. And finally on compliance, check two of the three boxes that we had. And for me, that gave it a total of 13.

David Vuong (26:44)
It was a total of 15 for me. what I also liked about Granola is that non-subscribers can sub can chat with the transcript. So in the, ⁓ on the Granola website, you know, I was not the one who had the free trial and I was still able to ask things of the transcript. So that was quite nice.

Ilan (27:04)
Yeah, it's a cool feature, the chatting with the transcript. I like that you can ask, hey, who said this? Can you tell me more about that? I also use it sometimes in interviews. When I'm interviewing somebody and I'll chat with it afterwards, like, how did this person answer this question again? And just bring up that part of the conversation.

David Vuong (27:26)
You could, you could ask it act as a hiring manager. Would you hire this person and just call it a day.

Ilan (27:33)
finally on the AI note-takers who were in the meeting we have Loom Loom their summary Left a lot to be Desired for me. I gave it a six out of ten. it wasn't the best. How about you?

David Vuong (27:48)
I gave it a seven. so it got the objectives as a proper sections. ⁓ but the key results were written in prose rather than, you know, punchy bullets with, which would have been ideal. ⁓ to me, what also mattered was that the action assignments were incorrect. So somebody was assigned the task to verify current metrics and it assigned it to the wrong person. So that, that's, that's no, no good.

Speaker 2 (28:14)
So action number one, write down our OKRs. Then

I think across both these, ⁓ another action we can take is begin creating shorts for each

Speaker 1 (28:27)
so who's going do that? I guess you can do that.

Speaker 2 (28:30)
What's another action item you're thinking?

Speaker 1 (28:32)
let's get concrete with our numbers and so I can do that.

Ilan (28:35)
On accuracy, I was surprised because Loom is such a market leader. But the accuracy of the transcript, I ended up giving it a two. And there's an interesting thing that you pointed out when we were chatting about this, a little artifact that showed the transcription provided by some random And I looked into it.

David Vuong (29:00)
a beat.

Ilan (29:00)
It turns

out this is an artifact that turns up in their speech to text transcription so not great when Loom isn't even cleaning up and just dumping whatever information they're getting from OpenAI.

David Vuong (29:14)
Yeah, that's a, that's really interesting. I noticed some of the transcriptions were really strange. I gave it a two. ⁓ so there were lots of mistranscriptions. ⁓ there was this one section where it had a question mark and a quotation mark just repeated three times. And it's like, how, how does one even transcribe that? So not too good there. I was also very shocked again, given, ⁓ like you said, the market presence of Loom.

Ilan (29:30)
You

David Vuong (29:40)
In terms of the free trial, ⁓ I gave it a two, ⁓ that was kind of rounding up. mean, you know, five meetings as the free trial. It's okay. ⁓ you know, it's not the best, but it's not time limited either.

Ilan (29:40)
Other

I felt the same. I also gave it a 2. It met my criteria of not being And on compliance, I checked 2 of the 3 boxes as well. For me, that was a total of 12. How about you?

David Vuong (30:07)
Yep. 13 total for me.

Ilan (30:10)
Also interesting about Loom is they acquired a company called Rewatch that was a meeting recorder and so they didn't have meeting recording capabilities until they acquired Rewatch. And as a Rewatch user I can say that the quality of the summaries went down significantly after the acquisition. I wonder if they changed the prompt.

David Vuong (30:33)
Yeah, the system prompt. I, you know, I wonder, ⁓ I was wondering about Loom because to me, ⁓ I know them as a short video recording tool, right? You send out a Loom video to somebody as opposed to it being a note taker.

Ilan (30:49)
Yeah, absolutely. I mean, at some point you gotta expand your market,

Ilan (30:54)
Finally, we have a bonus. David, you want to tell us about the bonus?

David Vuong (30:58)
That's right. So the bonus is that there's this site called Rev.

And Rev offers ⁓ free transcriptions. You just upload audio and it'll have a transcription service for you that'll transcribe what you've got. ⁓ Now we had a look at the transcription together. I I gave it a three out of five for accuracy. It didn't generate a summary. So just looking at the, at the accuracy of it. What was really interesting is that it, ⁓ it had flagged the human emotions.

around the conversation. So it would have laugh, for example, in, in brackets. And, ⁓ it would, when somebody said, would have it as like affirm affirmation, ⁓ in brackets. So that was quite unique. ⁓ the interesting thing about Rev as well is that it offers human transcription as a service. So if you wanted to insist on having humans transcribe something, they have a 24 hour turnaround service. So that's.

That's really different for me.

Ilan (32:01)
Wow, yeah, that's really cool. I gave it a 2 on accuracy. I also noted the human emotions and the laughing. But overall, I felt that the transcript was just it's free, so hey, got a 2 out of 2 there.

David Vuong (32:18)
That's right.

Ilan (32:19)
well David, we've gone through all of the tools now, we've given our scores. I think all that's left is to put these into tiers so we can let people know what do we recommend for them.

David Vuong (32:33)
All right. Now, before we get into it, just important to note that for some of these tools, they meet all three of the security and privacy criteria that we had set out. The SOC 2 compliance, the GDPR compliance and HIPAA compliance. And for those where these certifications are important, here is the list of the tools that we tested. It's Fireflies, and Zoom.

Ilan (33:01)
David and I are setting up our tiers, so it's a great time for me to tell you our other sponsor for this Colab.

Colab is a product that will help you validate and

a side is great for product managers or any professionals who are looking for a way set up their side business but don't really have enough do the initial market testing and building.

They'll hand you a working product and they'll even find your first paying customers.

click on the link below and let them know that David and Ilan sent you for $250 off Validate and Build.

Ilan (33:34)
Let's start from the bottom and work our way up. In my D tier, I have TL David Vuong. How about you?

David Vuong (33:41)
I've got a lot more. I've got tl;dv. I've got Grain. I've got Notta or Notta I've also got

Ilan (33:50)
So

moving on to my C tier, I had Notta and Grain.

David Vuong (33:55)
In my C tier I've got just Loom.

Ilan (33:59)
So that brings me to my B tier where I have both Zoom and

David Vuong (34:04)
Yeah, in ⁓ B tier for me was Sybill.

Ilan (34:08)
Well that brings me to my A tier. are all very good tools in my opinion. I have Circleback, have Sibyll, and I have Granola. Sibyll also sitting just above. You know, I was a hard marker, but you're a hard tier, David.

David Vuong (34:25)
I'm a hard tierer. You're a very generous tierer.

Circleback for me was also in the A tier because of ⁓ its higher cost. It is about double that of the other tools. So that landed it in the A tier. Also in the A tier for me were Granola and Sembly.

Ilan (34:45)
that leaves off with the S tier. In the S tier, that's We've got and Fireflies for me.

David Vuong (34:49)
Yeah, the S tier.

I only had one. I had Fireflies in S tier. So although it didn't score as high, it got an 8 out of 10 for the summary. Overall, it did very well. And like you said, it got all of the security privacy ones checked off. And yeah, with a $10 a month cost, it's hard to beat.

Ilan (35:20)
Well there you have it. Those are the 10 AI note takers. Did we miss any? You should let us know in the comments or send us a message.

Otherwise, like and subscribe, leave us a general note, and we'll see you next week.

David Vuong (35:34)
See you at the next one.