uSpeak Podcast: Why Human Review Remains Essential in AI Transcription

Speaker B [00:00:01]:
Hi everyone, this is Minnie with the latest episode of our podcast, you Speak. Today's episode is especially important for anyone interested in working from home, building a flexible income, or starting a career in transcription. There is no question that AI is changing the way we work. And for many people, that has created uncertainty. So especially in industries like transcription, a lot of aspiring transcriptionists are asking, is there still room for humans in this industry? Can I still build a real career doing this? And today we are going to talk honestly about that. Joining us today is Suyash Srikant, the co founder of Dicta AI, a company that combines AI technology with human transcriptionist to create high quality transcription services. We will discuss how AI is changing the transcription industry, why human transcriptionists remain valuable, which skills matter most today, and how people, especially work from home, moms and beginners, can position themselves for success in this evolving field. So if you have been curious about transcription, nervous about AI, or wondering where the industry is headed, this conversation is for you. Let me welcome Suyash. Very warm welcome to you. How are you?

Speaker A [00:01:30]:
Hi. I'm good, I'm good. How are you?

Speaker B [00:01:33]:
I'm pretty well. I'm doing well. So are you ready to.

Speaker A [00:01:37]:
Yeah, let's get started.

Speaker B [00:01:39]:
Okay, let's get started. So what made you personally want to build Dicta AI and what problem were you trying to solve in the transcription industry?

Speaker A [00:01:51]:
Dicta AI is essentially an effort to revitalize and expand the transcription industry and create a single umbrella where every possible need of transcription is addressed, where those who have already been in the industry for a really long time find a space. Even though AI already, even though AI has come and changed much of how transcription is done, we are working towards creating a balance where the human input and skills are not overshadowed by AI completely and where the AI lacks due to hallucinations and due to certain flaws in the way they are trained, is addressed by the human touch. That's what we're trying to do here.

Speaker B [00:02:59]:
Okay, so it's like a best combination of AI and human.

Speaker A [00:03:04]:
That's what we're trying to do here.

Speaker B [00:03:05]:
Okay, great. Okay, so AI is changing many industries today. From your experience at Dicta AI, how is it changing the transcription industry?

Speaker A [00:03:17]:
The existence of AI in transcription has made the industry grow in many folds. Because today almost every call, every professional call is recorded and transcribed. This was not the case a couple of years ago. The cost of transcription was so high that it wasn't even imaginable to do something like this today. Every business meeting, every consultation, every call is recorded and then it's transcribed by artificial intelligence. What this does is it creates a lot more volume of work for the industry and allows people to, allows a trickle down effect when there's a lot more transcription happening. A lot more human transcription is also happening because just to get it wetted to make sure the transcription is perfect, has 100% accurate. In industries where it's required, such as legal, such as medical, such as certain market research, the fact that AI does the first run allows even the transcriptionists to play along to a larger role because the entire industry has grown itself instead of where the price point of transcription used to eliminate the possibility of many people even considering getting a human review or a human transcription done at a later stage.

Speaker B [00:04:54]:
Yeah, that makes sense actually. My next question is how AI is actually helping transcriptionists rather than replacing them. Right?

Speaker A [00:05:03]:
Yeah. So what it does is it definitely reduces the amount. Transcription is an industry which already requires people to be very patient and very disciplined when they approach their work. It's hours of meticulous work, reviewing text and finding the answers. What AI does is it just does the first pass on it. There will always be a possibility of AI hallucinating and AI getting things wrong, mislabeling individuals who are speaking in a conversation. When that happens, a human review is the only sure shot way to make sure those errors are addressed. It changes the way transcription is done, it speeds it up, but it doesn't eliminate a human's role. And at the same time it allows the transcriptionist to build on the transcription and offer to upskill their own capabilities and offer more than just reviewing the words and offering a transcription. What they can now do is provide their inputs and observations along with the transcription, which wasn't really viable earlier because the amount of time it took to provide an accurate transcription didn't allow users allow transcriptionists to generate, to upskill them to the time to upskill themselves and provide insights, to find insights and findings in the text that they are so deeply familiar with. Do you understand what I'm, what I'm trying to say?

Speaker B [00:06:51]:
Yeah, yeah, so.

Speaker A [00:06:54]:
So that is something which moves towards more from a just, just manual task. The transcription can sometimes seem to a more celebr, more thoughtful task, more intellectual task, but you have to skim through the documents, understand what they mean. And along with the transcription and the accuracy of transcription, which only where only the humans can reach to an accuracy of 99% or 100% offer their findings which becomes really relevant in places like legal and medical, where if the transcriptionist is already familiar with the terminology, is already familiar with industry. They will be able to add their interpretations of the document along with just the document itself.

Speaker B [00:07:53]:
Yeah, and I think you nailed it. And you already talked about, you know, because my next question was why it is important for transcriptionists to upgrade their skills and learn to work with AI plus human transcription. So I think most of it, you already answered it, but would you like.

Speaker A [00:08:15]:
There are two parts to it. There are two parts to it. One is that the 100% accuracy cannot be achieved by AI, possibly ever. So that is a shortcoming where humans will always play a role. And the second thing is that the familiarity of the text that a transcriptionist achieves is unique. It's not something the client might ever get to. That puts the transcriptionist in a unique position in a special pair of shoes where if the files are, say, 12 hours of recordings or six hours of recording, it is unlikely that the client would ever go through those six hours of recording, but the transcriptionist does, and that makes their insights into the material far more valuable. So, yes, the familiarizing yourself with the specialized industry that you're transcriptioning is the best. The better you can do, the more unique insights you can provide, and that is the upskill. They should. They should, they should. They should focus on.

Speaker B [00:09:37]:
Okay. Okay, so Dicta AI offers an AI plus human transcription service. Can you explain how this process works from start to finish?

Speaker A [00:09:49]:
All right, so the client will first transcribe using artificial intelligence. If they're satisfied with the transcription's quality and they do not feel the need for it, the story ends there. But if they want greater accuracy and if the document is sensitive enough to validate that, to demand that, they can always choose to follow human review. And once they do that, we would pair the system pairs the transcript, the media, the media files, as well as the AI transcription file with a suitable transcriptionist who meets the client's requirement or who hopefully are from the same field as, as. As the. As. As the. As the material, as the media material. And then the transcriptionist ensures that we reach 100% accuracy along with, if required, give the inputs on the material, on the material offered. The client must choose the turnaround time, whatever it may be, and the transcriptionist must submit the file prior to that so that we may verify it once and submit it to the client.

Speaker B [00:11:12]:
Okay, I did not know that client could ask for, like, like their input, like, transcriptionist input, their expertise if, If.

Speaker A [00:11:21]:
If they so desire.

Speaker B [00:11:23]:
Okay, all right. So you also, you also provide that. Right? So what else Dicta AI offers beside this, do you offer any other services?

Speaker A [00:11:36]:
So every transcription is an automated summary of the transcription is created, which is emailed to the client, which just takes the. It's created by the AI. It allows you to use artificial intelligence to analyze the files. So say you have about 10 files regarding the same. Now we're talking more in the shoes of the client and not talking from the shoes of the transcriptionist.

Speaker B [00:12:05]:
Transcriptionist.

Speaker A [00:12:06]:
Yeah, yeah. So if, say you have a certain number of materials, say you have five files, six files, seven files, you can analyze them together and you can ask questions specifically related to those files in a dashboard which we call Dicta Leads. So imagine you are a researcher, you're a market researcher, and you have about 6, 7 files regarding a particular survey that you are conducting or qualitative research. Different people mention different things and you're looking for themes. You're trying to identify what points were raised repetitively by different individuals who were covered in the survey. That is something the Dicta lens will be able to provide you instantly. So you fill in the files, you write the. You put up the files or Dicta lens, once you have transcribed them, you can analyze them, you can ask the questions. And like, say you're discussing about, you're doing a survey about a logo change for a company, right? And there is a certain color changes which have happened in the logo and you're doing a survey about the color changes. You can ask the client, ask the Dicta lens, how many of these recordings did people criticize the logo, the color change, and how many of these recordings were the respondents pleased by the color change? And so on and so forth. It's just one example. But that is something that we also offer other than that we offer the capability to join every call. So whether this call or whether any other, any other call online, whether it's on Zoom, Google Meets, or the third one is Microsoft Teams, the Dicta AI can join, record the call, provide you the transcription, and provide you the summary of the file. When you're talking about Dictalens earlier, that's something which transcriptionists can work with, but we can come around to that when we go further in the. In the conversation.

Speaker B [00:14:14]:
Okay, all right. So. So it does do all. It's like highlighting the previous recording, whatever. You know, the main point, it can just.

Speaker A [00:14:24]:
The main points, you create a summary and it can also allows you to analyze multiple recordings together up to up to 11 hours of it. Just like ChatGPT, you can ask it questions.

Speaker B [00:14:42]:
Right? Okay, that's. That's Great. That's good.

Speaker A [00:14:45]:
Yeah. And it will not. It will, it will do its best. It's not bulletproof, but it'll do its best to respond only from the only factual information which are there in the record, in the recordings.

Speaker B [00:14:59]:
So in this model, what work is done by AI and what work is done by human transcriptionist?

Speaker A [00:15:06]:
There are two roles to it. Some of our customers are transcriptionist. What they do is when you want to provide the additional service, go above and beyond just providing transcription. What they can do and what they do is transcribe, make sure it is completely accurate. Right. And then analyze those files and offer the insights along with the, with the transcription which the. Which they can use the talents to do. That's one bit of it. Of course, as I was mentioning earlier, 100% accuracy is only possible via human review and wherever, in whichever fields and mediums that is a compulsory or like it's just required. The transcriptionist has to then manually go through the recording and review the entire transcription and make sure each word is correct.

Speaker B [00:16:10]:
Okay. So the transfer. Human transcriptions come like in the later phase and then they just go when it's.

Speaker A [00:16:17]:
When it's. When it's. When order is placed and it's asked.

Speaker B [00:16:20]:
Yeah, right. Okay.

Speaker A [00:16:21]:
Yeah.

Speaker B [00:16:21]:
All right. So for someone who wants to start a transcription career today, what skills skills should they focus on?

Speaker A [00:16:31]:
I would say time immemorial, the two main qualities for any transcriptionist has to be patience and discipline. Because it's, it's a super time consuming job. It does require a lot of patience, a lot of focus. You have to work for hours to go through. Like one can't be scared of hard work to be a transcriptionist you have to spend the time to go through the entire file and wean through every word present in the file. They must have some skill. They must know how to use softwares like Express Scribe, which allows them to. Which makes them detect the tech know how of doing softwares like Express Scribe, which allows them to transcribe faster. A course in transcription from an institute like TCI would be is always appreciated because it provides a certain amount of. We can presume that this individual comes recommended. Whereas compared to a layman who has never done such a course. Similarity with terminology of specific industries is another merit which one is always looking for, such as medical fields or legal fields. Yeah, that's about it. That's. That's what one looks for.

Speaker B [00:18:04]:
Yeah. So you mentioned tci. So recently DICTAI partnered with us, the Transcription Certification Institute. So what inspired you to do this partnership.

Speaker A [00:18:17]:
I think it's a win win situation for both parties in such a. When the relationship is built because the transcriptionists themselves cannot dismiss AI. I cannot live in the denial that AI doesn't exist. But they must adopt it to make themselves more efficient and offer more which Dictai is one of dictai's as AI's mission statements. So TCI fit perfectly in that dream and at the same time we can offer to. To further that goal, we could. We offered a special, special arrangement for TCI students where they can, where they can get essentially one month of a paid subscription free where they can transcribe up to five hours and pedal with explore this new field of transcription with AI and how it can help them. So where we get potential future clients, DCI gets to familiarize their student base with the upcoming AI tools which Dicta AI offers.

Speaker B [00:19:46]:
Yeah. So it's a win win for both companies. So hopefully good luck with that to both of us.

Speaker A [00:19:54]:
Yes.

Speaker B [00:19:55]:
Okay. So okay, now tell me how, how can TCI students benefit from this partnership and what opportunities does it create for them?

Speaker A [00:20:07]:
So upskilling themselves become much more easier when using AI tools like dictai. The 100% reaching 100% accuracy becomes much easier when you have already transcribed it once. With using a tool like Dictai when you it saves time to reach 100% accuracy because AI has already done the first draft. And then when the files already present with you and you already have them using things like lens, it's much easier for them to find the insights and find the solutions that the client is asking for, analyze them, make your reports, offer them more than just transcription. AI is a new technology. Using it well is something which the entire world is trying to figure out together. Nobody. I never like to talk about. I never like to talk about artificial intelligence in the way of like in the way of certain industries or certain skills or certain people will be left behind with the advent of AI because the know how that AI has, has that individual has, puts all puts them always in the special position to judge AI's output. So whatever the AI might be doing, a transcriptionist with the correct skills will be able to verify, add and make the output perfect. Dicta AI is trying to give the transcriptionists the tools to achieve that goal. So I hope that the TCI students who work who avail their free offer and get the taste of Dicta AI and figure out how to maximize their output.

Speaker B [00:22:06]:
Yep, we are helping them with a longer journey. Yeah, we are teaching them the basic of transcription. I mean, we are just training them how to be a transcription. We do have a lesson on AI, but if they partner with you, you know, you can, their skills can go to a different level as long as AI is. That's.

Speaker A [00:22:25]:
So that's.

Speaker B [00:22:26]:
Yeah, that's good. Good, good for both of us. Enough for our students for sure. Okay, so going to the next question. How are corporations using Dicta AI?

Speaker A [00:22:38]:
So a couple of things which Dicta AI has which makes it easy for them to adopt such a tool is that Dicta AI, as I was mentioning earlier, it easily integrates with things like Google Meets Microsoft Teams and Zoom. So there is no. The learning curve is really low. It's not a daunting new tool that you have. The adoption of it for a larger team will take a long period of time. I will join your calls, start transcribing them, start offering you the transcriptions, start giving you summaries. It ends up saving time because when you're having a call and you are talking about the subject matter, you are not worried about making sure you are taking the right notes down that the, the, the key things being sent by the client or in the team leader is not being missed. That note taking is already being taken care of by Dicta AI by the summaries that it's providing you. It lets you focus on ideation and like the subject matter itself, the another thing what it does is it's able to recognize when anybody is given a particular task or if somebody is accepting a responsibility in the call and creates a list of the action items. So say Matthew accepts the task of redesigning the logo as that example is something that we have already spoken about earlier. Dictai will notice that and in the summary it would have that listed and XYZ could have had accepted another task or the team leader might have assigned a particular task to somebody else, which makes working with your with your CRM much easier for you to just feed in those tasks. Okay, XYZ is doing this. XYZ is doing this. XYz is doing this. It saves times and makes meetings much more useful. Another thing that corporates can do is over a period of time while serving a client, they will have multiple meetings with the same client in which different things are said after different meetings. And if there's a dispute or if there is cross checking required at a later period of time, for whatever reason it may be, they can analyze all those meetings together, find what was said in which meeting and cite it if required. Another aspect of it is you may be able to upload. You can upload your own PDF files. So sometimes a lot of your work is discussed on calls. Sometimes you have project related scope of work files and other supporting documents. You can upload those to Dicta AI as well and you can analyze both these things together.

Speaker B [00:25:42]:
Okay, go ahead. I'm sorry. No, no, no, go ahead.

Speaker A [00:25:44]:
Which is. Which is absolutely new. Which is something which is changing the way the corporate work structure works. It wasn't like before the advent of AI such certain things like these things never existed before.

Speaker B [00:25:58]:
Yeah. So do you have any competition? Like what features you just told me are these features offered by any other platform?

Speaker A [00:26:07]:
Currently a lot of platforms offer a bits of the features that we offer.

Speaker B [00:26:16]:
Okay.

Speaker A [00:26:17]:
But there and people different, different companies specialize in different aspects of these, of what we are offering. There is pretty much nobody who's like, who's offering all of them. And what we are trying to do at Dicta AI is build an umbrella where everything transcription related can be found in one place. That's the goal that I do not see many companies around us trying to aim for. But as you mentioned, do we have competition? Yes, we do have competition. Even Zoom offers summarization.

Speaker B [00:26:56]:
Yeah, it does.

Speaker A [00:26:57]:
Yeah, it does. But Zoom lacks in accuracy to the degree that we can offer. Zoom feels at multi dual language conversations. So whether you if certain times when people are talking in Spanglish or people are if a meeting is happening in the English, Zoom has. Zoom's summaries and transcriptions struggle with. With dealing with such situations. Whereas Dictai does a much better job there. Zoom does not offer for instance capability of taking multiple files, analyzing them together or uploading your own files, analyzing them along with the call recordings. That is something that we offer. Whereas there would be a different company which would offer those things but will miss out on human review. So the dream is to be a 360 degree singular umbrella where all transcriptions, transcription needs are met. Yeah, yeah. And we are working towards that. There are new features being added every month. Our newsletter covers the new features.

Speaker B [00:28:08]:
Yeah, I was going to ask you about that. What are the upcoming features in Dicta AI that you are most excited about?

Speaker A [00:28:16]:
The one that I am. The couple of. The one that I'm most excited about right now is team team collaborative features where in a corporate structure or in a market research space there could be a team lead who has capability of viewing everybody else's work within their team so they can see their meetings, the transcriptions of their meetings, their recordings, their summaries the number of meetings taken, what was discussed in those meetings, which will allow a better control and allow less efficiency for companies. Another thing that we are super excited about is creating visual summarizations. So right now our summarizations are only in text format. But what if that was in a mind map? Visually represented AI generated summaries, which are easier to read, easier to stay up with, especially for those in creative fields where reading text might seem cumbersome. But a mind map, a visual representation of it would be easier to work with. So that's a feature we'll be launching soon.

Speaker B [00:29:33]:
I'm sorry to interrupt you, but what is mind map? I have no idea. What is that?

Speaker A [00:29:39]:
A mind map is like a graph. Okay, Right. So you had a meeting and a lot of stuff was discussed. Right. But so how. So you could represent it in a plain text manner, but at the same time you can represent it visually in a graph format, in thought bubbles, in charts, you know, so we could use dictai. We are working on making. Enabling Dicta to do that, so that every meeting, every call, every summary, every transcription task is represented in both ways, which allows it's easier to glance through, you know, compared to just a wall of text. Yeah.

Speaker B [00:30:22]:
Okay. All right. So when do you think, you know, all these features are like, I know it's coming, but do you have any timeline?

Speaker A [00:30:31]:
Yeah, yeah. So we are hoping to launch the team collaborator features by the end of this month.

Speaker B [00:30:35]:
Okay.

Speaker A [00:30:36]:
And the visual. Visual notes should come the month following that. And the golden goose, as I like to call it, is eventually offering our clients capability to turn. This is the third one, which I didn't speak of earlier, is to offer clients capability to train their own AI based on the documents available on all their calls and all the material, as well as their PDFs, and have it at their service all the time. So a unified intelligence that, that they can query whenever, just have it in their pockets. Like you are, you're brainstorming about it, you are going for a run and you are thinking about something and you ask yourself, all right, what if I do this? Implement this new idea. You can query Dicta AI there and then, and Dicta AI is answering in accordance to all the meetings that you have had before, all the meetings that you have had with your team. It's taking in consideration all the PDFs and scope of work documents and everything else regarding the project that you have uploaded. It's all there for you. And you can ask it, it will respond and it'll give you hopefully Help you get be more creative and solve your problems faster. So that is something we are working on, but that might take another six months or so to implement.

Speaker B [00:32:04]:
That sounds very exciting and very unique. So it will solve a lot of problems.

Speaker A [00:32:09]:
Yeah, hopefully.

Speaker B [00:32:10]:
Right. Yeah. Well, good luck to you on that. So that brings me to my last question. So now, for someone listening today who is thinking about starting a transcription career, what message would you like to share to encourage them?

Speaker A [00:32:29]:
I would first say that don't believe the naysayers who think AI is here to replace transcriptionist or jobs or labor in any which form or the workforce in any which form. It is only going to enhance the amount of work which is there, which you're already seeing in the transcription industry because the industry itself has grown. Do not be daunted by something like that followed by which transcription is going to be faster. That is something that we have to accept that the amount of hours and time that we spend transcribing is not going to be the same. However, it allows them the time to specialize in the particular field or in the particular space that they deal with. They do, they deal the most with as far as the transcription files are concerned. So upskilling yourself is vital. 100% accuracy is vital and that will always require human review. Spend the time, keep at it. We are in a time of turbulence as far as technology is concerned. Things are changing. But if one perseveres and one continues to. Find the right answers, there'll always be space for transcriptionists and there will always be space for clever people who are ready to apply themselves.

Speaker B [00:34:13]:
Well, that's all great advice. I hope my listeners, they took a note of that. And this has been such an insightful conversation. I think one of the.

Speaker A [00:34:27]:
Thanks for having me.

Speaker B [00:34:28]:
Yeah, you're most welcome. And you know, thank you for joining us and sharing your insight today. And this has been a great conversation. It's like the. One of the biggest takeaways from today's episode is that AI is not necessarily replacing transcriptionist, is changing the way transcription work is done. And for people who are willing to learn new tools, adapt and continue building their skills, there are still real opportunities in this industry. I also love the reminder that human skills like accuracy, critical thinking, listening carefully and understanding context still matter tremendously. So thank you, Suyash, for joining us and sharing your insight today.

Speaker A [00:35:12]:
Thank you, thank you, thank you for having me.

Speaker B [00:35:14]:
And to everyone listening, if you have been thinking about starting a transcription career, don't let fear stop you from learning and growing industries evolve technology changes, but there will always be value in people who are willing to adapt. That's all for today. Please follow us and engage with us on Instagram, Facebook, Twitter, LinkedIn and YouTube. See you next time. Bye Bye.

Speaker A [00:35:37]:
Take care.

Speaker B [00:35:38]:
Thanks.