AI has become the hot topic across marketing, raising questions about its potential impact on the industry. Abhi Godara, CEO of Rytr, an AI content generator, shares his thoughts on the future of AI, and explains the technology behind Chat GPT and how other platforms, such as Rytr, build on this technology. He also shares how to get the most out of AI-powered content and why being aware of its limitations is important.

Listen to the podcast now via the links below:

Transcript: Interview with Abhi Godara – Rytr

Speakers: Mike Maynard, Abhi Godara

Mike: Thanks for listening to marketing B2B Tech, the podcast from Napier, where you can find out what really works in B2B marketing today.

Welcome to marketing B2B technology, the podcast from Napier. Today, I’m joined by Abhi Godara. Abhi is the founder and CEO of an AI product called Rytr. Welcome to the podcast. Abhi.

Abhi: Thanks for having me, Mike.

Mike: So it’s great to have you on I mean, I’m interested learn about Rytr but first, you know, can you tell me a bit about your career journey? And how you got to the point where you decided to found Rytr?

Abhi: Yeah, absolutely. I mean, like most good things in life, nothing is like a linear path, I guess, to where you get to today. But I started my career as a professional consultant working in London in one of the big four companies back in 2007. Eight, did that for about five years, mostly in strategy consulting, bit of private equity work as well. And then I moved into startup space, pretty much for the last 10 years, that’s where I’ve been working and started as an early stage investor in one of the leading seed funds in India, worked with more than 50 startups, 150, founders across product marketing, fundraising growth, you name it, all those areas where founders need help, and then started my own sort of venture studio based out of Valley, late 2015 16. And that’s where I’ve been dabbling with a lot of homegrown ideas incubating quite a few product companies, mostly SAS companies, over the years, you know, some, I would say outright failures, a couple of moderate successes and a few whole brands. So that’s how pretty much the journey has been over the years. But yeah, you know, it’s my passion to work with entrepreneurs, who are, you know, solving big problems with innovative ideas. So that’s basically what I love doing.

Mike: That’s awesome. And I love the fact you’ve done this in different countries. So I think that international view is really interesting.

Abhi: Sure, absolutely.

Mike: So you founded Rytr, I mean, Rytr is an AI tool to help people write, unsurprisingly. And if anyone’s listening, it’s spelt ry T R. So that’s the product. What inspired you to build a tool to do AI generated written content?

Abhi: Yeah, that’s a great question, Mike. So as an entrepreneur, you know, I’ve always found content generation to be a pain, especially when you’re a small team that is just starting up. And it’s a fact that many startups and professionals fail because they don’t possess effective marketing and copywriting skills. And moreover, a lot of entrepreneurs, you know, potentially give up on the idea, due to the overwhelming nature of content creation. And I’ve been in the AI space, you know, for the last five years, started working on a chatbot platform for influencers and creators, which, you know, scale to millions of users at one point. But we didn’t have the technology like Chad GPT at that time, right, or GPT, at that time. So when GPT three came out, I think this was back in 2020, you know, so we realise the potential of this technology and the market, it could race, you know, copywriting, creative writing was one of the first use cases which kind of emerged from this, this tool.

So we looked around evaluated some existing writing tools, and we’re not the first ones in the market, we were definitely in the first, you know, few players, you know, who built something like this, but there were other players out there. But we found the experience very frustrating tools for delivering you know, subpar outputs, it was very overwhelming in terms of UX and UI, there was a lot of cognitive overload for users to get started. So at that point, we decided, okay, let’s give the market what it deserved. An intuitive a writing assistant, which offered the best quality of output at a very sort of fair price. So although we were slightly late to the party, but with limited resources, and small team we launched in April, I think 2021. And since then, we haven’t looked back Currently, we are serving close to, I think 5 million customers globally, with almost perfect ratings pretty much everywhere, and recognised as one of the market leaders in the space. So yeah, that’s that’s been kind of a journey that we’ve had over the last couple of years.

Mike: That’s a huge number of users. And I’d like to go back to that. But first, I think it might be worth for some of the the less technical listeners, you talked about chat GPT. And you talked about GPT. Three, can you explain what the difference is? And the technology that actually underpins Rytr?

Abhi: Yeah, so I mean, technology is pretty much like if you go to the really fundamental get a level that technology is called a transformer models. It’s called Bert, which was pioneered by Google back in, I think 2017 18. So all the sort of future evolutions that you’ve seen in terms of GPT 123 3.5. And now chat GPT is based on that underlying principles. And I would say model language model so to say, so that’s pretty much I think, powering all the applications in writing applications that you see around us. So charge GPT is just an evolution of Jack GPD. Three, which was like one of the, I would say, more mainstream models, which, which a lot of AI lighting companies started using, you know, bank starting from 2020 till the end of last year, and GPT 3.5, or chat GPT as this call, it’s just a more refined, sophisticated version trained on even bigger datasets than than its predecessor. So that’s essentially, you know, the difference between the two. So obviously, it’s, it’s trained on one data, it’s more powerful, it can give more sort of, I would say, better outputs, higher quality outputs than its previous versions. But yeah, the underlying nature of the technology language model is still the same.

Mike: And I think we’ve all you know, played with chat GPT, and been been impressed by its ability to communicate it in what feels like very natural English. But but I’m interested, you know, you’re obviously using, you know, this model to build a tool that specifically for writing. So what are you doing differently to what’s been done, for example, in chat GPT? To make it, you know, better suited to writing blog posts or adverts?

Abhi: Yeah, absolutely. So we have our own sort of training data. And this is what we have refined over the last couple of years, you know, again, with Chad GPT, or any other sort of piece of AI writing technology, it’s, you know, the basic principle of garbage in garbage out is still true. So if you, if you just throw some random inputs are sort of ill defined prompts, you know, the output that you might get is probably less than optimal, right. So we do a lot of pre formatting, you can see at the input level, and kind of post formatting at the output level, to make sure the output is aligned to the intended use case, or if it is, social media posts, blog posts, you know, your job descriptions, or song writing anything, there is a level of I would say intervention that we have to do from Rn, to make sure the output is customised. The second thing is the reliability of charge up like the the UX, the UI, whole sort of experience of people getting used to it, you know, it takes a little bit of time in the absence of any sort of education. So that’s where we have created this very seamless interface, very easy to use navigate, so folks can get started immediately, right, without having to learn the ABC of, you know, AI copywriting techniques. So I think that those are two things we have done. So we have abstracted away all the complexity that users have to go through to understand and use this technology, and to obviously, making sure that the use cases are aligned to the sort of intended needs of the end users. And the third is obviously, you know, the pricing and the value for money aspect. So we are still one of the most, I would say, value for money products out there in this space. And that’s how we’ve kept the whole proposition. Very, very oriented towards, you know, early stage users, smaller teams, you know, who do not have necessarily have the bandwidth and maybe the budget to go for, like, you know, more expensive solutions out there.

Mike: I think that’s that’s a really interesting point. I mean, you’re giving people quite a lot, because, you know, you talked about the underlying data, you’re adding extra data. So so your product understands adverts better than maybe chat GPT does. But you’re also, you know, almost providing this structure, this kind of wizard to help you create content. So I mean, what are your users really looking for? Is it the quality? Is it improving the speed of generation of content? Or, you know, what’s really driving the way that you’re introducing features for the product? Yeah,

Abhi: I think I think it’s a bit of both, actually. So I think if you if you just say, Okay, well, it’s about speed of content creation, with compromising the quality, I don’t think it works. You know, people want everything, you know, they want faster content generation, higher quality output, at a very affordable price point, right. So you have to take all those boxes. And, you know, luckily, nowadays we have, we’ve been doing all three of them at the same time. So you know, things like just a document management, the workflow management, again, going back to the point that we abstract away all the complexity, so you can, as a Rytr, you have to not just create content, but you have to manage the content as well. So creating documents, you know, sharing those documents, downloading that content, managing your team, allowing your team members access, seeing the analytics, history, all that stuff, is what you need if you’re running a proper business, right. And those are the things which you cannot expect in a standalone are sort of chat GPD kind of platform, which is more geared towards, let’s say, just casual use cases and, you know, end users who are not necessarily entrusted into those kinds of workflow management tools. So we provide that suite of features so that users can get the maximum value while at the same time they can create really high quality content with the least amount of time it takes to get there. So yeah, so you know, we have to balance out between those things. We are constantly adding features which can improve that workflow management for smaller teams, freelancers, agencies, and of course, keeping an eye on how can we improve the quality of output, you know, Every day, even if it’s like point 1% improvement, we try to make sure you know those interventions are added so that the quality gets better over time. So it’s a compounding effect.

Mike: I mean, presumably one of the biggest challenges you face is where you see a lot of AI generated content, you can begin to feel particularly from something like chat GPT, you just get a sense that it’s not a real human. So what are you doing to really develop the product to make it feel much more human when people are reading the output?

Abhi: Exactly. So I think this is more of a philosophy question like and that’s, that’s a good point here is because as a company, as a team, as a product from day one, our philosophy has been, we don’t want to encourage content factories to be built on top of this era, I think platforms, you know, the world doesn’t need more content, it needs better content, and motivated content. So if you if you look at how it works on Rytr, when you play around with the tool, you will notice that we don’t mindlessly allow people to generate content by pressing like just, you know, keep writing keep writing kind of button, it only takes in a limited amount of input, and then gives out a certain amount of output so that people can review the outputs when they come out. And they can edit and then refine it as they go along. So it’s not like you press a button, you have like a 5000 word blog post ready for you to be published. And I think that’s where a lot of people are getting it wrong. I mean, unless you spend time effort and reviewing and refining outputs, it will feel very mechanical nature in some shape or form. The second thing we do is we provide a lot of these granular controls, like we have a feature called readability score, which gives you the idea of how readable the content is. Second is we have an inbuilt plagiarism checker as well. So you can check the authenticity of the content. So you can just select any piece of text and then run it through our plagiarism checker, it will tell you whether it has any piece of copied content or references that you can edit. So we give all these controls. And again, this is this is what’s this is something which adds up to me that will won’t provide you out of the box, right? So all these things make the content writing experience much more, I would say emotional and practical for the real world use cases.

Mike: It’s interesting. I mean, what you’re describing is a product. That’s that’s not really designed to write content, but to accelerate that content writing. And I think it’s really interesting, you talk about pleasure, and I think a lot of brands are going to be very worried about plagiarism with with AI. I mean, certainly some of the early AI, generative text that we’ve seen, has has had plagiarism in it and has caused a lot of problems. I think CNET got into a lot of trouble recently, didn’t they?

Abhi: Right. Right. And yeah, again, I think you have to make that clear to the end users, and you have to give them the right tools so that they can address those things as they go along. So I think it’s ultimately responsibility, the platform to encourage, you know, the right kind of writing behaviour, I would say.

Mike: And I mean, another thing I think that people are concerned about is where, you know, AI generated content has data or facts inserted by the AI and whether the AI is actually correct or not. And I know, you know, Google recently ran an ad where they actually had something that was wrong. So, you know, I mean, treated me, Sam, when, you know, said that chat GPT wasn’t designed to be right. Are you doing things to try and make the output factually correct? Or do you see that as being something where really, because it’s somebody’s producing it for a project, it should be driven by the human and the human should be driving those facts and information?

Abhi: Yeah, that’s a good question. I mean, so like, again, I mean, we encourage people to use it as a bit of creative Rytr’s block, kind of tools to end the Rytr’s block. But at the same time, you know, when you get the content out, there is no guarantee that it will be 100%, you know, factually correct. So we encourage users to spend some time cross checking the facts and stats which are thrown at them. And, you know, again, let us this is part of some of the feeling that we do, like, on our site, the, you know, the prompt engineering, so to say, is to avoid throwing exact or specific numbers as much as possible, and leave that task to the end user. So they can decide what is the best, that are number of figure that can fit into that particular piece of content. But inevitably, you will come across cases where still AI would probably, you know, generate on its own some of the stats, which could be fake. So we encourage users to review and that’s another reason why we ask them to you know, go through things, you know, with a fine tooth comb to make sure there are no sort of random figures. And one thing which we are working on internally is called Fact Checker. So we are trying to work on, you know, these tools and features, which can allow users to fact check some of the numbers which are thrown by AI or generated by AI. So that could potentially really address this issue. Big time.

Mike: That’s interesting. I love I love the fact check it out. I think a lot of people would fill you know, reassured if there was some degree of checking, you know, what’s claimed in an article. I mean, another interesting challenge I think people have is Is that when you’re using AI, the AI is fundamentally trained on a training set, and kind of produces the average of what the training set is. Are you looking to, you know, somehow train the AI on the very best marketing material, the very best blog posts? Is that something that people in the AI sector are trying to do? Or is it all about volume of content?

Abhi: I think that’s an interesting question. So yeah, I mean, we, you know, some of the copywriting use cases that we have, we try to give those best, you know, kind of best practices, so to say, the swipe file kind of examples, so that AI can produce content, which is aligned with that, that sort of examples and samples we have shown, but still, there is a high probability that it will just generate based on the earnings, it has had, you know, based on the underlying data set. So it’s difficult, but again, you know, with a lot of fine tuning a lot of examples that you can provide, it obviously gets better and follows the guidelines that you have provided, and tries to stick to, you know, those kinds of examples, one of our sort of sister companies, Poppy Smith, they have a very unique approach to addressing this issue, where they only work with like bigger companies, enterprises, instead of taking their existing content and trying to fine tune the AI models. So the content that is generated is very customised to their brand, voice, their sort of product and description that is already out there. So yeah, so there are ways to do it. But again, we want it to be a little more open ended, and less, I would say, one particular brand or sort of use case focused.

Mike: One of the things I’m interested in, you know, just moving on to some of the applications. Is there an area you think that that generative AI today is doing really well? I mean, do you think, you know, using a tool like Rytr is best for, you know, short form social media posts for ads, or for blogs? I mean, where do you think it really shines?

Abhi: I think you’ve hit the nail on the head, like when you say, you know, creative writing, content writing, I would say, and I think that’s what Simon was alluding to, maybe in the quotes that you mentioned, it’s not meant to be like 100%, factually correct. It’s meant to remove that writers block that you face, in your creative content generation process. So if you’re writing blogs, if you’re coming up with video ideas, or add ideas of social media posts, I think that’s where AI could really help you as an assistant, to throw new ideas and new sort of direction of thinking, you know, so to say, and I think that’s where it really excels. So whether it is next generation, or image generation, or any sort of similar things, I think it really opens up new possibilities in terms of ideas that you can explore as a copywriter, or a content writers. So that’s where it excels. So I wouldn’t expect it to write novels end to end fully formatted, completely factually accurate. I don’t think that’s the intended use case, at least as of now, you should think of it more as a tool in our repository to sort of just get rid of that writer’s block and come up with new angles to write about or think about.

Mike:  I think I think that’s really interesting. I mean, you know, looking at it as a tool to help the writers today is fascinating. I mean, some people are almost saying, you know, writing is dead, it’s all gonna be AI. And clearly you believe that writers have a lot of value to add. I mean, how do you see AI changing over the next five years? I mean, do you think it’s gonna get dramatically better? Or have we seen a big jump in performance, and now it’s maybe going to hit a bit of a plateau?

Abhi: I think you can probably see some of the possibilities already in front of you, right? I think the vision of AGI doesn’t feel very far fetched now, with how the technology is evolving. I think the use cases will emerge in other industries as well. So I think what we have seen is just barely scratching the surface in terms of content generation. But I think where you will see more of it being used is other day to day tasks. So things like predictive analytics, you know, doing tasks on your behalf, automating a lot of internal tooling, in a company answering, you know, questions on your site. So these are things where maybe, you know, content creation, or new ideas, or less of a use case, but more about, you know, how AI can actually do tasks, different kinds of tasks, in a much better simplified and efficient way for a variety of use cases. I think that’s what I’ll see more. I mean, I think we’ll see more of over the next five years, whether we’ll we’ll get to see that dystopian world some people have, you know, probably envision is yet to be, I think it’s still it’s still far fetched. And I don’t think we’ll we’ll get there. It’s a new piece of technology, which we should embrace, try to embed it in different parts of our lifestyle and different tools that we use, and that’s how I think it will become over the next five years, just like an invisible piece of technology is there to help you and guide you. A lot of new kinds of categories of jobs and skills will emerge. So I think some of the concerns are overblown, some of the potential. You know, I would say impact is also overblown, maybe in a dystopian sense. But, you know, I think we have to use it wisely and use it for the right use cases, I think it can be really powerful piece of tech.

Mike: I’m pretty interested. I mean, the way you talk about this, it’s all about, you know, speeding up that process of generating content. I mean, do you do you have a number or a guide as to how much quicker someone could write a blog post, if they’ve got support from from a product like Rytr versus, you know, trying to do it all themselves or an ad or anything like that?

Abhi: Yeah. Finally, actually, we we’ve had a tool on our website like homepage, from pretty much the early days when we launched. And this is, this is exactly what we went, you know, it’s just an indicative sort of assessment of how much time and money you can save with a to like Rytr. So it basically takes in the number of words you write, and we have some sort of logic in the background, we tries to calculate, okay, if you write this much content, then you’re probably spending this much time and you know, each hour of your time is probably this much in dollar amount, right on average. So that gives us a sense of how much money and time you’re saving by using a platform, right Rytr, based on how many works you do, right? So it’s there right on the website. In fact, one of the stats we show when you land on it, is how much time and money people have potentially saved by using a platform like Rytr.

Mike: That’s awesome. And I think, you know, I do feel sorry for people who, who are, you know, writers as a job, because traditionally, they’ve had very little investment in them. I mean, you know, you buy them a word processor, and that’s it. And so it must be fairly easy to show massive ROI, you don’t have to improve speed that much to to get value from a tool like Rytr, I think it’s fascinating. Right, right.

Abhi: You know, again, just touching upon that, I think, I think if anything, it will have a positive impact on the content creation process as well. So, you know, I firmly believe that people with highly, I would say, sought after skills are people who are really good at what they do, whether it is copywriting, blog writing, or just coding or anything for that matter, they would probably benefit from this, because now you can probably appreciate their value even more. But I think some of the middle management and mediocre skills, like just people writing content, for the sake of it, nothing original nothing, you know, inspiring, I think they will probably have a hard time because that can easily be replaced by something like, you know, GPT, for example, or AI can do it for you. So I think it becomes important to upskill yourself, if you are one of those sort of, you know, middle layers to try to, you know, get to get to, I would say more close to the client requirements, understanding the end user personas and writing content, which is really authentic and original and inspiring, which is good for overall, like I would say, for the whole space marketing space.

Mike: I really like that positive view of things, I think it’s it’s good to see that as producing, you know, as output fairly average content, if you’re above average, you’re going to be more valuable exam. So if you upskill yourself, I think I think that’s great. We’d like to ask a couple of more general questions. So it’s really interesting. I mean, you’re on the forefront of some massive change in marketing in terms of bringing AI to marketing. If a young person was thinking of marketing as a career, would you advise that or having, you know, seen a lot of startups and work with them? Would you advise them to do something different?

Abhi: No, absolutely, I think I think even more, so I would encourage them even more. So now with this technology, because like I said, if you’re really champion of your skill, then I think your value is going to go up, even with this piece of technology. And if you know how to use this tech to your benefit, then it is even manifold the impact that you can create. So I would definitely encourage, I think, I always believe that the first principles, the fundamental needs never change, you know, marketing, still marketing, you need to put content out there, you need to target certain people with the content, and you need to sell the solution. Right. So the best piece of marketing advice, I think I got was, don’t think of it as a marketing, you know, as a different function, it should be an extension of what you’re doing, like a product you’re selling. So the best marketing is something which doesn’t come across as marketing, it comes across as educational, it comes across as helpful. And just as an extension of what you’re actually selling and making money on. So I think that skill is still going to be even more valued going forward with with AI. And I think if you know, your way around using AI, then you will be even better positioned going forward. So, you know, keep at it, I would say

Mike: That’s great advice. I mean, I’m sure people listening to this will be quite excited and you know, pleased to hear that actually Rytr’s there to help them rather than to replace them. If they wanted to try right. How would they get a chance to to actually use a product and experiment with it?

Abhi: Yeah, absolutely. So we again, we take pride in being one of the most seamless and easiest way to get started with here I think, you know, space so Just go to our website, right a.me You know and start writing, you will see easy to sign up process, just sign up with any of your social accounts or email accounts. And then as soon as you’re inside, you can just start generating content for a variety of use cases, we offer a very sort of healthy, I would say, free plan. So you don’t have to put any of your payment information, you can generate up to 10,000 characters, and use all the features that we offer pretty much. And if you need extra credits, then you can sign up to our zero plan, which is again, very, very generous, just $9 a month, and you can generate up to 100,000 characters and some images as well. And then if you really want to up your game, then we have an unlimited plan, which is $29 A month or Yeah, and you can generate as much as you want. So it’s it’s fairly easy to get started.

Mike: Yeah, and I think most people, if they’re like me, they’ve sat down, tried to write something and been faced with a blank page and writers block. You know, that sort of pricing is pretty cheap to avoid that pain.

Abhi: No, absolutely. And yeah, and that’s why I think it’s such a lifesaver for a lot of people because, you know, you’ll get tonnes of value for the money that you’re spending, there is a lot of value of getting from A to like Rytr.

Mike: I really appreciate you been a great guest. If people listening to this would like to know more information, I’ll get ahold of you, what would be the best way to reach you.

Abhi: I mean, you can connect with me on LinkedIn, or you can just drop me an email at abhi@rytr.me. That would probably be the easiest way to get in touch with me directly. And yeah, you can follow me on Twitter as well. Abhi_Godara is my handle. So if there’s anything I can help you with a writing space using Rytr or anything else, just feel free to reach out to me, please.

Mike: That’s very kind. And thank you so much for being a guest on the podcast. I really appreciate it, Abhi.

Abhi: Thanks, Mike. Appreciate you having me on the show. Thank you.

Mike: Thanks so much for listening to marketing B2B Tech. We hope you enjoyed the episode. And if you did, please make sure you subscribe on iTunes, or on your favourite podcast application. If you’d like to know more, please visit our website at Napier B2B dot com or contact me directly on LinkedIn.

artificial intelligence in B2B marketing content marketing media page podcasts