Building an AI Copywriting Tool with Jasmine Wang

ERIC: Hey guys, it’s Eric from foundersbeta.com. I’m delighted to have Jasmine from copysmit.ai. Jasmine, thanks so much for joining today.

JASMINE: Thanks so much Eric for inviting me on.

ERIC: I’m looking forward to hearing your story so let’s dive in. Can you tell us a little bit about yourself and copysmith.ai?

JASMINE: Definitely. So I’m from Edmonton, Alberta attended school at McGill studied computer science, but I actually started out in comparative literature. So I love writing, love words, and really fell in love with CS. And the fact that you could represent words mathematically. So I became really involved in NLP research after doing a few software engineering stints in Silicon Valley, worked at Square, Lyft, and then just really fell into the world of natural language processing.

So I worked at Mila, which is or did research at Mila, which is the largest deep learning lab. Academic deep learning lab in the world is headed up by Yoshua Bengio, who all of us Canadians know very well, godfathers of deep learning. Also worked at Microsoft research in India then worked at Open AI as a researcher there.

So really navigated a few different roles, both then research and engineering as well as policy. And Copysmith really started when, out of a side project, actually so I was working on an e-commerce store with a best friend. I really love jewelry. I’ve like five rings on my desk right now.

Just thought it would be a fun project and it was a huge pain. It was not fun anymore. After I had to write my 10th product description and my second blog post and get out at some, it was my first experience doing marketing and not only was drafting the copy, really painful, but actually managing all of it was as well.

It was really hard to know like what I have running at any given point. And I thought, okay, big companies must have solved this. Like they must not be staring at spreadsheets like I am. And it turns out they absolutely are. And they’re doing the same thing. They’re staring at spreadsheets all the time.

If you’ve ever tried to collaborate, spreadsheets is a huge pain. You have to find the yellow triangle at the top right corner of each self and so Copysmith was born. We started working on i it. In late August and we just launched on October and happy to chat about it more but yeah, it’s been a wild ride thus far.

ERIC: So I guess your background from computer science and literature brought did altogether How did you go about it? Were you thinking about hey, this could be something because I experienced it or or did you talk to other people in the beginning?

JASMINE: Yeah, a hundred percent. I didn’t want to solve a problem just for folks like me. Although I saw there was a need there for folks who didn’t know how to do marketing at all. I also want you to understand, okay, what are the things that people experienced the pain points from folks where at the top of this field, I understand it very deeply.

So I actually went and chatted with, one of the top copywriters, I think in Canada. Brad Tiller at partner stock, who I worked with at breather. And initially he was really skeptical about Copysmith, which I found really interesting was like humans. The copywriters, I think have a really tough time to sending their sort of work, which is hugely important, within a company because people have this conception, that words are easy.

Cause everyone speaks right and everybody writes. So they’re like, Oh, were there easy? Actually copywriting is hugely, hugely difficult. And like Brad knows us as an expert copywriter. And he was really doubtful that AI could come in and do this. But when he first saw the product, he was like, holy crap. Not that it would replace his work, but rather that it would make it, much easier to cover more creative territory.

And we’ve seen this again with doing interviews with folks who are like very experienced. They’re like, oh this is super interesting. Not that I would take this right away and deploy it, but I didn’t think of that before. So I really think, Copysmith for users like myself, who haven’t done a ton of marketing.

Yeah, go ahead. Deploy it. It’s better than what you had, right. It’s way cheaper than hiring an agency. It’s easier than using a freelancer a hundred percent. Like that’s, it’s a solution in that way for you. But what we’ve discovered as we have talked more, experienced marketers, and focused on growth, is it it’s more about the ideas and brainstorming partner aspect. It’s just really sucks staring at a blank page, right? Right.

ERIC: That’s a really good point because I think it can be a platform to enable the user to do more great things in marketing. So because I think a lot of people just get skeptical about like AI replacing jobs and those things. But I see it more as a tool or something that would make your job maybe a lot easier. So speaking of AI and like marketing where do you see the industry is heading?

JASMINE: Definitely. So right now, I think there’s. There are different bottlenecks that I see in marketing. And I’m sure I’ll discover more as we continue building this product. But one huge bottleneck that we see is bottleneck in experimentation. You just can’t test enough variance for enough segments of the population.

ERIC: Right.

JASMINE: And most of that is copy bottlenecks. Like how do I write copy for like 20 to 25 year old women who liked this page and they’re based in California. And should that copy be different from every other specific segment, but you might have, let’s say hundreds of these segments. There’s no way that like an early stage startup team can test variants across that hundreds of segments.

And that’s the bottleneck that we’re looking to unblock for folks who are dealing with that sort of volume. So I definitely think we’re moving towards paradigm where if we want to do this micro targeting and really specific segmentation humans, you don’t have to take, I wouldn’t to call it the backseat, but really a supervisory role where they look, they can edit and sort of hit, okay.

I don’t like this variant let’s discard that, but really playing an editorial role instead of being in the driver’s seat of producing the actual copy, because it was just way too much. There’s too many segments to target. To write four hundred, it’s much easier to edit a hundred pieces of copy than to write a hundred pieces of copy.

So I see one definitely like a path forward for removing bottlenecks. But the second thing I might mention here is, I think we’re going to see increased personalization throughout a customer individual customer’s journey as well. So again, another bottleneck with, people spend so much time rewriting their home page and AB testing various of the homepage and people get more experience.

Also start AB testing variants throughout a customer journey, but extreme version of that hasn’t been achieved, which is, imagine a world where pages are dynamically generated based off of users like either query friends or just things we know about that specific user in a much more personalized way.

Then before so like for example, imagine we had access to your like message history in some sense. And we even see if there are relevant keywords to this product and we can generate a blurb. That’s a pain point. That’s not very general. It’s not a targeted at people who live in Toronto between X and Y H no it’s targeted Eric, who has mentioned a specific problem before and his message and that’s dynamically generated. So I think we’re moving towards that sort of world, which is really interesting and also has like, obvious implications for like potentially fake news and how people are digesting information. But yeah, I think, I think the way forward is definitely increased personalization.

ERIC: You had the idea and went talking with some users of course, and getting some insights. What was it like building your team? how did you meet that person that actually kind of believes in you? How did that conversation go?

JASMINE: A hundred percent. This is always the hardest part, right. And the thing that’s like the most important to get right is getting the right team together. And I feel really blessed have an amazing team. So I think a lot of the velocity came around, launching on Product Hunt.

I should back up a little bit. My CTO is my sister. So she trusts me a little bit, you know. But yeah, we’ve worked on projects together, actually, that e-commerce store that I mentioned last year. He helped me with that as well. So she, and she helped with like some of the social media and she was like, this is like really, really hard.

So she also experienced this problem and she’s an amazing software developer. So she came on board after I had been working on it for a few weeks and trying to convince her like, and others, an idea, like I think a lot of people would find this really useful. But it really things really started, getting rolling when we launched on Product Hunt, because I think a lot of people saw it and were just like, this is helpful.

I want to help. So, for example, Ian who’s is our chief creative officer has like 20 plus years. agency experience, he’s like an international best selling author. He found us through product hut and he reached out. I was like, I would love to help, like in any form, I’ve been thinking about AI and creativity, my whole life.

I was like, so I think there were a lot of folks who just like saw it and were like, whoa, this is a thing I want to work on. I feel really blessed about like, this is such a set. The idea is compelling enough to draw people in. And then there were the traditional channels where, our amazing, designer, um, was introduced to me through like a close, like founder community fall interact that’s based out of Silicon Valley, like lots of, um, just like top researchers, like, top researchers like founders, they select like 50 every year.

It’s one of my, um, Most important communities in my life, those introductions through that sort of community. Um, so all these bits of different bits and pieces, and last, we also hired, someone who we found while going through the startup Edmonton program, which is a local incubator. And they helped like promote our job postings. We as well, so definitely different channels, but yeah, just super blessed, like love my team.

ERIC: Oh, for sure. I think that one of the biggest key thing is that it always begins with the founder, but I also think it’s the product too. Cause your traction, it’s super impressive.
So I think those things always help. Because I think a lot of startups struggle in the beginning because, hey, they haven’t proven the traction but if there is traction, I think recruiting always comes easier. Let’s talk about the MVP, you know, like what was the first version of like?

JASMINE: Our MVP MVP was not deployed on product hat. Our MVP was a tight form where we sent ads or the text back to people via Gmail. Like it was, is Typeform plus Zapier integration. So people get what they wanted and then we would send them back like a few dozen variants of like whatever type of content they wanted.

And it was all hooked up through Xavier backend. So completely no code. We just had like a lab. And even with that, we built like a few hundred people on the wait list about 400, which, who later did not convert to our product, but it was okay. We held up that wait list. So in terms of building a MVP that gave me already interesting enough data and customer users who are willing to talk to me concretely about something that like a pain point that they were experiencing, but I would probably call that the minimum viable products maybe, but not the minimum lovable product. So level of product, with our Product Hunt lunch and building the MVP was humongously painful because, I haven’t done react development a long time. Hopped in like last minute, we pulled in help, lots of favors from friends who are engineers.

It was a very, very hacky process. So in terms of like tips, when it comes to building an MVP, I definitely think we’d launched later than we should have. And this is something that everybody says, like, you probably launched your MVP too late, but and the heuristic was when we launched on Product Hunt.

I wasn’t embarrassed about it and there’s sort of this startup frame that if you launched and you’re not embarrassed by your product, you’ve launched too late. And I think we lost far too late. We could have launched two, three weeks earlier and I think it would’ve been good to launch earlier so definitely that would be my number one tip.

Like if you have something in your mind and you really want to build it, go, go make it happen. Like use a no code tool. If you’re not technical, go hire a contractor. You really believe that there’s the market for your product, and you’ve already validated that you can spend that sort of money, like just get it out there. Cause I reflecting back, like we could have built MVP very quickly as soon as we had the initial designs and release in September instead of October huge difference.

ERIC: Wow, those are great pointers for our audience looking to build their MVP. So let’s talk about the Product Hunt launch. That was a massive success. I saw it, hats off to you. What were some key ingredients? How did you coordinate it? Any tips on such a massive successful launch?

JASMINE: there was a ton of playbooks online that I would encourage folks to look at. So Julian Shapiro of curve released a really good playbook on relaunching on Product Hunt that we followed.

I would highlight some key tips here. One get a great Hunter. So we were lucky enough to get Chris who’s like the number one. Product Hunt Hunter, because this means that when they hunt you all their followers are notified that they’ve, um, and ideally you get top Hunter on your product because they can feature you.

Like you go straight to first page. We actually weren’t lucky enough to get this because we were the third product. Chris had hunted that day. You only get one featured product a day. Lucky, actually, we were at a disadvantage, but you actually need to plan your Product Hunt launch, like. A month prior, because there are a ton of assets you need to create.

You need, you need to secure that number one spot with your Hunter to be featured, but you also need to create a video. You need to create image, assets and music, create a logo. So start chatting with hunters like one month before you plan a launch on chronic, how we did it in a week, that’s not recommended.

Number two, create a calendar invite or like a Facebook event and invite everybody you’ve ever talked to about your, to that, um, to that event. Just go through your calendar, look through all your meetings and buy all of them and just have a little blurb. That’s like. Hey, I’m like you supported my journey and obviously make it such that people can’t see what other guests are on the site.

Time it for like when people wake up and just put a calendar block, like instead of messaging people, one by one on the day off, this is just way more effective. If it gets on their calendar, they know that you’re launching something. And you can link the link as soon as it’s up. You can say like, could you retweet this for reach?

And all that information is in the calendar invite in the same place and they don’t know how many people are invited. Like I think before our calendar invite, we had like almost a hundred people on it. Cause I tried it so many people by that line about the product. And I did another Product Hunt launch recently, actually for company to box, which reached number one.

And we employed the same strategy, but we have to do something different. We have two invites, one for folks who are in a European time zone. So we had one very early in the morning and at like 1:00 AM right after the midnight launch of Product Hunt.

The third tip. I’m trying to only give tips that you won’t read online because it’s like you should launch a high traffic blog launch at midnight. Last would be offer a Product Hunt specific offer. So we got about 4,000 folks visiting our site on the first day, just off of Product Hunt. And I know we offered a limited time. We are going to be a hundred credits for the free trial.

If you sign up, we’re going to give you like a discount code that the first 50 folks who sign up from ProductHunt, get to use. And just within the first day we had three people convert into paying customers. And I think that was $200 in MRR, like even before their trial ended, because it was like my box, like this coupon code, you don’t know when I’ll run out. Because there’s only 50 seats. So those would be the three things I would mention for sure.

ERIC: Wow, that’s awesome. I think that that’s great because when you launch these things, it’s good to plan a little bit and not just to rush it out there, especially for sites like Product Hunt where they’re so big. I think it’s good to have some planning there.

JASMINE: I think in traffic, from Product Hunt so I definitely think it’s a worthwhile platform to invest some time into.

ERIC: Absolutely. That’s great so let’s shift gears, you know, as a founder, obviously many projects on the go, always working on new things. What’s the key to keeping it balanced. What are tips you might have for just balancing out?

JASMINE: Like practices, self care in different ways. And this is one of my main like takeaways from 2020, which definitely had its own ups and downs. As you know, like my family went through a health scare due to COVID and really what I’ve focused on is just what are the things that make me forget about other things?

And like when, when will I not be checking my phone because I’m absolutely phone addict, I’m addicted to checking our like user count increases like the graphs. Recurring, right. The things that work for me are, weightlifting. And so while the gyms were open, I lifted weights every single morning of fourth house.

Now I don’t do that, but I just do body weight exercises also in the morning. I meditate so not for a long time, over 20 minutes a day. I love waking up by Sam Harris. It’s like definitely the best meditation app. I’ve tried them all. And the last thing I do is actually just take a really long bath every week.

And I read in the bathtub, which is horrible and it’s bad for the books but it’s absolutely my favorite thing to do. Those are definitely the top three things, as well as trying to have at least one day a week where I do less work. So it definitely has a founder like I’m workaholic and I worked 24 seven and I love working, um, And I really try to take myself away from that at least once a week, I try to spend time with friends, like go call, like do calls, catch up with people and try to put that all in one day. And that’s my break, like being, being with people, um, who I care about.

ERIC: As a founder, it’s always a challenge juggling things and staying healthy and staying focused because if the health is not there, the start up won’t be going anywhere. I always ask the guests on foundersbeta what’s the best advice you’ve ever gotten, whether it’s startups, whether it’s philosophy, what would you say?

JASMINE: So Michael Nielsen actually told me this who’s like a, um, amazing physicist, um, but just does amazing work across a number of verticals. Um, pay attention to what you find. Beautiful. And I found this incredibly moving, because I’m a pianist and a writer and artists. And during that period of my life, when he gave me this advice, I’d sort of forgotten about those parts of myself and how core they were to my identity.

But I think it applies to everyone, not just folks who are artists, because it’s really about like, what do you find? To be the most moving parts of life. Like if you reflect like one where the points at which you felt like a sense of awe and like, how can you like return not returned to those points, but like what, what was the core of that experience that made you want to go, wow, like this is sort of what life was about or like, this is one of the best moments of my life.

And that helps shape your sense of taste helps you guide your everyday. Like, we can’t experience that every day, but it might teach us something about our orientation. So, for example, some of my, um, I’m not a religious person, but some of my, most beautiful experiences have been when I’ve been listening to music and in like a religious, building like that, those moments are hugely moving for me.

And this has led me to realize that I like that sort of music and I can replicate that for myself within yeah. The ton of COVID. So sometimes when I, I have this song that I play, that feels quite spiritual to me, although it’s not written in that way called Census Bureau by list. And I tag it as one of my, um, songs to listen to whenever something important happens to me.

So whenever I play it, I’m like I remember everything good basic the last three years. So paying attention to like, what you find beautiful, developing your sense of taste so you can create that more in your life every day. Absolutely the best advice I’ve been given.

ERIC: Right. What’s the next exciting thing for your startup, what should we look out for in 2021 of course?

JASMINE: there’s a lot of amazing, product updates happening that I won’t reveal yet, but, um, folks come to us saying I need to launch 30,000 products on a client site, or I need to manage multiple clients. So we’re excited to support folks who have bulk needs work with some of the top retailers in the world. So we’re very excited to make, those sorts of announcements.

We’re also hiring, so we’re hiring for a senior engineer, as well as sales and growth. So adding team members are super exciting. And we have some exciting potential funding related updates in Q1. We are all discussed yet, but those are the main three things.

ERIC: Amazing, amazing. So for our audience that want to learn more about your startup about you, where can they go?

JASMINE: Yep. Just go to copysmith.ai. There’s a hundred percent free trial. We don’t gate it at all. You just have to sign up and you can reach out to me directly. I’m jasmine@copysmith.ai.If you want to partner, if there’s like questions you have, super happy to chat with anyone from the foundersbeta community.

ERIC: Awesome. Jasmine, thanks so much for taking the time to come out today, share your expertise and we’re super excited to see some development with copysmith.ai in 2021.

JASMINE: Thanks so much, Eric, this is fun and apologies for my grandma walking around in the room.

ERIC: No worries. It’s all fun.Thanks again, Jasmine.

To learn more and connect:

Website: https://www.copysmith.ai/ (Get your first 3 months at 50% off with the code FOUNDERSBETA.)
Twitter: https://twitter.com/copysmith_ai

FoundersBeta