Shopify AI Reality Check: 166 Stores, 1 Truth | Hitesh Matlani | Rachit Magon | Chainet Podcast

Welcome to the 30th episode of ChaiNet, where we learn from builders before the machines take over. I'm your host, Rachit Magan, and today we're talking about something every Shopify merchant or Shopify dev is asking, that what does Shopify's new AI actually mean for us? Right. So Shopify has just released more than 150 different features. And those are mostly all of them are AI features. People are literally going crazy over them. There are influencers talking about it. There are people hyping it up saying that, you know, it can, it can basically act as a full-blown agency. People are saying now that you can open your Shopify store and start earning millions in months. Of course, these are all, there might be some truth to it, but we're actually going to find out with someone who's managed more than 166 Shopify stores. has a 25-person team, and he actually started his agency during COVID where everyone said that e-commerce agencies were doomed. And while the rest of us are reading press releases and watching Instagram influencers talk about things like Sidekick and Agentic Storefronts that Shopify is working on, Hitesh, our guest today, is watching this AI rollout in real time across hundreds of merchants. So let's welcome— let's start with your story. Began selling click-to-call and click-to-chat services to merchants, and then of course Bigger merchants, and you've been at it since really long. So you've seen the time when there were a lot of platforms, right? There was Volusion, there was BigCommerce, there was Shopify, there was this kind of platform wars. And then I think eventually Shopify has of course taken the lead. So give us some thoughts about that and take us back to the moment when you actually started the thought that, okay, now let's— I need to start my own agency.

When I landed up with my job as a business development representative. So I used to sell click-to-call and click-to-chat services, and that was on behalf of ATG, which was later acquired by Oracle.

Right.

So the things were different, right? You did not have a lot of plugins, a lot of apps that you can simply install on your website and you're done. It was not that era.

Correct.

You had to actually search for softwares that meet your needs and accordingly have them configured through a developer. It was not a DIY approach, which we have today. So it was quite different, right? We used to contact enterprises and tell them we'll increase customer loyalties and all that stuff. Finally, we used to get them to purchase our services. It was that journey. And then we—

This was which year?

This was, this was in 2010, '11.

Wow. Right. And then I would've been in school at that time.

Right. And then the era of SaaS-based e-commerce started. So Volusion came into picture, BigCommerce, Shopify. So like Shopify break all, breaks all the bars, right? So now Shopify has 2 million plus stores. Very good, right? The other were left behind. Part of the story. So, so the thing is like we used to work with all those technologies at that time. We used to serve the clients and the clients were pretty happy. The strategy that kind of Shopify followed was almost like Apple did. So they had a phone plus the app on the App Store. So similarly, Shopify has the website and you have App Store where you can install the apps from.

Right.

So 2 successful companies, almost a similar story, you can say. Yeah, one of the reasons is the App Store, of course, which is the app ecosystem that they have. They have 12,000+ apps and the partner network. So the partner network was very strong, right? I met Virgab in 2014 when he was just starting to expand Shopify in India, right? And I, today I see the kind of efforts that he has put in and the way he has expanded the whole partner ecosystem is amazing.

Seeing all of these platforms grow and then during COVID I think there was a pivotal time in your life, right? You went from zero to managing 166 stores, but most people would say COVID was perhaps the worst time to start something of your own, right? There was so much unpredictability in the market. What did you see at that time that others did not?

So I was switching from one job to another, right? I didn't know whether I will start something of my own. Like, like you said, companies were not able to predict whether they'll get more business or hire more people, so they didn't shut down, right?

Even survive. They were—

yeah, even survive, right? So So they hire, they didn't hire me, right? And so I started this, right? I got into this and I slowly realized that more and more people, so they wanted to go online because they want to sell, they want to survive everything they want to do, right? They want to get into the business again because shutting down their shops, online shops, physical shops was a big pain for them. So that's how the journey started, right? And it's shaped into a company. Company got registered in September. And then it's been 5 years. I've been continuing to help people, providing them the right kind of solutions to help them grow, right?

Don't wait for—

So AI has been around for quite a, quite got a bit of time, right? So it's not a new thing to happen, but these features were quite big wherein the clients can just think of a solution, type it, and get the solution right into their dashboard, right? So it's using Sidekick or any other frameworks that they have launched, right? So it kind of like comes to the agencies, like whether Shopify wants to build more of a DIY engine. They don't want agencies, but that's not the right thing to say, judge, right? Because it's not about— you get the whole toolkit, but then you should be able to know how. That's the main thing. So if you have the experience, you have the right kind of skillset, then you will be able to work with the AI and make things happen. the way you need it. So that's, that's one thing that I, I'm really positive about, right? So people do get the, like, these features with AI, but if they're not able to use it the right way, then it's not going to work. So the agencies that are selling their services are furniture stores, and the agencies who are selling the solutions, or rather proposing the solutions, are the interior designers. So they would first understand what your needs are, understand what your business objective is, and then they'll tell you these are the solutions that are right for your business. And if you implement them in the right manner, it's going to provide you the right results.

Right, correct. So for a merchant, develop the website for them and done.

That's not going to work. So if you're trying to judge someone based on the budget, then it's not going to work, right? But if you are trying to speak to someone, and if you speak to an agency, and if they try to ask you more questions and then provide you the pricing or a proper scope of work with the right quotation, what you're looking for, and then they provided you the solution.

Question. So we were talking about the IKEA analogy and, you know, how some agencies are furniture stores and some agencies are interior designers. And it depends on which one, you know, one specific agency is, and that's probably what's going to define whether they will survive or not. That's, that's a very interesting ideology. Let's talk about merchants using AI, right? So let's say there are 2 stores. One store is doing $5,000 a year, whereas one store is doing $10 million a year, right? Do they use AI differently? And are the outcomes or the impact of these AI tools which Shopify has launched, is it different or is it the same for them?

Different approach. Some of them use AI to validate what they say is the right thing to do, right? Some of them will try to research who the right kind of agency is on AI, right? And someone will say, um, find out what the right solution is and then put up a job somewhere in the marketplaces saying this is what I would need done, right? And yeah, so So people use AI differently for each type of business they are into or each type of different solutions that they require. There's no specific trend, I would say, between someone who is doing less revenue as compared to doing a higher side of revenue. But the clients who are doing a higher side on the revenue, they are quite mature in understanding that AI cannot do everything for them. They will have to hire someone.

Correct. Correct. Yeah, I think, yeah, that makes sense. But give me a little more concrete example of how a merchant tried to use Shopify AI. And initially it looked brilliant, but in production it failed. Yeah.

So I was approached by a client, right? He was in Germany, right? He wanted to develop his website, right? So he said like, I want to develop this, this kind of website. I said, okay, we can do it. And suddenly he vanished. And thereafter he approached me a few weeks later saying, I've already developed whatever was needed. I used ChatGPT to do all of that.

I did it.

So I said, very good, right? So he said, I need some tweaks to it. So I said, okay, let's go on a screen sharing call, show me what you've done, and accordingly I can see what you've done and we can try to optimize it. So now, and I can see there are all sections with custom liquids and each one has static HTML codes and static AI codes from ChatGPT.

Right.

Now, right, I told him that this is impossible to optimize because you've not followed any framework, you've not followed any methodology to develop the solution that is required for this, right? You've just used AI code, HTML, CSS, and pasted it right here. So this is very impossible to work with. She told me, no, like, you'll have to help me. I said, no, we won't be able to really help you. If you need a whole theme rebuilt, definitely we can do it in the right manner.

And that's, I think that's what I've been seeing with the vibe coding as well, right? You can reach to 95% and you feel that, you know, there's just, there are just these 3 things which are not working. And if I'm able to fix it, then I'll have everything ready. But those 3 things will need you to write the whole thing from scratch. Right. So it's great for MVPs and, you know, for showcasing your idea that, okay, this is something which I dream of, but it's still very far from production-ready code. Right. On versus, you know, like I use Cursor to write code. It really helps. But then you need to make sure that it's following the right thing, it's following the right pattern. You need to have that knowledge and then you can use AI to implement whatever you want. It's very interesting that, you know, there's a similar thing on the Shopify AI side as well. What about on the buyer side? You know, like there are, like on the storefront side, there are a lot of new AI tools coming up, right? So what AI features are stores using on the storefront side?

On the storefront side, there are features like AI chatbot that people are using. So we use a tool called Timmy AI that works really well. We can give him— give the AI all sort of documents and give them all the understanding about our product, about the specifications, details, and everything. And the AI understands it and accordingly responds to the customer. So that way you have less number of customer queries coming to you and more satisfied customers.

Very interesting. And but do these AI chatbots actually help increasing like sales conversions? And you know that I've seen a lot of claims, but do they Yes, definitely they help when you chat with them.

Who would, let's say we did this for a store who's selling kegs in Australia. Now with, with the kegs, what happens is like you, you can have a keg fridge at your house, right? So you can first decide which fridge you need for the house, which fridge you would need for the bar, and how many kegs can fit in, whether you need a 50-liter one. Each would fit into the fridge. Now these kind of questions are very difficult to— so they would chat with the AI chatbot. It'll recommend the right kind of fridge and it'll recommend the kegs as well. And average order value is about $1,500, which was previously just $500.

Oh, that's, that's amazing.

Say technology to use, if you use it in the right manner, definitely AI is useful. Yes, that's right. Yes. Right. You have to keep on monitoring what answers AI is giving, right? They, they have the option in the dashboard so you can monitor what sort of answers have been given. And you can optimize and make the AI learn better. There are, right? So there's a tool that we use called Material Retail. It connects with your catalog and it generates metafields based on the product attributes.

Interesting.

That way, if the people are searching for, I need shoes, 7 size, and that's for men, so it will deliver the right kind of results to that search because you have the right kind of metafields and product attributes that are mentioned to the product. Metafields help you with everything. So it helps you with the search on the website, it helps you filter out the products, it helps you categorize the products and all sort of things. So metafields are like, there are different uses of metafields. You can use them when you're using like a reviews app. It's a reviews app will store your number of stars into the metafield. So it's like a temporary field or an additional field on top of the fields that Shopify allows you to have.

Yes.

For customer service, as I said, Timmy AI is the one that we have been using, right? That has worked out pretty well. TDI is also a similar sort of chatbot that works for customer service. Gorges has came up with their own AI version as well, which is kind of working out.

Wow.

So people are slowly getting into AI too. Like, you can definitely use AI when there's tedious work involved, when there is repetitive work. It really works well when entrepreneurs really do some of their work which they don't want to do. People are apprehensive. whether the AI would give the right kind of answers.

That's what I thought.

Yes, that's right.

Why can't I talk to a human?

Yes, and that was one of the features that Shopify launched as well, right? So Shopify launched a human chat button on their support portal because people were kind of not so happy with the AI search and then the human coming into picture. So they launched a button saying chat with the human advisor. Yes, definitely. So there's a tool called Angular AI So Angular AI helps you predict what's going to happen based on your past data, based on the data science understandings that they have, right? And also they offer you a data science customized service if you're going for their enterprise plan.

Yeah, definitely.

So AI plus human is the future, which we, we are already seeing, right? So if you have AI, human intervention is definitely required to understand what the AI has created and accordingly refine it before it gets to the public. That's where I see the whole thing going, right? The repetitive tasks that we are probably doing at the moment, most of them will be automated by then. So you don't have to really go in and optimize your listings. More AI things will do that for you. Then currently some, say, stores also do image resizing, making them into proper shapes and sizes as for Facebook, and then different ones for Amazon and all that stuff. So they don't really have to get into all of that stuff that That work has to be automated and that work takes a lot of time at this point. That is the thing, right? There are people getting into AI-based searches and then people are asking us to optimize their catalogs and their updated. So we are implementing LLMS TXT for the websites that would help us in getting better into the AI search engines. I would say use the AI wisely so it helps you in the right direction. Do take advice or solutions from the agencies that you're working with since a long time. They won't give you the wrong kind of advice. AI is helpful for tedious tasks, right? But for more solution-oriented approaches, I would recommend that approach an agency. Yeah, so I don't have any particular advice for them. They are quite wise to take their own decisions. Just make sure you take the decision right. But you got to understand what AI provides you is maybe right, may not be right. If you have the right kind of understanding that you need to really make things happen, right, then you just have to get it right.

Thank you so much, Hitesh. It's been wonderful. For all our listeners, check out avidbrio.com to see what agencies look like when they evolve beyond just writing, you know, building websites. Thank you. Thank you once more, Hitesh. And for everyone else, Until next time, we'll keep learning from builders.

Thanks so much, Rajesh.