Headless Technology and
The Future of Omnichannel

Jordi Torras:
Welcome to the Future of Customer Service Podcast. We are lucky enough today to have with us Marcelo Lewin. Who’s going to tell us about Headless Content Management. Marcelo, how are you?

Marcelo Lewin:
Excellent. Thank you so much. I have to say I’m lucky to be here. Thank you so much for inviting me. I appreciate it. You were also on my podcast, right? On headlesscreator.com, and that was a great conversation. People really loved it. We talked about AI, and chatbots. People should definitely check that out.

Jordi Torras:
That was amazing. Thank you for now coming here to our podcast and we have a lot of questions. We are very intrigued about the concept of Headless, but first would you Marcelo, tell us about you?

Marcelo Lewin:
Sure. Let’s see, God I’m so old. I don’t think I should start way at the beginning because it’s going to take too long for that intro. But I’ve been in technology, in the web mainly. That’s where I really started back in ’97. Literally when the internet was just starting out and I was working way back then for Toyota Motor Corp, a corporation here in Southern California. I saw that change: going from your regular network to the internet. I jumped on that and I’m so glad I did that because that just changed my life completely. It’s been a wild ride ever since, I’ve gone through the dotcom era, the dot-bomb era, when everything crashed, then the resurgence, and there are so many things that have transcended from way back then to today. It’s wonderful to see how the web is today.

Marcelo Lewin:
How people use it, how people take it for granted. I love that the fact that people take it for granted because that means it’s just part of your daily life. Right? But back then, I remember driving on the freeway with my wife and seeing a sign on a billboard and it said something.com. I forgot what it was, but I’m like, “Look, look, it’s the internet”. I mean, back then, you wouldn’t see that kind of stuff. My wife thought I was crazy, but now she doesn’t. Well, she still thinks I’m crazy. So I’ve been with technology for a long time, switched many iterations. I’ve always been with content creation. I love creating stuff, content meaning anything, right? It could be podcasts, videos, live streams, code. I used to be a coder. That, to me, that’s all creation. Delivering something and seeing people either consume that content or use it to some sort of benefit. So that is kind of like my quick wrap-up of my many years in this world.

Jordi Torras:
Amazing. Thank you. Thank you so much Marcelo. So when I… The first time I heard the word Headless, I couldn’t help, but think about one of the scenes of Harry Potter, when one of the ghosts is hanging around with his head hanging out. I think there’s a lot more than that. So could you tell us what is Headless CMS (Content Management System, I’m assuming) and how is that different from regular or older generations of CMS?

Marcelo Lewin:
Yeah. Yeah, definitely. I had the same reaction than you, to be honest with you, whenever I heard Headless, because I got put into a project saying, look, we’re looking for a Headless CMS. I’m like, ‘that sounds illegal. I don’t want to deal with that’. It’s like, ‘whoa, are we cutting people’? I just, I literally didn’t understand what that meant. Prior to that, I was into this traditional or the monolith CMS, right? I ran a development team. We created a monolith CMS, but when Headless came along and I was asked to jump into that, it intrigued me, and then the light bulb went on and I’m like, ‘my gosh, this is the way it should be’. This is how we should have done it from the beginning. If you think about it, it’s a very simple concept, but it took all these years to get to the simple concept.

Marcelo Lewin:
And the big difference between a Headless CMS and a traditional or monolith CMS is that the head is not there. What do we mean by the head? The head is your delivery channels. That’s why they call it omnichannel delivery. Displaying your content on the website, displaying your content on a mobile app, displaying your content or not displaying, but saying your content, let’s say on an IOT device, right? Those are all the head’s right? Those are the delivery channels. What the Headless CMS does. It separates that delivery channel from the backend, from the content infrastructure, from the authorizing tool and allows a developer to basically pick any framework they want, work in any technology they want, to deliver the content that is stored in the Headless CMS.

Marcelo Lewin:
So it really… You’re going into the API first microservices kind of realm where you’re decoupling all of these things. So you can use the best of breed of what you want for the delivery channel, for the backend, for search, e-commerce, and then we’re going to talk about all that, but you can pick all that. So that’s a big difference, we’re going from a monolith to more of an API first microservices, decoupled system, allowing you to deliver the content anywhere.

Jordi Torras:
Got it, got it. That makes a lot of sense, right? I remember Vignette, a traditional Content Management System, and back in the days.

Jordi Torras:
It was all about putting the content for one specific website that you had in mind and what’s going on now, it’s like, well, you didn’t know that content is going to be, it’s going to be different channels or heads. Now we’ll learn that. That’s amazing. Tell me, there’s headlesscreator.com. It’s a company or an initiative that you were working on. What is this? Will you tell us more about your project.

Marcelo Lewin:
Yeah, sure. So whenever I become passionate about anything, I immediately create, you’re doing the live stream podcast, tutorials, all those kinds of things. Like I mentioned before, I became very passionate about Headless because it just made sense. So I started creating a lot of content, but it got to the point where the content grew, the site grew. So then I decided to create a website called headlesscreator.com. It’s basically a place where you can go for free to learn all about a variety of Headless CMS’.

Marcelo Lewin:
And there are a lot of them from Contentful to Content Step to Graph CMS, to Agility CMS. I mean, there are tons, literally, there are tons. In fact, I’m putting a spreadsheet together of so many that there are to compare, but it’s a place where you can learn. It’s like a school basically where you learn how to administer a Headless CMS, how to develop for a Headless CMS, and how to create content in it. So it’s focused on authors, administrators, and developers. It’s a hundred percent free because it’s thanks to all the sponsors that they helped me keep the site free. So the content keeps growing the sponsors keep coming on in and supporting the site. So it’s a win-win for pretty much everybody.

Jordi Torras:
Oh, that’s a beautiful project. I encourage the listeners now to go to headlesscreator.com. That’s an amazing place to learn.

Marcelo Lewin:
And by the way, it’s early morning here. So I hope you don’t find this rude, but I got to have some coffee

Jordi Torras:
That’s good. You look… I’m not, I mean, you are smart, but with your coffee, you look even smarter.

Marcelo Lewin:
It’s the glasses. If I remove the glasses, it goes away.

Jordi Torras:
That’s a good one. So listen, I was doing some research about the idea of Headless and of course, there is the Headless CMS. Then I found something interesting, which is that Headless e-commerce, and I was wondering, what is this, what does that mean in the space of e-commerce?

Marcelo Lewin:
Right. It’s the same thing as a Headless CMS, right? In terms that a Headless CMS decouples the front end from that backend, Headless e-commerce does the same thing. You’re not stuck to templates that are provided to show whatever template designed for your products, come with it, right. In Headless e-commerce, you can now mix a Headless CMS with Headless e-commerce and then mix both of them up. Because a lot of what you can do today is not only bringing products from your Headless e-commerce but then also bringing content to augment those products. Because a lot of commerce is going towards content-driven commerce, right? Where you learn about certain things and then you end up buying them. So use a Headless CMS to augment your Headless commerce products with content. You can mix it all up and then on the front end, you can deliver it however you want to.

Marcelo Lewin:
Whether it’s on your mobile app, whether it’s on a website, again, whether it’s through IoT devices, and in the future VR and AR. It’s actually already been done. People are delivering content to a virtual reality headset, right? So it’s the exact same concept. I did want to add something. What’s really important with Headless is that you’re not only separating the front end, but you’re separating design from content. The big problem with the traditional CMS’, or the monolith CMS’, is that you had design and content all mixed up together. The problem with that then, and this goes towards something I know you are obviously passionate about is chatbots and AI. But when you have designed with content, your content is not intelligent. You can’t have relationships between different pieces of content, right?

Marcelo Lewin:
Like a blog article written by an author. But if you mix up HTML, it’s just an HTML tag that says, it’s the author name. You don’t really know it’s an author, right? But when you separate the design completely from the content, now you can add metadata to this content and make it intelligent. I think that’s where the real benefit of Headless is when you’re focusing on your content and its intent, and I know we’re going to talk more about that, is when it becomes really, really powerful. that’s a big thing where people are coming from a traditional CMS, including myself as I mentioned, it’s hard to grasp at the beginning because you always want to say, “Well, no, no, I want to put, I want to make this word red”. So let me make it work. Let me make this word red.

Marcelo Lewin:
But then if you really think about it, it’s like, why are you making it red? What does that, what is the meaning? What is the intent behind the red word? Is it important? Oh, okay. So you want to highlight words as important. Now we’re adding metadata to words. Now we know that these words are important. Not that they’re just red because tomorrow the important word may be blue on a different delivery channel. Right. So that’s really important. I wanted to clarify that. That applies also to e-commerce right, where you’re focusing on the content, as opposed to the design.

Jordi Torras:
And as you were telling all these stories, I couldn’t help but remember stuff like XML, right? That was a big… That was back in time was a big thing. Then there was, okay, HTML and then CSS trying to separate format from content. Then the idea of the semantic web that basically you would develop your content using XML, not HTML. So it looked like that all these concepts have been cooking for a long time until this idea now of the Headless CMS, which is amazing. My question is about the channels, how can you use a headless CMS to embrace your omnichannel strategy?

Marcelo Lewin:
With a Headless CMS, you start out with a content model and the content model is what defines your content. Now, when you think about content modeling, when you create the representation of your content in your headless CMS, you got to think big, not just ‘we have a webpage and we need this blog article to appear on the webpage’, or ‘we need this product to appear here’. It will appear there, but you need to think about the product or the blog article itself or whatever you’re modeling and think of the intent and think of not just the intent behind it, but where do you want to see it in the future? Because right now you have it in a blog article, later on, you may want this piece of information elsewhere… Let’s take, for example, a knowledge base article, many websites have apps and well… I’m sure you have a huge knowledge base, right?

Marcelo Lewin:
So customers can run Inbenta properly. So you have a knowledge base article and you’d write this article that answers a few questions. Well, yes, you want to display it right there in that knowledge base article, in the help, right? help.inbenta.com or whatever it may be for you. Right. You want to display it there, but what about inside the actual engine. Do you want to display something there, in context? What about if you have a mobile app, do you want to display it there? What about a tooltip? If you have a tooltip and you want to give more information, would you like to relate that tooltip to this knowledge base article? So you can see how, when you’re modeling your content for Headless CMS, you shouldn’t think of just the one delivery channel that you’re trying to model, which is normally going to be your website, your webpage.

Marcelo Lewin:
You want to think ‘Where else can we use this information?’ and ‘what are the relationships between this piece of content and other pieces of content?’ Do we want to relate a knowledge base how-to article with a knowledge base FAQ article? Do we want to relate a tooltip with any of these articles? So then you start building this web of relationships, and that’s when you start thinking omnichannel delivery, you go beyond the obvious, which is your website, your webpage, whatever you’re modeling. You think of all of these relationships. On top of that, you think about, are we going to integrate this content? Do we want to feed these knowledge base articles to customer support so they can help our customers? Right? So now you’re integrating this piece of content. So that’s why it becomes so important not to type design to that piece of content, because design means nothing when you’re integrating it into Salesforce and delivering it to your customer support agent or when you’re displaying it in a mobile app or IoT. Do you know what I mean?

Marcelo Lewin:
So that’s when you start thinking about all of these omnichannel deliveries is by thinking first about what is the meaning of this piece of content and how do we expose it and where do we expose it?

Jordi Torras:
And then you talk about knowledge articles, right? And you know that Inbenta is very much into knowledge and providing relevant answers for the questions, which at the end of the day, it’s all about content. So the question here is when it comes to customer support, that I’m assuming that this idea of Headless can actually help to build knowledge that has been created to satisfy user questions, to provide as much as possible, as a sales service and the quick and relevant answers. So I’m assuming that’s going to be something very relevant in the space of customer support. So what is your experience or your vision on how Headless and let’s say knowledge base or knowledge-centric customer support, how do you see these two dots connecting?

Marcelo Lewin:
So I think the biggest frustration when you go into a chatbot is when it provides the wrong information or worse, it provides outdated information. So when it comes to Headless CMS, one of the key concepts piece besides separating design from content is also that it is a single source of truth. By creating these relationships. If you say this blog article is written by Marcelo Lewin, Marcelo Lewin is an author. It’s an author content type. When I create a hundred articles, I related to just that one Marcelo Lewin entry, I can update the content, Marcelo Lewin and whatever else, my bio or whatever. Then a hundred articles that are using my name now automatically get the main information, right? Apply that to an FAQ. An FAQ article has 20 questions. One of the questions is ‘How do you manage users in Inbenta’?

Marcelo Lewin:
Right? Well, you create that as a single source of truth question, meaning that there’s one entry that says, How do you manage users in Inbenta? Here’s how you manage users in Inbenta. Now you use that FAQ in a thousand other knowledge base articles, how-to articles, whatever you want to. So now you change how you manage users. Well, you just go to that one entry and wherever else it’s being used, it will be automatically updated. That means that maintenance in content is much simpler in a Headless CMS, because it supports a single source of truth. What does that translate into? Whenever I’m searching as a user, in your knowledge base, or in a chatbot or whatever, I’m always going to get the latest, most updated information. What I find horrible is going into a place that gives you an answer and then you go into the app and it has nothing to do with what it’s telling me.

Marcelo Lewin:
It’s completely off because, in the traditional way where you don’t have a single source of truth, where you would create that FAQ question a thousand times, you have to update it in a thousand different places. So I think that’s one way of making customer support much better, a different way is where you can actually start relating, like I was mentioning before, information. Because now you have relationships of FAQ and tooltips and knowledge base articles and everything else like blog articles, you now say, “Hey, here is a piece of information that may answer your question, but here are some others that you may think will answer the question because it has that”. So I think that’s another way it could help: by exposing the right content at the right time to the right user.

Jordi Torras:
Absolutely. I love this idea of a centralized source of truth. Probably that has happened to you, right? So you go to this website and there is a homepage that says, “Hey, we announced our new offer for XYZ”, right? And then there’s a chatbot and you click on it and it says, “Hey, I can help you with anything”. Wou say, “Well, what about this offer on XYZ?” And it goes: Can you rephrase that? What are you talking about?… I mean, I’m not even mixing channels, here’s the same channel, but it looks like the chatbot is disconnected from the website, which is really frustrating and shocking for users. I’m assuming that’s one of the symptoms of this disconnection. When you have several places and several editing systems and putting all that together, it’s impossible without some kind of Headless strategy. Right?

Marcelo Lewin:
Exactly.

Jordi Torras:
So, and then there’s the moment where you say, okay, I have all this, but Inbenta is all about relevance, search, semantic search, artificial intelligence, natural language processing. So what is your vision about how that content could work? How are semantic search and the idea of semantics playing a role here in the idea of a Headless strategy?

Marcelo Lewin:
Yeah, totally. Well with semantic search, the key is intent, context, and relationships between words, right. That’s how, when I type in something in Google, it doesn’t take into account just the words that I typed in, but what is my intent behind it, and what is the context around those words as well as what are the relationships with the other things I’ve done, right? To bring up the proper result, which normally gives me what I want. Well, those words that I just used are exactly the same words in a Headless CMS, right? You want to focus your content on intent. You want to have context. Then remember I kept talking about relationships between different content types. That all helps with semantic web, right? Because you are storing intelligent content in semantic web users intelligent content to provide the proper results. So I think it’s just a perfect match.

Jordi Torras:
Yeah. Yeah. That makes sense, it’s like, okay, I have my content and now I can deliver it anywhere. I have an intelligent system that is able to find that content. Now about conversation and chatbots. If you look at the tools out there, and Inbenta is no different, there’s a place where it basically says, “Hey, what is the answer for that intent?” Right? You have to program at the end of the day, chatbots are based on pre-built answers. Artificial intelligence is just about trying to guess what is the best answer given a specific question. So my question here is how could chatbots and conversationally benefit from the idea of Headless?

Marcelo Lewin:
Sure. So I’ll talk a little bit on chatbot, AI you’re the pro and people should go to that podcast that we did because that’s not me and even about chatbot, you’re the pro. So I don’t even want to come through as like, I know a lot about chatbots, but at the end of the day, because you are separating content from design and you’re adding intent to content. That’s what chatbot is all about, right. Is having a conversation and understanding not what they typed, but what they want. You can match that up with a properly designed content model. At the end of the day, with any headless CMS project, it all starts with a content model that will support your data, your content needs today into the future, and making that content intelligent. So if you design your content model properly, then you can expose all that information through a chatbot. Let’s say through the conversation, and I’ve seen actually chatbot APIs, that you can create the conversation within a Headless CMS.

Marcelo Lewin:
So you’re actually the author, because at the end of the day, this is the other thing that I think people need to understand. Headless CMS is very much associated with developers, oh, a developer, developer, developer. At the end of the day in a headless CMS authors are the ones that live with this. The authors are the ones that manage content. So you want to empower your authors to manage things, content. In a chatbot, you’re going to have conversations. In the conversation, and you brought up a lot of this during our interview, have sort of some sort of attitude, the company attitude, and right in the way they speak and the way they express, but that’s all handled by some sort of author, right?

Marcelo Lewin:
Somebody has to write some of that stuff. In a Headless CMS, you can empower the author to be able to manage these conversations and the outcomes based on intents. If they ask for this and the intent is this, then show all of these things or answer it in this manner. You could model all that in a content model for Headless CMS, and then expose it through your chatbot. So I think it’s really perfectly made for that. I’d love to hear your thoughts on that too because obviously, you’re the chatbot king.

Jordi Torras:
So that’s the thing, right? We have some customers that say, Hey, we love Inbenta, the way you make artificial intelligence easy by implementing this pre-trained linguistic model. So our customers don’t have to invest time in training with utterances and examples and all of that. But then they say, “Hey, my content is in Salesforce”, right? and ‘I want the chatbot to use that content’. It’s like well, wait a minute, our tool is designed in a way that you enter the content there, but you’re starting to see that. Well, they have another tool in this case, Salesforce, which I believe that’s a head much…! But we start to see the need to have that one source, and have authors working in one environment then say, okay, that is for the help side.

Jordi Torras:
This is for the website. This is for the chatbot. Just have only one place. So we see that more and more. So my opinion is that vendors of conversational AI will start to have to think very soon… Although we are already doing that: ‘How can we basically couple the logic of the conversation and the logic of the content and the answers from the authoring tool in a multichannel context? So I believe that’s coming and at Inbenta we are already working on that.

Marcelo Lewin:
So a headless chatbot.

Jordi Torras:
Of sorts, yes. That sounds more terrifying for somebody-

Marcelo Lewin:
I know it does. It does.

Jordi Torras:
So we see now headless CMS, headless e-commerce. What else can go headless?

Marcelo Lewin:
Well, pretty much everything that you want to decouple of the front end from the back end. Right? So search, I think Algolia is a perfect example where they’re a headless search, right? So you can pass information to it. It returns information, you display information, however you want. So really anything can go headless because what it’s doing, it’s decoupling the logic and the content from the back end in the design, as we said, from the front end delivery channel. That way you can expose it wherever you want. Again, it’s omnichannel, omni-device delivery.

Jordi Torras:
Got it. It is related. Right? So the idea of Headless and API is somehow related as well. Right? So when there is now stuff like GraphQL, right? The way a way to kind of organize APIs, and we saw the Restful and then GraphQL. How is it all about connecting with data?

Marcelo Lewin:
Most Headless CMS’ right now will offer you an API. An API, which most people know is an application programming interface, is a way to get into the content that’s stored in that headless CMS. All of them will let you get to that via Rest. But many of them also give you Graph QL, which is an initiative, started by Facebook many years ago, to speed up the delivery of content into mobile devices. Right? So that way it would load up much quickly, and you would ask exactly what you want, and it would return exactly what you want without having to change your endpoints like in Rest. If you had one, a different kind of piece of information, you have to create a new endpoint in Graph QL, and an endpoint is a point in a location where you say, “Hey, go there and give me all the users”.

Marcelo Lewin:
Well, it may return all the users with all the information about that user, but really all I want from it is to give me all the users’ first names. I don’t need their last name, their address, and all that kind of stuff. Because I’m just going to say, “Hello, John”. Well, Graph QL allows you to say, “Hey, give me all the users’ first names”, that’s it. It will return all the users’ first names. When in Rest, you could do it, but it requires programming and Rest is kind of tied to that front end, because depending on what you want on the front end, you tell Rest what you need. Where Graph QL is more generic. It says, just give me what I want, it returns it and that’s it and you’re done. So many, many people… Remaining Headless CMS’ now have the ability to get the information via the Graph QL, which is another way of getting that information.

Jordi Torras:
It’s an API, but instead it gives you all the information because that’s how the API has been set up. Well, you tell me what to infer. So QL comes from Query Language, right?

Marcelo Lewin:
But Graph is an important term because it’s really the relationship between all users graph, right? It’s all the content related to this user. See, we go back to relationships, which is the basics of Headless CMS with content modeling. All of this ties together at the end of the day. Right?

Jordi Torras:
Correct. Absolutely. So that’s fascinating. So I’m wondering, tell me, what do you think the future holds in terms of headless technology and the Headless CMS?

Marcelo Lewin:
Sure. So I’ll tell you what I would like the future to be since I don’t have this magic lamp that will tell me. I wish I did. One of the benefits of the old CMS’, the monolith, was that the head, the front end, made it very easy for authors to manage content. With a Headless CMS, the authoring tool, although many people are striving forward in creating a pretty good authoring tool, it is still not up to par with where the traditional model’s authoring tools are. It was very easy and very simple for an author to go in and create content. They would see it immediately, and most of the time they would edit it right on the page. With headless CMS’, because you are disconnecting the front end, whenever you want to do any kind of live editing, you have to involve a developer for that. That connects that front end to your backend and so on.

Marcelo Lewin:
So with Headless CMS’ we found we’re about 90 to 95% there with content, infrastructure, developer framework. All of that is there for developers, but for the authors, I think we’ve got a while to go. To make the experience great for an author, which at the end of the day, we got to remember, these are the people that live with that Headless CMS and the content day in and day out, every day, right? The developer’s done, they move on to something else, but the author has to live with this.

Marcelo Lewin:
We have to improve that authoring experience and make it less abstract. I love content modeling, but sometimes you can get a bit too abstract. From the author’s perspective, they get lost. They’re like, well, I don’t understand the relationship between this screen and that screen. So we need to work on that authoring experience. I’m hoping that as an industry, all the Headless CMS’ strive towards improving that authoring experience, so it’s much easier to maintain content from the author’s point of view. That’s what I hope the future brings, a much better authoring experience. As I said, many are moving in that direction, they’re doing a great job, but we still have a while to go, I believe.

Jordi Torras:
All right. Yeah. That makes a lot of sense. It’s a very pragmatical vision. At the same time, really complex, because here’s what you said, you have the authors, these guys know everything about content, but they’re not technical. Right.?Then the developers that they know about coding, but not much about content. So how can we make these guys live together and be more productive? That’s indeed the missing part here. That makes a lot of sense.

Marcelo Lewin:
Yeah, totally.

Jordi Torras:
All right. Marcelo. I could be here asking you questions about Headless and CMS’ for hours but we want to make that podcast consistent when it comes to length. What I would like to ask you is how could our audience find you, in case they wanted to know more, and learn about headless CMS, and talk to you. What should they do?

Marcelo Lewin:
Yeah, definitely just go to headlesscreator.com. You can create an account there. You can reach me if you want. I always tell everybody to reach out to me at [email protected]. Many people put two L’s, but it’s one. So they can reach me via email, or just sign up for an account. Feel free to reach out to me on Twitter, LinkedIn, it’s headlesscreator on Twitter, LinkedIn, Facebook. Yeah, just search for Marcelo Lewin. You’ll find me. I’ve been on the web for many years, so you’ll find something.

Jordi Torras:
All right, Marcelo. Thank you so much. For the audience, thank you for listening to this. We have all learned a lot about CMS. I want to thank Marcelo again for being here with us, and now we are working already on preparing our next episode of the Future of Customer Service. Thank you so much. Have a good day.

Thanks so much for tuning in. This podcast was brought to you by Inbenta. Inbenta symbolic AI implements natural language processing that requires no training data with Inbenta’s extensive lexicon and patented algorithms. Check out this robust customer interaction platform for your AI needs, from chatbots to search to knowledge centers and messenger platforms. Just go to our website to request a demo at inbenta.com. That’s I-N-B-E-N-T-A.com and if you liked what you heard today, please be sure to subscribe to this podcast and leave us a review. Thank you.