This week on the Future Ear Radio podcast, I’m joined by Anna Rosen, Enterprise Solutions Engineer at Voiceflow. Earlier this week, Voiceflow debuted a big new update to its conversational AI design and prototyping platform, called V2. So, in order to fully breakdown the big update, I had Anna on to share with the listeners and I all the key new features included in V2, and how this update transcends beyond Alexa/Google and the smart speaker ecosystem, into a full suite of conversational AI channels and modalities.
Voiceflow enables teams of conversational designers and developers to collaborate remotely, in real-time or asynchronously, to prototype, design and deploy conversational experiences. It’s both a no-code tool that requires zero-coding ability, as well as a full-service developer platform with sophisticated feature sets and capabilities. As the video below illustrates, it’s where these conversational experiences are brought to life and deployed.
What’s so exciting about Voiceflow, both the tool and the company, is the speed at which it’s innovating, which was on full display during V2. In just a few short years, the tool has become one of the primary platforms that conversational designers and developers rely on to build their experiences (just ask all the Amazon & Google employees who use it themselves). In many ways, Voiceflow represents the future, which is a heavy dose of remote collaboration, be it in real-time or asynchronously, for a swelling workforce that’s dedicated to making our interactions with computers more ambient and conversational.
My Takeaways:
- One of my favorite points that Anna made was around this point of creating a “unified” language between developers and designers. So much of what was revealed in this update was based around features that provide for better communication between these two disciplines. For example the “canvas markup” feature is pretty neat, as it lets teams add labels, images and hyperlinks directly into the design template. In a remote-working environment, communication is paramount.
- Voiceflow CEO, Braden Ream, was included in Voicebot’s 2021 list of predictions from people around the industry. His prediction: “We’re going to see two major trends continue in 2021: the merging of chat and voice as one ‘conversational product/AI’ team within enterprises, and the continued rise of independent assistants.” I strongly agree with this statement and it’s obvious now that he was alluding to Voiceflow as being the tool that caters to this merging of chat and voice.
- I loved the bit of our conversation around Smart IVR (interactive voice responses). Everyone reading this has probably had a poor experience with an IVR system (press 1 to do X, press 2 to do X). IVR systems can be designed more efficiently and effectively through Voiceflow, and can be made more intelligent through the context-switching abilities of conversational AI. Talk about transforming an old legacy technology into something new.
- In my opinion, context is at the heart of the next phase of voice assistants. The more contextually aware these systems become, the more natural we’ll be able to speak to them. The more natural it is to interface with computers through our voice and language, the more compelling it will be to migrate various “jobs” that we do on our phones today. Contextually aware assistants will enable the combo of hearables + voice to truly shine.
- One of the biggest announcements at V2 was the Voiceflow SDK (software developer kit). The developer community seems to be really fired up about this and understandably so, as the SDK will allow for an entirely new level of customization and integrations into one’s voice experience. This will really unleash a wave of creativity for developers and open the door to entirely new applications and types of voice experiences and integrations.
-Thanks for Reading-
Dave
EPISODE TRANSCRIPT
David:
Okay. So we are joined here today by Anna Rosen. Anna, tell us a little bit about who you are and what you do.
Anna Rosen:
Hi, thank you so much for having me. Yeah. So my name is Anna Rosen, I am an enterprise solutions engineer here at Voiceflow, and I help enterprise clients figure out how to solve the problems that they’re trying to build for, whether it’s custom drive-through assistants, voice banking, smart IVR systems, or anything Google and Alexa experiences.
Anna Rosen:
So, Voiceflow is a design build and prototyping tool for every conversation, like I mentioned, doesn’t matter the output, whether it’s chat IVR or a custom intent assistant.
David:
Love it. Yes. Voiceflow. We’re talking about Voiceflow today, Anna is there, and honestly, one of my favorite companies that I’ve been following really in any of the spaces that I cover, I’m cheering hard for you all, I think it’s just such an awesome sort of way that you all are tackling this space. And we’re going to talk all about what you all just announced yesterday.
David:
It is February 23rd, so on February 22nd, you guys released Voiceflow 2. Big update that I know that you’ve all been working on for over a year and it did not disappoint in any way, I think this is just so cool to see. I think one of the most important tools as it relates to the evolution of the voice space kind of being built in real time before our eyes.
David:
And so, I’ve got connected with you, I think it’s really cool that you have this background in both developer, but also in enterprise sales. Because I think that it’s always really cool to hear about what is resonating with businesses. What are the enterprise accounts thinking about? They’re probably going to be on the forefront of ways that these types of experiences get deployed. And some of the big things that we see in the market initially as this space continues to progress.
David:
But before we jump into Voiceflow and Voiceflow 2, share with us a little bit about how you even got introduced to Voiceflow, what led you there and what your time has been like in the period of time that you’ve been there?
Anna Rosen:
Yeah, that’s a great question. So I’ve been aware of Voiceflow for a number of years, back when I was at an agency building mostly voice experiences for clients all over the world, mostly in North America, but certainly across the pond as well. And I looked into Voiceflow at the time, but it wasn’t what I needed for what I was doing.
Anna Rosen:
Since then they’ve done a completely revamped code base, so that tool that I looked at all those years ago is not the same as it is right now, and certainly not offered today with our latest releases. But essentially I stayed in touch with Braden. I think I gave him some feedback about the tool back then, he tried to hire me and then he stayed in contact for about a year before I agreed to join the team.
David:
He finally got you. Yeah, the thing that I find really cool about Voiceflow is it used to be Storyflow, right? And so they were building turnstile story flow. You would have a create your own experience, and then if you choose this, then it would open up a new path and that would then lead to a different part of the story.
David:
And so each path would lead to different ways, but what’s fascinating is, they sort of, I think as I understand it, realized that the bigger opportunity was around how you actually create these experiences. I think they ran into a lot of issues initially in terms of the development of their own skills, and they saw that if we’re having these issues, then there are probably a number of others that are as well.
Anna Rosen:
Yeah, that’s certainly how they got into the space. And it was originally this really great no code tool to service the smart speaker industry. And that’s part of where we got our start, which is, it’s still something that we do. And even with the latest updates in V2, we’re always aware of that part of our user base, making sure that we’re still taking care of them in terms of the needs that they have.
Anna Rosen:
But it actually transfers really well into where we’re focusing now, which is on more enterprise deals and on extension beyond Alexa and Google smart assistants. Because the ability to design in a sleek environment, in a way that’s easy to understand and where you can collaborate with your team members and people can jump on and test things and give feedback. It works really well for teams as well, it works really well for enterprise experiences, it works really well in a COVID world where everyone’s remote, and… there was one more point I had. It’s especially important for non-technical folks, which are members of design teams, conversation teams, UX folks, PMs to be able to jump on this tool and not be so overwhelmed by, Oh, it’s too technical, I can’t understand it. That kind of thing.
David:
Yeah. I mean, for myself, I am that person, I am not technical, I don’t have a developer background. But as a Testament to Voiceflow, I was able to create an account, get in there, and it’s just very intuitive. Like you said, it’s kind of part of the no code movement which is really intended to be something where you can, through building blocks, kind of like Legos almost you can create something. And underneath the hood, what you’re really doing is you’re taking pieces of code and you’re more or less designing your own experience without having to actually go in there and code.
David:
And that really resonated with me, I mean, that was the big thing for me, where I was like, this is a really powerful tool that in my opinion, the formative stage that we’re at right now with all of this, particularly the voice assistant experiences, whether it be Alexa, Google assistant, I think that in order for that to really I think take off, you will need to enable small businesses, all kinds of different enterprises that don’t traditionally have big budgets that they can allocate toward something like the blue chip fortune 500 companies have the budget and the wherewithal to do, kind of similar to I think the web in the early nineties, I think that you all are building something that allows, and it kind of democratizes the enablement of the creation of all this.
David:
And I think that as this space progresses and as it becomes, obviously we’re already seeing a sort of proliferation of all the different devices that support these types of experiences. So then the next thing is, you need to create the supply side, all the different modalities are in existence now that can be a conduit to these experiences, so then it’s a matter of how do you make those experiences easy enough to generate so that you get this big supply of them.
David:
And that’s what I think that you all are really doing, and that’s what was on full display with Voiceflow 2, was this ability that really anybody I think can create something in there and I think that’s really cool.
Anna Rosen:
Yeah, it’s about access, it’s about empowerment, and one of the things I love about Voiceflow is how we’ve been able to do that for so many people for so many different reasons. So we have users that are in their teens, who are students, but I was chatting with one of our users who’s 65 and she’s building a Google auction, and it’s nice to be able to see that large of a spectrum of users.
Anna Rosen:
And Braden mentioned it yesterday, but one of our users was able to buy a car with all of the extra money that he earned building Alexa skills on Voiceflow, and just being able to see that and have that feedback I think is definitely a big thing for the team.
David:
Yeah. And not to mention that he also said that you have developers in a hundred different countries now, and I think that’s another really big point here too, is that this is something that is not limited to a handful of markets, this is something that is being widely deployed and I think people are finding lots of creative uses for it.
David:
So, I think that’s a really, really cool thing too, and to go back to what all you were displaying and you said, this evolution of moving beyond just Google Actions and Alexa Skills, I’ll link it in the show notes, but that showing versus telling video that you all created was so powerful.
David:
Just to kind of give the audience a little bit of an idea of what I’m referring to here, basically it is on full display showing that you can kind of like prototype in real time the building of all these different channels, so you think about Alexa, Google, Siri, Microsoft teams, Twilio, Facebook Messenger, SMS, Slack, right? And then you have all the different modalities that these things can be ported to. So you have cars and smart speakers and phones, right? And the different apps contained in them.
David:
So, what was really interesting, so it shows somebody kind of in this two minute, really, really well produced video, building a banking skill, so it’s like the key use cases check my balance, right? And it’s showing how you can basically produce this in one fell swoop for all these different channels. And then it shows how it would appear on all these, so, you can check your balance while you’re driving, you can check it on your phone, you can check it in Facebook messenger.
David:
And that’s again where I think this is getting to be really interesting, it’s going beyond voice, and it ties into what Braden said in his voice spot 2021 prediction where he said, and I’m sure that this was all on his mind, because he’s like this is what we’re building, but he said, we’re going to see two major trends continue in 2021, the merging of chat and voice is one, conversational product AI team within enterprises.
David:
And that’s kind of the way that I walked away from this too, is it’s like, that’s what’s happening here is you have this merging of more or less computer languages, whether it be natural language processing, or it’d be the chat bots and all that. So that in a very natural, in an increasingly more natural way, I can communicate with these things and they’ll understand what I’m saying, and they can communicate with me back in these different channels, and that to me is like, this is, like he said, this is a really bigger thing than just voice, this is more about kind of a conversational AI suite of services.
Anna Rosen:
Yeah. And it’s all powered by one centralized tool, that being Voiceflow. And the amazing part about that video, which was my favorite part as well, is that Tyler actually built all of those different channel outputs, so that was real, it wasn’t even just a production trick. He went ahead and did it all just because he knew he could.
David:
I love that.
Anna Rosen:
Yeah. So I think it’s amazing. And it really blows your mind when you start to about all of the different possibilities that you can use Voiceflow for at this point.
David:
Yeah. I completely agree. I mean, so going back to this idea of one platform for it all, I think this is the really important thing for everybody to understand is that, this tool is, I think at the heart of it, it’s a prototyping tool and it’s a collaborative thing, just like you said, that we’re all working remotely.
David:
Again, I think there’s a bigger picture here, and I was so struck by this, during this yesterday where it felt like such a representation of the future of work, because you see the way that you can work and collaborate in real time. Like we’ve all started to get accustomed to with some of the different Google suite of services, when you’re in a Google Doc with other people and you can kind of just see each other working in real time.
David:
I just thought that’s really fascinating, where you might have a team that’s comprised of two designers and a developer or something like that. And you’re basically within this tool, you can be designing together, I saw that there’s this canvas markup feature that you’ve added, where you can add labels, images and hyperlinks as you’re going.
David:
So again, you can really communicate together in I think a really efficient manner, and that’s I think the name of the game here is speed and closing as many of the communication gaps as possible. And I think that this idea of being able to in real time, sort of just like on the fly create these things and modify these things and just consistently be refining them as a team.
David:
Again is, I can see this as being just such an important tool, for what again as Braden described, this conversational AI new division that I think is probably going to be built within a lot of these different enterprise organizations, and we’ll probably see a trickle down and it be a focus much in the same way that a lot of the way that the web went and SEO initially became something that enterprises were focused on and then everybody became focused on it. It became wide widespread knowledge of, you need to kind of understand the way in which these search engines work, because that’s the way that the user behavior is kind of set.
David:
So, that to me was such a big part of this, and I’m just curious to hear your thoughts, that designing and the advancements that you guys I think are making in that space is just absolutely fascinating to me.
Anna Rosen:
Yeah. So there’s a lot to unpack there, I think the last part that you were touching on, I mean the big guys, they always have the deep pockets to move faster, that’s the reality. But for the little guys, it’s an opportunity for you to see what the big trends are and where things are moving when you start to see the bigger companies jumping on certain trends.
Anna Rosen:
So it’s important to pay attention to the fact that so many large fortune 500 companies are starting to invest in conversational AI and everything like that. Mostly working in the dark around what they’re actually about to launch, but yeah, important to pay attention to that stuff.
Anna Rosen:
In terms of what you were talking about before, it’s a design and prototyping tool at heart and really what it’s about is creating this unified language between designers and developers, which is historically a relationship that always has this gap. There’s always this sort of misunderstanding between what I want to create and what’s actually possible.
Anna Rosen:
And speaking from my own experience as a developer, not using Voiceflow to design conversational experiences or starting to build them, I was looking at a mishmash of multiple tools, I was having to look at many different documents in order to understand exactly what the designer wanted me to do. This was pre COVID. This was us sitting beside each other for many months, being able to talk all the time, so it wasn’t really an issue.
Anna Rosen:
That’s not really possible anymore, nor is it really the best way to do it anyways. So when you have something like Voiceflow, I think it’s important to be open-minded about a different way, because I find a lot of the time you have designers and even developers who are very hesitant to move away from the modality that they have created around these symbols and these shapes and these colors, so they’ve sat down in a room together and they’ve decided that this decision diamond means X, and this is an API call and green and dah dah dah.
Anna Rosen:
But the whole point is that you had to design this language that’s not even universal, it’s only in your company and perhaps only in your department in order to communicate something that voice flow does with a prototype, because at the end of the day, the communication gap largely is due to how the logic flow of the conversation is going to work.
Anna Rosen:
And so on Voiceflow because the designer is able to still have that creativity within the confines of structure of what’s actually possible to develop, you’re able to create something that’s actually executable code, and you have all these other amazing features in terms of canvas markups for stakeholders making sure that things are shareable and consumable by people who have no idea what conversation design is. And then of course, the advanced features around conversation design, specifically around state management and context, which is one of the most challenging things from a development perspective, and one of the most necessary things in terms of a great design perspective.
David:
Yeah. I mean, I hadn’t really thought about it the way that you described there, where it’s creating a unified language. And that to me seems extremely, again, it kind of transcends just this, it’s I think something that speaks to a much broader theme, which is voice in the conversational AI space, I think it’s almost, because it’s so nascent and because it’s something that is, it’s all being built as we go, and it’s not inundated with a lot of the legacy systems in the methods and all this, we get to kind of start from scratch.
David:
And I find this to be so interesting that this appears to be one of the first things that is recognized as being maybe a legacy issue that we’ve had in the past, right? Where part of the issue with any sort of development and design comes from that communication breakdown between designers and developers, and being able to create a tool that you are able to collaborate together using a more unified language, I think speaks to this broader theme of we’re getting a second chance or a third chance, or a fourth chance, however many times you would cite that we’ve been through these development cycles. I just find that to be more representative than just this particular space and more of a representation of kind of like a rethink of how we should be doing these things in general.
Anna Rosen:
Yeah, it’s interesting as well, because it’s such a young industry, however, you still have certain dichotomies around what people prefer in terms of how they design. I mean, I remember I was on a panel last summer around this whole design development experience for conversation design. And one of the questions was, how closely do you work together?
Anna Rosen:
And as a developer, I was like, very closely because of everything I said before, in order for the two of us to put together something that was actually what the client and the designer wanted, that communication had to be flowing. But all of the other panelists answered that it was, I finished my design and I give it to the developer, and then if they have comments, we talk about it the next sprint.
Anna Rosen:
And I always thought that was an ineffective way to go about it, and it’s certainly not something that we think happens at Voiceflow, because of the nature of collaboration that I believe is necessary when you’re working with this kind of app develop… or this kind of design. So, yeah.
David:
No, I agree, I think fascinating. So, clearly a lot of the people that are listening, I think the folks from the voice space are familiar with some of the conversations I’ve had with voice, have a general idea of some of the different ways that people are designing and developing Alexa Skills and Google Actions. I’m really curious about some of these other channels that you’ve all really highlighted and emphasized.
David:
So IVR, I think this is a really interesting one, interactive voice responses, I think anybody listening has interacted with an IVR system, it’s like press five for the next menu, so it’s basically just a big layer of different menus that you have to dial into. And it does seem like this is a really ripe area for a better way of doing things.
David:
And I think that conversational AI probably is a perfect match for this, so can you just share with me, what’s going on in the IVR space? Why is that becoming an emphasis of you all and how are you all looking at IVR as part of your total suite?
Anna Rosen:
Yeah, so IVR is certainly an area that is ripe for disruption and certainly an area where most of our enterprise clients are coming in from. So one of the biggest examples is banking or finance, because they historically have these very large legacy IVR systems that offer, not so great customer service, I think that’s kind of what they’re known for, but the real reason that they’re coming and looking at Voiceflow and conversational AI is because of the cost savings that are available to them if they’re able to streamline the conversation such that the experience happens faster and more seamlessly.
Anna Rosen:
So, to give you an example, like you touched on, typically an IVR system has a lot of different menus and you go into sub menus and by the time you’re even at the place that you want to be, you’ve probably been on the phone for seven or eight minutes already. Average talk time is probably between 14 minutes to 28 minutes and that’s on a good day. I’m sure you’ve been on call waiting for longer than that.
Anna Rosen:
So, there’s this large opportunity here in terms of reducing that call time and also in terms of reducing fulfillment to an agent. And so if we’re looking at and thinking about KPIs for IVR systems, we want to get people to the resolution that they want, customers want to solve most problems on their own. And most of the time we want to deflect away from them having to talk to the agent anyways.
Anna Rosen:
So, there’s a couple of different ways to do this, and we’ll talk about first how you would reduce that sort of a menu experience. So instead of you having to press those buttons and going through that very, I tell you how you can proceed experience, you would change it to something that’s more conversational. And that would be something like Dave, you would call in and the assistant would say, how can I help you today? And then you would be able to say something like, I want to check my balance. Which happens to be one of the number one call drivers in Canada for Canadian banks.
Anna Rosen:
So, that kind of call driver, so a call driver for those that don’t know is the reason that the user or client is calling in. That is a example of a situation that can be easily handled by AI, right? I can tell you what that balance is, I can authenticate you, and I can get you to that resolution. And all of a sudden your call hold goes from 28 minutes to four minutes and you’re happier. And so that saves the bank on average in Canada about a dollar a minute, if they get 44 million calls a year, that’s going to save them quite a bit of money.
David:
I love that. I mean, I think again though, this ties into the broader suite, right? And this goes into what that showing versus telling video displayed was that this can be part of a much larger suite of channels. And I think this is what’s so compelling to me is like, just as you said, you might have, I mean, clearly if this is one of the main reasons that people are calling into their bank, there’s lots of different ways that you can check your balance today, right? You can go onto the website, you can use an app, but still a lot of people are preferring to use this, rather traditional method of doing it.
David:
And I just think that the ability to be in more places and provide that the better. And I think that this is what’s going to be really interesting to see is that, this idea of Chatbots, IVR systems, mobile apps, Google assistant, Alexa, all these different things, I think they really will be able to perform that same function. And this is again what speaks to the power I think of Voiceflow is being able to do it all in one platform and create all of that and prototype all of it and see okay, so we’re creating this, what’s my balance function and everything’s running smoothly except the IVR one. And you can see in the actual tool, you can run these diagnostic tests to figure out why is that particular channel broken.
David:
And again, as a person that is sort of an outsider here, that to me strikes me as a massive driver of efficiencies, this seems like just a huge advantage in terms of the speed and the accuracy, and ultimately just the wholisticness of being able to do this all in one.
Anna Rosen:
Yeah, it’s having this… It’s almost like a hive mind or this like brain that powers at all, but it’s not being developed in this monolithic experience where it comes from this life central place, and then everything comes out of that. It’s more that Voiceflow is this workspace where you are able to have all of those different experiences, all those different channels. And you can reuse across channels where it’s applicable, but then you also can customize across channels where that’s necessary as well.
Anna Rosen:
But to your point, there is opportunity for multimodal experiences across channels, there’s opportunity even with IVR to tap into multiple systems and kind of do data dipping or whisper affects as well. So if I know that you were just trying to reset your password on the website, and then now you’re calling in, they just have a lot of information on you, they know that kind of stuff. And so then the IVR system can adapt and say, Hey, Dave, I noticed that you are trying to reset your password, is that what you’re calling about today?
Anna Rosen:
And then again, getting you to fulfillment so much faster, you’re happier, the bank saves money and everyone walks away happy. And so IVR smart IVR is definitely one of the biggest trends that I’m seeing people come to Voiceflow for.
David:
Yeah. I mean, as an enterprise sales person, you would be the one to know what are the things that are resonating with these big accounts? I mean, are the first conversations I know that we’re talking about banks right now, but in general, I’m curious, are some of these accounts coming to you all to solve IVR things, chatbot things, are they looking at it more holistically and saying kind of to Braden’s point the conversational AI suite, I’m curious, where they’re beginning in terms of the way they’re thinking about this.
David:
And then as the conversation progresses, are some of the different accounts that you’ve been working with having a higher realization of this is bigger than what we anticipated initially, and it’s a much bigger project or there’s much more potential here, something like that.
Anna Rosen:
Yeah. So most folks come to us for a better design experience, it’s typically that they’ve run into some issue with another tool that they’re using and they kind of like what Voiceflow looks like, and they’re looking to learn more specifically just about design. And then once we get in there, we start to identify what some of their problems are. Then they start to ask all the questions, can you do to this? Can you do to that? And can you do this? And then that’s when all of their ideas start to go all over the place and it always ends up leading to more than just design.
Anna Rosen:
So, it depends, but that’s kind of what I’m seeing, and it’s interesting because it also is where they’re at in their journey of, if we’re still talking about IVR, how modern their IVR system is becoming. So some IVR systems in Australia can go as far as to do a full refund on an IVR experience, which is not something that you’re seeing in Canada yet. And so there’s all sorts of different places that different companies are at in terms of how smart their IVR systems are becoming.
David:
Love it. So let’s talk a little bit more about Voiceflow 2, that big announcement yesterday. Tell me what your favorite aspect that you all released was, what did you all unveil yesterday that you personally love the most?
Anna Rosen:
Yeah, so aside from the big reveal in terms of our SDK, which just allows developers that flexibility to code in such a modular way and kind of plug and play the different kinds of tech stack that they need in order to fulfill whatever it is that they want to fulfill.
Anna Rosen:
I would say it’s visual prototyping, I think that’s definitely something that a lot of our users are really happy to hear about, it’s something that they were asking for, that we are listening and we really came through for them and everyone was really hyped up about that. So to give you an idea, we released all sorts of new customizable options on your prototype, so now I can send you this prototype, maybe it’s a drive-through and I can also include the visual about the different things that you’ve ordered, and I can customize it based on screen and size, and I can also customize it based on language, and I can send you a specific part of the conversation instead of sending the whole thing.
Anna Rosen:
So there’s all sorts of ways in which we’ve improved the experience of sharing and gathering feedback, but then also in terms of user testing, which is a big thing for conversation designers.
David:
Okay. That’s really, really interesting. So let’s talk a little bit more about this. So, you’re saying that visual design, so it’s helping teams to more or less better articulate what they’re trying to… I’m trying to solve this particular design dilemma, and here’s just a whole lot more robust information as to what I’m trying to describe to my fellow teammates?
Anna Rosen:
Yeah, it’s essentially a one-to-one experience of what you would be experiencing if it was released in the wild, so you’re getting that same feel of here’s a replica of this experience. Whether it’s from [crosstalk 00:32:37]-
David:
Love that.
Anna Rosen:
… stakeholders typically are a big one, they get the shareable prototype and then users, so I’m testing this IVR experience on you, I want it to actually look like an IVR experience so that I’m really setting the scene for this realistic testing scenario so that I can gather the data that actually is relevant to me. That kind of thing.
David:
I love too that you’ve used the term stakeholder, I think that’s again, from wearing your enterprise hat, this is really interesting too, because I think that as this becomes I think more relevant in time. And I think that there will be a lot of demand in terms of people in departments, in corporations to legitimize these things.
David:
And so if you can make it so that they’re easier to just like you said, so it doesn’t look like it’s like some prototype and it looks more kind of like the finished product, I think that’s really enabling people from an internal standpoint within their organizations to validate and legitimize the projects that they’re saying, like this is what we need and this is what it would ultimately look like.
David:
I think for the folks that are not immersed in these worlds, the stakeholders that might be the powers that be that have the decision-making powers might say that, Oh, okay, now I see exactly what you’re trying to articulate to me. And so therefore it just further enables their ability to get these things passed internally.
Anna Rosen:
Yeah. It’s how you get internal buy-in, and I know from working the other side of this relationship, you do typically have to demo every couple of weeks, depending on the relationship at the end of sprints, and before you’re at a place where you can even, let’s say we’re not using Voiceflow, I’m doing it all myself. It takes some time to put a prototype together, you’re not always able to just demo the first four weeks of the project if it’s just very basic stuff.
Anna Rosen:
So, being able to bring stakeholders in and have a prototype that they can touch and test and feel and hear and actually see what’s going on, it does bring them into the experience a lot more easily as opposed to creating some sort of adhoc, automagical, here’s a video of what it would do when I’m done developing it, versus here’s an actual prototype of how it really works.
Anna Rosen:
That again, in terms of improving communication is a really important piece, because it’s one thing when you’re knee deep in the conversational industry, in the conversational world, it’s quite another when you’re just a person who needs to kind of give the approval and go on with your day, we’re playing inside baseball all the time, and you do need to be able to present something in a consumable format for those that aren’t, which is also where the markup comes into play in terms of being able to add these little touches and pieces to just help that communication.
David:
Yeah. That’s really well said. Okay, so continuing on, what else in this updates really stands out in your mind, maybe what’s something that was flying under the radar a little bit that you think is actually quite important or really will resonate with people as this all becomes something that people will start to really use?
Anna Rosen:
Yeah. So bringing it back to design, I would say flows. So flows was something that was in the design announcement, and the reason that that’s important is because it just opens this of box of possibility when it comes to how advanced your conversation is able to be. So flows is a way that you manage the conversation and you manage context, you can kind of think about it like functions.
Anna Rosen:
So, if I’m building an order pizza flow, let’s say to seize that little classic example, I can have all of the different secondary flows that are part of that experience, like what are your store hours and help and payment flows all happening in a different canvas, so that everything’s kind of like modular and tucked away. And the reason that that’s important, one, is because you can have multiple developers building on different flows at the same time, you can have reusability across flows and across projects, so I can build this, here’s how this help experience works for this kind of drive-through experience and I can reuse that across different experiences.
Anna Rosen:
And I can test it, so I can test just this chunk of the project that I want to test as opposed to testing the entire thingy, having to start from the welcome and go through all the way, which if you’ve ever done user testing without a tool like Voiceflow and you make one wrong step [crosstalk 00:37:43] over and it ends up taking you so long. So, flows are a really great way to just improve, like I said, how articulate and how advanced a conversational AI is able to become. And then from the team’s perspective, it’s great for testing and it’s great for usability.
David:
Yeah. Love it. All right, let’s do one more. One other thing that you all released that you think we need to draw attention to?
Anna Rosen:
Yeah. So I don’t know if I’ve already talked about the SDK, but I think-
David:
Yeah, let’s talk about the SDK.
Anna Rosen:
… landing on the SDK, which has definitely been flying under the radar for a while, we do have a couple teams using it. It’s just the list of opportunities are endless, so you can still use it if you did want to use it for Google or Alexa, but you can also use the SDK when you’re building for in-car assistance, you can use it when you’re building for Microsoft Teams, Facebook messenger, whatever it is, whatever channel that you want, it just gives you the ability to literally do whatever you want and add all those customizations in, add all of the different analytics, add whatever NLP if you want to swap that out. And I think that’s important, because you do want to offer developers…
Anna Rosen:
You don’t want to take away their creativity, just like you don’t want to take away the designer’s creativity. And you want to be able to support them through what it is that they need to build for their specific reason. And because conversational AI, you touched on this before? It’s applicable across so many different verticals. Anytime you have a situation where a customer is interacting with your company, that interface is basically an opportunity for conversational AI.
Anna Rosen:
So, whether we’re talking about drive-throughs, manufacturing, finance, anything, you’re going to have different technical needs, and the SDK is able to account for all of those different needs.
David:
Yeah. I mean, I think that the, just like you said, all of this I think in my mind boils down to coming up with more ways that people can naturally communicate with your systems in ways that your systems will respond back in a more natural manner.
David:
I remember before even Voiceflow 2, when you all released that demo that was showing the new NLP capabilities in the high fidelity smart prototyping, and it gives you the ability to, on the fly, just change, like hi, I’d like to order a pepperoni pizza, and then midway through the conversation, you’re like, actually, you know what? I want that to be a cheese pizza. And you throw in all these different variables, which for us, we do context switching all the time naturally.
David:
It’s just like, there’s so much implied things within conversation that you draw back on, so that you know, naturally when you’re just speaking with someone, you’re like, Oh yeah, they were referring back to, two questions ago when I was asking for the kind of pizza. But that’s what the big thing that’s happening right now, that’s what I keep trying to communicate on this podcast, for the folks that aren’t as immersed in talking inside baseball like you said about voice is that, the state of where we’re at with this technology in my opinion is that we’ve gone from a very brittle sort of framework that breaks all the time, or it’s like, I’m sorry, I didn’t understand that. And then people walk away and they’re like, this thing sucks, and then you have a bad taste in your mouth.
David:
And I think though what’s happening is that all of the underlying stuff is getting significantly better, meaning that it’s translating into just a much better experience, it starts with the accuracy of understanding what you’re saying goes dramatically up, it’s like now it understands more or less what I’m saying.
David:
And then the next big thing is going to be around the actual, natural, conversational AI element of it, where it actually can kind of in the same way that we communicate, it can refer back using all the context switching that we do, and so, again, with this talking about a prototyping tool, I just think this is so cool that what we’re kind of seeing in real time, surfacing is a tool that’s just so much more dynamic, it’s so much more conducive to the way that we all actually communicate, because ultimately that’s what we’re striving for is just so that what you said, when you have a bank, clearly there’s going to be a percentage, a small percentage of people that really need to speak to an operator.
David:
And so it’s a matter of, how do you filter out the 95% of people that we can more efficiently just get them what they need and get them on their way without having to tie up our operators. And a lot of that, that’s I think what IDR was ultimately trying to solve, but it wasn’t the optimal solution. And I think that we’re kind of now in the midst of a lot of these new optimal solutions being built, because what we’re ultimately seeing are computers that are more or less learning how we communicate in our language. We’ve always learned computers languages through all the different, Java and all the coding languages and all that, ways that we have to get computers to talk back to us is always through their language.\
David:
And now what we’re doing I think is, we’re creating a better system that can register the way that we naturally communicate, so that in a few years I think, or I hope, you’ll be able to communicate in any one of these different channels and all these different modalities in a very natural way, and it will just feel like it’s not something that’s… it’s like alien to us, I think it will kind of feel natural.
Anna Rosen:
Yeah. There’s so much that you said that I love. The number one thing being the drawback. I think that’s such an important piece of conversation design and what makes a good experience a good experience, and it’s something that, it’s only a problem that you realize you’re going to have to account for, once you get far enough down the road of conversation design.
Anna Rosen:
But you’re right, I mean it’s the ability to design these natural conversations, human-like, not to the point where you’re tricking people, but just to the point where you’re comfortable enough to speak naturally as opposed to the way that we still kind of speak to our speakers which might be a little bit more, I’m enunciating and I’m speaking slower than I naturally would, that kind of thing. Rob showed us yesterday in V2, you can say, I really, really, really, really want to order a pizza and it’s going to understand you and it will just sort of account for the different ways in which we communicate across languages, across cultures.
Anna Rosen:
And I think what will come from that in the next couple of years will be the persona development. So it’s being able to design specifically for who I’m speaking to as a conversational AI. And if we go back to the banking or even let’s talk about healthcare, those are the kinds of situations where you might want to switch personas and be a little bit more compassionate in the way that you speak to someone when you’re talking about their knee surgery rehabilitation, versus when you’re just telling someone how to reset their password or whatever it is.
David:
Yeah. And I think that that all boils down to context. I mean, when I first started to look at the voice technology space, the thing that really registered with me was this idea of contextually aware assistance. And this was back in like 2017, and so, it’s taking longer than I think I initially thought, but that was just my naivety, I didn’t realize how many challenges there were at the infrastructure, the plumbing level of making the systems work in such a way that you can do that. But I think the big first step is making it so that they are a lot more, and it’s just like you said, it’s not in order to trick people and like, man, we’re going to be living in a world where it’s going to be impossible to discern one from the other, I don’t think that’s really the point, it’s just to make it so that you can more naturally communicate with them that you’re not having to deliberately speak differently in such a way, because you’re operating off of the assumption that if I speak in a natural way, it could break this thing.
David:
So, just to your point as that happens, and I think that’s well underway right now. What we’ll probably start to see is a lot of the contextual understanding, is that the way in which we communicate will be context dependent, it will be just like you said, where this particular conversation that I’m having, I use a different inflection in my voice, I use a different cadence in the way that I speak, because it’s a more sensitive subject or something like that.
David:
Or if I’m like Rob, where it’s like, I want a pizza, and it’s like, just get me a pizza, and I want this, this and this. I just think that it’s all going to be boiled down to what exactly can we infer from the context clues here? And as the systems get intelligent of how they can retrieve all that context, that’s what will enable, like you said, kind of the more persona element to this, which I think is going to be a really fascinating iteration of this too.
Anna Rosen:
Yeah, I think that’s definitely one of the parts I’m most excited about, and it’s just interesting because all of these extra pieces that are going to be added into the experience are part of the ways in which we communicate as humans, that we kind of get for free, just things we’re able to infer. For a lot of different reasons that you’re not consciously thinking about all the time, you adapt your tone and you adapt how you’re speaking to a person depending on who it is. And it’s the AI just kind of catching up to that in order to meet us where we are in terms of our preferred way of communication.
David:
Well said. Well said. Okay, well, as we sort of wrap up here, closing thoughts, I mean, anything else that you want to draw attention to from Voiceflow to, one of my favorite things about you all is that it’s always like at the tagline, at the bottom is like, we can’t wait to see what you all build. And I guess as we close, what are some of the things that you’re expecting? You don’t have to go into super granular detail, but I’m just amazed at the possibilities here.
David:
And again, going back to the beginning of the conversation, you have people that are developing these in a hundred different countries, and it’s just going to expand in time. And I think that we’re kind of seeing something special here, we’re at the early onset of a totally new type of subset of developers and ways that we can design these different experiences, and it just feels kind of like we’re at the beginning of something really cool.
Anna Rosen:
Yeah, I feel the same way, I’m super excited about all of the possibilities and everything that we’ve set people up to be able to build and I know that the team just feels so strongly about our community and supporting them in those things that they’ve been asking for. So we’re excited to just see what folks do with the SDK, what kind of ideas and creativity kind of comes out of that.
Anna Rosen:
For me I think the thing I’m most excited about is seeing some of these in-car assistance that teams have been deciding, of course, will kind of come to life in any kind of experience that you think about it, where audio or voice just offers a superior experience in the way things are happening now, I think those are the ones that fire me up the most, and those are the conversations that I’m most excited about having.
David:
Yeah, I would say that audio and voice are probably the embodiments of the internet right now, it seems like that’s where all the action is. And I think that the two go hand in hand, there’s going to be a lot of really interesting ways that the two work together, because I agree with you, I mean, I think that these two spaces are so deeply intertwined, because again, it’s how you listen and how you speak more or less.
David:
And so, so much of that is how we now listen to computers and how we then also can speak to computers. And I just think that technology really is feeling like it’s kind of moving into this more invisible layer, more or less. And it’s allowing in a weird way a lot more of a sort of, I don’t know, there’s just more ways that we all as each other and as humans, we can all communicate more effectively now.
David:
I mean, I think that so much of this is removing a lot of the task work, a lot of the grunt work of getting from point A to point B, it’s like the jobs to be done thing where so much of that is just, if it’s my phone I got to dig through it to even get to what I’m ultimately trying to get to. And so I think that’s the big pie in the sky value proposition is, well, what if you could just speak it? I mean, what if you could just interface with each other through these different things?
David:
I’ve talked about clubhouse a number of times on this podcast where it’s like, what’s so fascinating about that? To me it’s just like, we can now just quickly hop on and speak with one another one to one, five people, 50 people, a group of 5,000. I mean, it seems like we’re kind of entering into this era where technology is more or less just enabling a level of communication that I don’t think we’ve ever seen before.
David:
So, I think it’s getting interesting and things are definitely starting to take shape in that regard. So as we wrap up here, Anna, share with us where we can connect with you, learn more about Voiceflow, and just continue on with everything Voiceflow related.
Anna Rosen:
Yeah, so you can find me on Twitter, whatsupanna. You can find Voiceflow at http://www.voiceflow.com and you can find our tutorial videos at voiceflow.com/tutorials I encourage should check out all of that stuff and feel free to reach out to me if you have any questions about anything that I’ve talked about today.
David:
Brilliant. Awesome. Well, thanks so much, Anna. Thanks for everybody who tuned in here to the end and we will chat with you next time. Cheers.
