Conversations with Hans van Dam
By Hans van Dam
Conversations with Hans van DamNov 23, 2021
Helping brands create their own voice assistants
Join host Hans van Dam for his conversation with Matt Buck, the banjo strummin’, guitar pickin’ CTO of Voxable, a company he founded with his wife, Lauren, in order to seize upon a market opportunity.
Most of the voice solutions Matt and Lauren were seeing were pre-packaged, ready-for-implementation products; the design work was already done and little to no room remained for customization. But how, then, could users ensure their voice assistant reflected their brand? They couldn’t, really, and this is how Voxable came to be, offering a tool that enables users to design their own ideal voice assistant from scratch.
Voxable is the conversation design platform for teams that want to build better voice and chat apps. In December 2020, Voxable pivoted from conversational agency work to focus on a product informed by the Voxable founders' extensive industry experience and user research conducted with top conversation designers.
Matt describes his transition from a developer, to a businessperson and salesperson-cum-engineer as Voxable set about bringing its product to market: the day-to-day of running a company, building relationships, marketing and raising awareness, listening to users, and positioning Voxable for future opportunities.
It’s an interesting conversation about being at the right place at the right time and having the vision and wherewithal to seize upon opportunities, some of which may only be emerging, and some of which may not even exist yet, but will as voice technologies increasingly become part of our daily lives.
Don’t miss this one!
- 05:00 Matt Buck: From engineer to salesperson.
- 11:30 Recognizing the opportunity.
- 16:30 Voxable’s target customers.
- 20:00 Making it user friendly.
Bringing Voice To The Masses
Perhaps you remember that moment in the 1990s when the internet and email transitioned from things only computer science majors talked about, to things everyone knew and used? It just sort of… happened. Maarten Lens-FitzGerald thinks that we're approaching a similar transition with voice assistant technology, and he’s a passionate evangelist for the value of voice.
Sure, there are already smart speakers all around us—from the phone in our pocket, to our automobiles, to the Alexa device on the kitchen counter—but, as Maarten reveals, voice technology is still very much in its nascent stage.
In this episode, Maarten talks with host Hans van Dam about his work helping organizations understand the value of voice, and navigating the challenges of implementation. Challenges like language, age, and standardization. Sure, for English speakers, voice technology may already seem commonplace, but what about multilingual environments like Europe, where not only is there more than one language, there are also regional dialects of those languages?
Maarten also explores the question of older people, who are not as technically savvy or computer literate as your typical early adopter? And, just like things were with the internet and email in the Wild West 1990s, voice technology is currently a mélange of proprietary technologies that will require agreed-upon standards and protocols before it can become truly mainstream.
There’s still a way to go, and still lots of work to be done, but Maarten is confident that voice assistant technology will soon be as commonplace as, you guessed it, the internet and email.
“Alexa: Taco's maken.”
- 02:15 Introduction, voice evangelist.
- 04:39 It’s all about change management.
- 07:10 The point of emergence.
- 09:47 The need for standards.
- 15:30 Project Zilver.
- 19:20 The Dutch Voice Coalition.
- 33:00 Finding compelling use cases.
- 40:00 Mindset, skillset, culture and systems.
AI: Make It Easier But Don’t Dumb It Down
Per Ottosson is a man on a mission. As CEO of Stockholm-based Artificial Solutions, Per wants his clients to be as productive as possible in their AI journeys.
Join host Hans van Dam and Per discuss the challenges organizations frequently face when deciding to incorporate conversational AI.
It starts with leadership that understands the value of AI and is willing to make the required investments. But then what?
As Per reveals, it’s a road of trial and error that’s been well travelled by now, and there is no longer a need for most companies to start from scratch, or having to make do with off-the-shelf solutions that don’t reflect their brand. Thanks to the solutions provided by a company like Artificial Solutions, the journey is easier than many chief executives might think.
By leveraging internal talent in collaboration with advisors like Artificial Solutions, and by applying concepts like localization and hybrid learning, organizations wanting to get into the AI game can do so more easily and economically than ever before.
Per also discusses the challenges of operating in multilingual environments, and provides some case studies that reveal solutions to irritating-yet-all-too-common problems, like AI that clumsily translates language with no nuance or context. (At present, Artificial Solutions works with 86 languages.)
He also discusses how AI technology needs to be flexible to adapt to cultural, linguistic, and even generational differences.
Essentially, it’s about making AI easier to use, while keeping it engaging and relevant: a challenge that Per Ottosson and Artificial Solutions are passionate about, and a discussion you’re sure to enjoy.
- 02:17 Per’s background.
- 03:24 AI disappearing as a concept.
- 06:01 It’s like a bowling alley.
- 07:24 Your AI needs to come from within.
- 09:52 Be prepared to make the investment.
- 11:05 Leadership.
- 22.55 Multilingual environments.
- 29:10 Hybrid learning.
- 43:00 Don’t dumb it down.
AI: Listening And Understanding Are Key
Empathy is a word that gets thrown around a lot, but what do people mean when they say something like “you need to be more empathetic”?
In this episode, host Hans van Dam gets philosophical with author and autonomous systems architect, Phil Harvey.
As is discussed in his new book—Data: A Guide to Humans—Phil talks about how empathy has been a buzzword in recent years, but not a well-understood one.
To understand—and benefit—from empathy (especially as it concerns AI technology), Phil talks about the difference between biological empathy and cognitive empathy. He also explores what we can do to improve our ability to be empathetic, including venturing outside our comfort zones into the world of art and music to develop our ability to listen, and understand and interact with the world around us.
Ultimately, if we can build machines that are better at listening to and understanding the people trying to communicate with them (i.e., being empathetic), the more authentic and productive our relationships with the technology will be.
It’s a deep discussion, ranging from data and systems, to biscuits and cakes, that will help you appreciate empathy as a skill that can—and should be—developed in order for us to succeed in our chosen field, but also in life.
- 01:54 Phil’s background – an Arts degree in AI.
- 03:08 Empathy, a word that gets thrown around a lot.
- 07:16 Biological versus cognitive empathy.
- 10:24 What is data?
- 16:30 Systems optimization and too much data.
- 21:17 Positive interdependence versus transactional isolationism.
- 22:34 Conversational AI.
- 27:22 Potential abuse of empathetic AI.
- 29:39 Bringing it all together.
- 32:22 Using art to improve our cognitive empathy skills.
When It Comes To AI, Personality Is Everything
We’ve all engaged with a virtual assistant or chatbot that isn’t very helpful, or is even downright annoying. So how do we improve this technology? What goes into making AI more intuitive, more natural, more helpful, more… human?
Join host Hans van Dam in a fascinating discussion with Jason Gilbert, the former film maker turned character designer who specializes in creating interactive personas for AI technology. Having worked with the likes of Facebook Messenger, Skype, Amazon Echo and Google Home, there’s a good chance you already “know” Mr. Gilbert’s work.
Jason is currently working with Intuition Robotics on robot assistant technology to help vulnerable senior citizens, work that, as he explains, goes well beyond just the technical challenges. How, for example, do you give a machine an authentic character so it forms a trusting relationship with people.
How do you teach a machine to be empathetic? Where is the line between interacting with something and interacting with someone, and what happens if that line gets blurred? Hans and Jason discuss these questions and more, and their talk is sure to leave you pondering a few of your own.
Jason discusses his current work, his work with AnnA (you may have spoken with AnnA if you have an Amazon Echo), the background of AI character development (including a mention of the Turing test for all you Blade Runner fans), and the technology’s future potential across myriad applications.
As it turns out, trying to make machines more human might just teach us a bit about humanity.
Inbenta Uses AI To Create Meaningful Chatbot Conversations
Online marketing, eCommerce, and call centre managers face some pretty tough challenges these days.
Tasked with transforming emails and calls into web traffic, this task can take a long time. And if a company’s website is lacking, visitors will leave in the middle of a transaction!
This sounds like a no-win situation, but Jordi Torras, CEO and Founder at Inbenta, has a solid solution for senior execs in this predicament: Inbenta.
Jordi and host Hans van Dam explore how Inbenta helps companies automate conversations by chatbots.
Initially, Jordi tackled the big problem of frequently asked questions (FAQs) on websites. Most companies have them, but users aren't fans, mostly because even with a search engine the results are all the answers containing every word in the question, and that can be a pretty long list!
Back in the 2010s, Jordi and his team developed successful technology to match user questions with answers. When the age of conversational AI arose, the FAQ-search engine problem remained the same.
Jordi’s early work placed him and his team in the perfect position to create solutions. Inbenta can match user questions with intent, and there’s no need for any training of company staff. In fact, with customer data Inbenta can go live within 24 hours.
Ibenta understands the meaning of what someone says, and that can be mapped to intent. Inbenta is built on a linguistic model, rather than a statistical one.
Inbenta has the linguistic model that requires no training and it extracts potential intent. It then uses machine learning for disambiguation, based on user behavioral patterns.
Former Google Conversation Designer Explores Persona, Personality in Chatbot Writing
Conversation designer Jasper Klimbie focuses a spotlight on the incredibly wide range of backgrounds writers can bring to conversation design!
Jasper trained as a lawyer, studied journalism, worked as a comedian, and performed as a musician. He studied screenwriting at UCLA Film School, wrote ads, and content copy.
“I was really learning the power of language, and it’s so interesting to see what well-written language can achieve, and that well-written means different things for different applications,” explains Jasper, who is the Chief Procurement Officer at the Conversation Design Institute.
Jasper and host Hans van Dam explore giving conversation design personality, infusing conversations with delight, avoiding repetition in chatbot conversations, and why excellent copywriting is an essential part of conversation design.
Jasper entered conversation design after taking on the role of creating the Dutch personality and content for Google Assistant.
The project presented some unique challenges, in part because in the Netherlands people don’t like hierarchy compared to some other countries, so the Google Assistant could not be perceived as a servant.
In addition, cultural differences became evident, because Google has such a strong American ethos. For example, the Netherlands centres its December festivity on Sinterklaas, but Google found this strange and so Sinterklass had to be explained to Google U.S. Ultimately, Google had to make a leap of faith, allowing local markets to be trusted to make these decisions.
This is where Jasper’s background in law became an asset. Many questions were routed to him so that he could make judgment calls about whether something may offend.
In addition to all the challenges, the Google Assistant required a persona, a personality that makes the content more engaging.
Converting Information Into Conversations Is The Mission At Kare
In the pre- and early Google Docs age, tech companies were focused on building platforms and developing AI. Within organizations, companies were building their own internal systems.
Today’s guest, Tim Porter, looks back to that age and recalls: “There was a huge focus on artificial intelligence and machine learning for the end consumer.
“But what people weren't focused on was really leveraging business knowledge that customers need or business knowledge that their employees need,” Tim tells host Hans van Dam, co-founder of Conversation Design Institute.
Back then (and still today) a lot of time was wasted on going through emails or databases looking for relevant documents to answer customer or employee questions.
Tim has solved this problem. He is the Founder and CEO of Kare Knowledgeware, an UK-based software company that makes information understandable and accessible. Kare helps users turn information into knowledge, which can then be turned into conversations.
Kare does this by converting and dynamically maintaining an organization’s legacy knowledge base instantly. It is a proprietary platform, built after years of research using artificial intelligence, machine learning, and natural language understanding.
Kare provides businesses with a simple way to understand what customers need answering. It can bridge platforms and then almost instantly pull answers and convert the information into conversations. The platform can respond to either voice or typed queries.
Using Kare means that businesses can serve up exactly what customers need to resolve issues. Tim and his team took quite some time to arrive at solutions to make Kare work. They had to rethink the whole approach to intent and search. The ultimate solution was looking at use cases and designing the technology around those.
The magic doesn’t end there! The software is self-improving, and will automatically recommend how the system can be better. All of this happens in a controlled environment, where Kare asks the appointed team member for permission before installing its recommendations. All the solutions are completely editable.
In addition, Kare will identify when questions or inputs are starting to cross-reference or diverge, and will recommend that a user actually split an answer so that the information becomes more specific.
Tim says that Kare can work really well for any kind of system in any industry. The system has a web crawler where you literally just put in the website URL and it’ll crawl thousands of pages if you need to, it doesn't matter what the scale is.
If businesses don’t want to launch like this, a user can put Kare in escalation mode, so that information is improved and updated.
For a small to mid-tier business, it can literally be set up out of the box and be up and running within about 20 minutes. For larger businesses, it takes about two to three hours to set up.
Ultimately, it saves money. A business will spend less time searching for information and developing documents containing information, and more time on the real things that matter—including delivering value to customers.
Looking to the future, Tim says Kare is going to enhance the voice tech components of Kare. The enterprise will also aim to become more proactive, meaning that it’s going to build predictability into its searchers. This is possible by following a customer’s journey on a website, tracking what the issues might be, and making changes to Kare over time.
Translation, Localization, And Conversational Design In a Multilingual World
Many forward-looking companies are embracing chatbots and working with skilled conversational designers to engage new markets.
But global enterprises—even smaller businesses—risk damaging their reputations by not making local translation (called localization) a priority.
Conversations host Hans van Dam explores this fascinating challenge with Thorben Stemann, a Conversational Linguist with Kocarek, in Germany. Kocarek is an experienced translation service provider for small and medium-sized businesses. The focus of its translation activities is on technology, legal, economics, marketing and communication. Many forward-looking companies are embracing chatbots and working with skilled conversational designers to engage new markets.
But global enterprises—even smaller businesses—risk damaging their reputations by not making local translation (called localization) a priority.
Conversations host Hans van Dam explores this fascinating challenge with Thorben Stemann, a Conversational Linguist with Kocarek, in Germany. Kocarek is an experienced translation service provider for small and medium-sized businesses. The focus of its translation activities is on technology, legal, economics, marketing and communication.
Translation isn’t a new issue for global businesses. They have been considering tone of voice, brand presentation, and cultural nuances for decades now.
However, in voice tech Thorben points out that at the core of all issues with language technology is the fact that everything is demoed and prototyped in English.
He explains: “Most of it works great on English. Just adding one more language . . . usually you realize: ‘Oh, that sounds rough’. You always have to add these slashes to add both genders and it doesn't look so great.”
So because different languages have different grammatical structures, designers may need to subtract or add information.
“For example, in the Slavic languages, when you speak in the past tense the verbs all take a gender. That means you need to know the gender of the person talking. And ideally, you would need to know the gender of the person you’re talking to,” says Thorben.
“As well, will the bot be using formal or informal speech, and will it universally use the same forms to address someone? Those are just technically sometimes also challenging, not just from a design standpoint. But how do I actually make sure that's working in the infrastructure of the bot?”
Thorben says the number one mistake businesses are making in voice tech is placing localization at the end of the conversational design step. A translation, he explains, is a rewrite in another language. The whole process is being done again, so ideally, it should be taken into account right from day one.
Inclusivity is another important issue in voice tech. Hans notes that copywriters need to understand nuances within a culture, so that everyone understands what is being said.
Kocarek is working to solve these and many more challenges. Thorben’s role is to be a bridge between designers and linguistics, working with engineers and copywriters to ensure that solutions to linguistic problems are found, and that the finished dialogue sounds natural and makes sense when in use.
“We want the linguists to be able to adapt to the technological possibilities, and maybe refrain from doing things that are just outright impossible in the technological tool that you're using.” At the same time, Thorben works to sensitize engineers to language so that they recognize when something sounds weird or doesn’t evoke emotion.
Find Thorben on LinkedIn.
The hurdles that conversation designers and developers face: client team alignment & how a conversational bot can fulfil their needs
The two main challenges facing conversation designers are client team alignment and ensuring that clients have a good understanding of how a conversational bot will fulfil their needs.
To ensure success, a conversation designer needs to get client team members on the same page, so that there’s a shared understanding of what a bot can achieve. Client teams also need regular progress reports so that they can give frequent feedback. Actually, accomplishing these things is much harder than it sounds.
One way to resolve these challenges is to actually build a bot, but is rarely a viable solution. A client probably won’t want to pay for engineering work if they are not ready to make a decision to proceed with the work
In this episode of the Conversation Design Institute podcast, host Hans van Dam talks to Obaid Ahmed about some of the hurdles conversational designers and developers face, and now Obaid founded Botmock to solve many of these hurdles.
Botmock is a platform that allows for the rapid creation of prototypes and design conversational experiences across multiple channels – both voice and text.
A conversation designer and entrepreneur, Obaid came up with the idea for Botmock while he was working for a design agency and developing Facebook bots for clients. He and his team kept hitting a roadblock - difficulty in giving clients a strong sense of what a good conversation on Facebook Messenger would look and feel like.
“One of our team members thought: ‘Maybe we can handle this better’,” recalls Obaid. “We created a simple script that would take input from Excel and other places and convert it into a dynamic chat.”
Botmock was born.
Obaid outlines the kinds of people involved in a project, such as a writer, developer, and a data scientist, noting that as soon as you have more that one person involved in a conversation design project, you need a system to map and keep everyone on the same page.
Exploring the societal and philosophical implications of AI in the modern world
Sina Kahen is a voice strategist based in London and Co-Founder of TheVoiceCourse.co and Vaice. He has spent the last three years merging psychology and biomimicry with voice technology, presenting to brands such as BBC, O2, and Amazon about how to make voice experiences more emotionally intelligent.
But his career started far away from voice tech. Sina began his career in the medical technology world. After being sponsored for an MBA by his company, he discovered voice technology and the power of voice.
Sina still works full-time in medical technology, but the MBA experience saw him develop a strong interest in voice technology’s place in branding and consumer needs. After being part of a pro-bono agency as part of his MBA course work, Sina launched his own company - Vaice.
In this episode Co-founder and CEO of the Conversation Design Institute Hans Van Dam talks to Sina about voice technology, its future, and the role AI is playing in technology interfaces. In this far-reaching episode, they explore the societal and philosophical implications of AI in the modern world.
Sonic branding for voice assistants, with Diederik van Middelkoop
Diederik van Middelkoop is Executive Creative Director at Amp.Amsterdam (Amsterdam, New York and Mumbai), a music agency specializing in sonic branding, music production and music supervision with a great love for the innovative.
Diederik has been a welcome speaker at many festivals, including Cannes Lions, Spikes Asia, Promax, LIA, and ADFEST, to name but a few.
His clients have included The Olympic Committee, Google, Samsung, Nike, Adidas, Coca Cola, Heineken, Jet Airways, Qatar Airways, Audi, Mercedes, BMW, Axe, Johnnie Walker, Zespri and many more. In this conversation, we focus on some key concepts behind sonic branding and what they mean for designing AI Assistants. The key question in the age of voice, is how do your customers know it's you they are talking to.
The conversation explores how brands can be recognizable and trustworthy as they seek to leverage the opportunities of voice interfaces.