AiThority Interview with Sofia Altuna, Global Product Partnerships at Google Assistant
by Sudipto Ghosh on March 9, 2021
Hi Sofia, please tell us about your current role and how you arrived in the field of voice AI / VOICE Talks / Google.
I am a member of the Global Product Partnerships team for Google Assistant, leading go-to-market partnerships strategy and also the host of VOICE Talks, a monthly show focused on voice technology. In my Google Assistant role, I’m working on initiatives that make everyday tasks on mobile faster and easier via voice first experiences.
A native Spanish speaker, the power of language and how conversations can be interpreted is fascinating to me. I’ve always been passionate about making technology accessible and easy to adopt for more people, and voice allows that. Speaking comes naturally to humans, so voice makes it simple for people from all parts of the world, from all ages and genders to use this technology. Prior to working on Google Assistant, I was involved with other Conversational AI projects at Google.
Since joining the Google Assistant team three years ago, I’ve worked across multiple product features globally and with many partners. This has given me a broad understanding of the voice tech ecosystem, the possibilities and challenges across the platforms, and the opportunities for brands and users. Being at the intersection of product engineers and partners provides a unique perspective to understand both the technical complexities and our partner brands’ vision, goals, and requirements.
What is the most contemporary definition of a voice bot? How does your Voice solution help in understanding customer journeys?
A voice bot is any voice or chat interface that relies on Natural Language Understanding (NLU) for interacting with users — not on static scripts. It can be synonymous with conversational experience, conversational interface, conversational UX, conversational app (this is not an exhaustive list of synonyms).
Google Assistant brings together all of the technology and smarts we’ve been building for years, from the Knowledge Graph to Natural Language Processing. Users can have a natural conversation with Google to help them in their user journeys. In the last episode of VOICE Talks on Starting a New Decade with Voice & AI we share how for the Google Assistant to have a natural & smooth conversation, it should be able to understand when it’s being spoken to and should be capable of responding to several requests during an interaction.
And beyond understanding just the words and intent, the Assistant needs to deeply understand the social dynamics and context of a conversation — having a back and forth conversation without having to say Hey Google every time, while also making sure the Assistant responds with relevant answers catered to you vs another person in your family (e.g. Voice Match).
You saw us starting to tackle some of these challenges with new technologies like Continued Conversations, which lets you have a natural back-and-forth conversation with the Assistant without repeating “Hey Google” for each follow-up request, Multiple Actions, single queries that include a combination of up to 3 commands, and Duplex (among others), which completes tasks on user’s behalf. We’ll continue to expand the possibilities of what’s possible with Voice, delivering greater conversational intelligence and speech innovation to provide solutions that further help users in their customer journeys.
What are the leading brands that leverage your voice solutions, including Google Assistant?
There are many brands that are leveraging voice solutions in very effective ways, but I’ll name a few below that I’ve particularly been impressed by:
Since we launched Google Assistant, Spotify has been a leading brand on the platform. Playing music has become much simpler and faster through Voice. In fact, smart speakers owners listen to more hours of music a day compared to the general population. However, Spotify is not only a leading brand in Google, Amazon, Apple and Samsung’s virtual assistants, but they’re also exploring Voice in new ways, and are an emerging innovator in the voice industry. For example, they are running voice-activated ads in the UK and will use voice to suggest songs based on emotions through biomarkers in the voice data.
Snap is another great brand innovating in the space. Last year, Snapchat was one of the first brands to use our new App Actions features on mobile, enabling users to use voice commands to take a photo hands-free, create a Snap with a lens, and search for profiles. In their app, Snapchat also added “Voice Snap,” a new voice interface that enables users to find the right lenses and music — all hands-free in partnership with SoundHound’s Houndify voice AI. In 2020, Snapchat acquired Voca.ai, which is an AI-based voice assistant developer, further solidifying their investment and innovation in the voice sector.
Walmart has also been a leading brand in the e-Commerce industry integrating Voice across an array of its brands and services. On Android phones, Walmart has integrated with Assistant to enable users to access their app via voice, check orders, book a pickup time / reserve a time slot, and check in to your Walmart order. They’ve also launched experiences across Google Assistant devices that let you create your shopping cart hands-free, and this year, they’ve added voice search to its subsidiary, Flipkart, in English and Hindi that were built in-house. Its initial release in January is already drawing over 5M voice searches per day.
Lastly, I’m a big fan of the usage of Voice in cars as it provides a much safer way of getting things done while driving, so I’m very excited about the new Google partnership with Ford that will leverage Google Cloud’s AI, ML and data analytics technologies in millions of future Ford and Lincoln vehicles.
VOICE Talks debuted in April 2020 and is now about to celebrate its one-year anniversary. What can we expect from VOICE Talks this season?
You can expect a recurring theme around who and what will power the future of voice. We all know what voice is capable of today in our homes, cars and mobile devices but voice technologies are just getting started, particularly as AI, machine learning and NLU are becoming more sophisticated. We recently did a show on what the next decade of voice will sound like from this technology perspective, bringing in experts from across the industry to make their predictions for what many believe will be the “roaring 20s of voice.” Our March episode will be dedicated to mobile applications, but also celebrating some of the women who are the driving forces in the voice technology industry. Your readers are invited to tune in at 2 pm March 25 at voicetalks.ai.
This year we are excited to partner with more subject matter experts in the ecosystem, such as Brett Kinsella, Dr. Teri Fisher, and Google Developer Expert’s, and to bring more instructional and hands-on content to viewers via hackathons and code camps.
You can expect to hear from experts in the Conversational Design and Android Developer ecosystem on how to build voice experiences that are designed for both personal and shared spaces, conversational and multimodal, frictionless personal and multi-user, and built on top of existing web and Android ecosystems. We also look forward to showcasing brands who are on the cutting-edge of bringing voice to their products and accelerating innovation in the industry.
You can also expect VOICE Talks feature stories in education, gaming, technology, and application development while making the tech space accessible for everyone.
Can you give us an example of some of the more unique case studies or products that you’ve highlighted on the VOICE Talks show?
In the first episode of VOICE Talks I was very impressed by a company we highlighted called Voca.ai, which partnered with Carnegie Mellon University on the “Corona Voice Detect” project to enable fast diagnosis of COVID-19. The project combines AI and voice forensic technologies to find “specific patterns in voice, tone and other sounds that we produce as we speak, that relate to unique illnesses and other human factors.” (source)
One of my favorite episodes last year was episode 4, focused on Diversity, Inclusivity and Accessibility. In the episode we shared a few case studies and products that can help not only the 1 billion people in the world who experience some form of disability, according to the World Report on Disability, but also…well everyone, as we all have situational / temporary impairments (e.g. we shouldn’t be looking at our phone and reading text messages). One of the products that we highlighted was having the Google Assistant read web pages out loud on Android phones when you say “Hey Google, read this page” and you can adjust the language, voice persona, and how fast it reads. While this is great for people with visual impairments, or reading disabilities like dyslexia, everyone else can also benefit from it, allowing you to get an “audio version” of a blog post, news article, etc.
Lastly, throughout many episodes we’ve also highlighted Nike multiple times. Nike is proving to be adept at embracing voice from several angles simultaneously. They integrated voice into their apps allowing you to start tracking your run with Nike Run Club by just saying “Hey Google, start my run with Nike,” they’ve also integrated voice to let you tie / untie Nike Adapt shoes via voice, and they did a promotional campaign during a live NBA broadcast of the L.A. Lakers vs. Boston Celtics game where 2MM+ viewers were prompted to “Ask Nike” via Google Assistant for the a chance at a free pair of the Nike Adapt BB before of its public release.
What were the key lessons you learned during the peak of COVID-19 months? Would you like to share these with our readers?
Although we began our journey towards voice technology long before this current crisis, COVID has provided a new perspective of the importance of this technology. First, as more people were staying at home, voice assistants played a bigger role in work productivity, education, and family activities.
Secondly, people wanted to avoid touching shared devices (or any device), so I think Voice was and is poised to be part of the solution that helps shape our new normal and make our lives easier and safer. This is something that makes me excited about this space, of all the opportunity there is and the impact that we can have.
Some interesting trends we learned:
– Communication: The number of text messages sent with Assistant more than doubled in 2020; Many of us also turned to group video calls through Zoom, Google Meet and Duo to host virtual celebrations, attend workout classes and to simply catch up with loved ones.
– Cooking: Recipes searched with Assistant grew over eight times this past year, and thanks to quarantine baking trends, banana bread remained the most requested recipe on Assistant.
– Family engagement: Last year, the number of stories told by assistants increased over three times.
– Smart home controls: Not surprisingly, the most popular ask was to turn the lights on or off.
– Dictionary and translations: the amount of requests made to Assistant for help with spelling, using the dictionary and translations more than doubled this year. The top translation request this year was for “I love you”—and the top word that people needed help spelling? “Quarantine.”
How has your outreach strategy with the Google Assistant partner ecosystem evolved in the past year?
The pandemic changed the way people interact with Assistant, so as we were starting to see and foresee some of the new trends mentioned above, we evolved our strategy to have a stronger focus on work productivity, education, and family activities. We immediately released Google Meets on smart displays and partnered with companies like Zoom, to facilitate meetings at home, we launched a new Learning Hub to make it easy to discover Educational content, and for users that are looking for inspiration, they can just say “Hey Google, teach me something” and see an array of educational content, and we also launched many new games and tools for families, including Family Bells, to help families stay on schedule in an interactive and fun way.
It was also very important to find creative ways to engage the developer ecosystem to stay ahead of those trends and identify potential opportunities to create impact voice experience with users. With the cancellation of major in-person events like Google I/O and Mobile World Congress, we had to rethink our approach to keeping the ecosystem connected and informed during this time. So, in April 2020 we launched VOICE Talks, a monthly series aimed at the voice developer community. Every month, we bring insider content from the world’s leaders in voice technology – from industry trends to case studies to business tips to product demos and announcements. This also provides brands an opportunity to share their insights with the community: Matchbox.io, Sony, Bamboo Learning, AmEx, Verizon, Headspace, Vizio, iRobot, Nike, Dunkin have all been on the show.
I’m proud to say I’m part of one of the fastest growing internet shows in this space. It premiered for a second season in January 2021 with more than 110,000 subscribers, and 34,378 hours of content consumed.
Why should more consumer brands embrace a voice strategy?
Voice has taken a major leap forward and it has emerged over the last couple of years as a new foundational interaction model in computing. As users start to have access to this technology everywhere, and this behavior becomes more normalized, if brands want to meet the users wherever they are, they’ll have to start incorporating voice technology into their strategy.
Voice technology also allows brands to engage key audience segments in personalized conversations through more natural and seamless interactions, which can ultimately drive retention and business growth.
Brands that are using voice technology as part of their strategy today, are not just creating new experiences for their users, but are beginning to learn & invest in the future of customer interactions (i.e. they are developing the technical know-how to navigate the new computing era — the first-mover advantage).
Any advice for brands thinking about developing for Voice?
The first question brands normally ask is: how should we think about what experience to build? Users are not just looking to access a brand’s website in audio form (at least not now). Voice is a much more “intent” base (i.e. use case base). Brands should spend time thinking about those moments where they can be truly assistive with voice and create re-engagement.
At first, it’s important to think about how to help users in sustained, often daily/weekly/monthly repeatable interactions. For example, it’s become common for food ordering apps to start their voice journey around use cases like “reordering,” as well as for banks to build an experience to quickly check your account balance or bills vs purchasing a new credit card or opening an account.
Tag a person in the industry whose answers you would like to see here
Jessica Dene Earley-Cha and Braden Ream!