PentaTech

Natural conversations with Alexa Multi-value slots

Marisa Espinosa — Thu, 18 Feb 2021 11:27:57 +0000

One thing that makes conversations more natural is the ability to speak out multiple values of the same type in a single sentence, similar to what we do in our daily conversations. For example, “I would like to order a pizza with tomatoes, mozzarella, and bacon”

Let’s compare 2 examples, the first one without multi-value slots and the second one with them.

After this simple example you can see the power of simplification that this new technology feature can bring to your solution. Removing friction from interactions is key to improve customer experience.

Multiple simultaneous voice agents means many brands in one device

Marisa Espinosa — Thu, 01 Oct 2020 13:38:03 +0000

By the way, what is an agent?

In plain English, when you use a wake word such as “Alexa”, you are invoking an agent. If you enabled a second agent, you would have another wake word to activate it. Imagine the many opportunities available to have several wake words / agents running at the same time, each one specialised in a particular brand or domain.

The Voice Interoperability Initiative (VII)

Early in 2020, Amazon and more than 30 leading technology companies announced the Voice Interoperability Initiative (VII), a new program with the mission to ensure voice-enabled products provide customers with choice and flexibility by supporting multiple voice agents simultaneously on a single device. In September 2020, Dolby, Facebook, Garmin, and Xiaomi have joined the initiative.

Multi-Agent Design Principles

There are 4 driving priorities:
1) Building voice-enabled devices that promote choice and flexibility through multiple, simultaneous wake words
2) Developing voice agents that can work seamlessly with others, while protecting the privacy and security of customers
3) Releasing technologies and solutions that make it easier to integrate multiple voice agents on a single product
4) Accelerating machine learning and conversational AI research to improve the breadth, quality and interoperability of voice services

Customer choice is a bedrock principle

Customers can choose from available voice agents by enabling the use of multiple, simultaneous wake words when more than one agent is registered on a device. Multi-agent products help customers find the agents that are available and explore their capabilities.

So, what does it mean for your customers and your business?

This technology looks fantastic. So, what? Well, imagine your brand having your own wake-up word. You don’t need to call Alexa, so then you talk to your brand. You remove friction and have a personalised engagement interaction.

References

[1] Voice Interoperability Initiative Welcomes Facebook, Garmin and Xiaomi and Releases First Multi-Agent Design Guide

How are your customers using Voice Assistants

Marisa Espinosa — Mon, 24 Aug 2020 03:14:38 +0000

Customers are looking for new convenient and entertaining ways to spend their time. Voice assistants are becoming ubiquitous and trending quickly; these devices range from smart displays, smart speakers, mobile phones, tablets to cars, wearables and even microwave ovens. This article presents the most common ways your customers are using these devices and how your business can leverage this opportunity.

Smart home or conversational experience?

First, let’s separate the 2 common usages. Voice activated applications can be grouped into 2 major segments:
(1) Smart home, home-office or office activities
Manage IoT devices such as smart light bulbs, air conditioning, garage doors, coffee machines, etc. from your voice-activated device. For instance, Hey Google, turn the lights off.
(2) Conversational experiences
Provide voice applications based on interactive content, video, games, facts, etc. These type of applications rely on other services usually API based, databases for rich content persistence, images if using a smart display, pre-recorded audios and a wide range of techniques to deliver a high quality experience.

Your customers are accessing these devices for different reasons. Some of them are:

Video conferencing

Some of these devices come with built-in video conferencing capabilities, and it can be combined in a routine to dim the lights or make the call. With a visual display, it can be integrated to display your documents or PowerPoint presentations.
One example is Alexa for business which simplifies meeting room experiences for your employees. You can control conferencing systems, check room availability, and book rooms with just your voice. [1]

Play music or tune in the radio

Voice-activated assistants act as a personal DJ connecting to music services such as Spotify, SoundCloud, Tidal, TuneIn among others [2]. Once integrated, users can ask for playlists or a particular song. Majority of these services work on a freemium basis; providing some basic functionality for free and premium available.

Play news and podcasts

Playing the news and podcasts is popular in these devices. They can be combined with routines so if you say “good morning”, then it can inform the weather and combine it with your morning news. [3]

Play videos

Depending on the device, it can also play videos coming from Netflix, YouTube or a particular skill. Alternatively, it can instruct to play it on your TV if integrated.

Drive hands-free

Smartphone distractions while driving contribute nearly a quarter of crashes. In these cases, the driver was using a phone within a minute before the crash occurred, and perhaps even at the moment of the crash. [4]
With your phone sitting on the seat or centre console, say, “Hey Siri, where’s the nearest gas station?” or, “Hey Google, give me directions to Antigua.”

Read out loud

In many cases, the content your customers want to access is actually written in a website; such as blog, recipes, wikipedia, etc. Having your website ready to be readable can give your business an edge on the voice-activated race.
This is a demanding feature for websites, that in some cases require some modifications to the content in order to be found and consumed easily. However, websites don’t need to do anything special to enable this functionality. You simply ask “Hey Google, read this page”. [5]

Answer all sort of questions and queries

Voice-activated devices can answer all sort of questions and queries in a reliable way. Frequently Asked Questions (FAQs) can be presented in a conversational interface and offer better results for your customers. And then provide a way to relate to your business.

Calendar integration

Let’s assume you are a dentist and you want to have an automated way to pick up appointments automatically 24/7. Voice applications can integrate to calendars, existing appointment systems, and set reminders. [6]

Phone calls

These devices allow you to make phone calls; so your business voice application can provide a direct call option, removing friction. [7]

How can your business leverage these opportunities?

There are several ways that your business and your brands can leverage these devices and their capabilities to join existing customer journeys or to develop new ones.
Each of the different touchpoints contribute to your customers experience and these skills should be part of a whole strategy and fit accordingly.
If you care about getting new customers or retaining existing ones, voice applications can give a competitive advantage to your business.

Contact us

If you want to understand more about the described opportunities, please contact us.

References

[1] Amazon Alexa for business. Meeting room experience
[2] Google Assistant help. Listen to music.
[3] Amazon Alexa. Create Alexa Routines for Smart Devices.
[4] The Washington Post. More evidence that smartphones and driving don’t mix.
[5] Google Assistant help. Easier access to web pages: Ask Google Assistant to read it aloud.
[6] Google Assistant help. Create calendar events & ask about your schedule.
[7] Amazon Alexa. Make Alexa Calls with Your Voice.

Boost your voice applications beyond voice – into multimodal solutions

Marisa Espinosa — Fri, 08 May 2020 22:01:11 +0000

Voice solutions are not just about voice. This article compares applications based on voice first, only voice and voice added.

Single mode

A solution with a unique mode of operation or interfacing only offers one way of interaction. For instance, a Google Home or an Alexa Echo Dot only allows voice as input and output. You could argue that these devices also present some visual cues to highlight when it is listening, muted or processing a request converting them into multimodal. Let’s now dive into multiple modes of operation or interfacing.

Figure 1. An Amazon Echo Dot is an example of single mode (voice) with visual cues

Introducing multimodal

The different ways to input and output are called modes. Devices have voice, display, buttons and visual cues to interact with the user. When you combine all these different modes into a single application, we call it a multimodal solution.
Each mode can serve a different purpose. For instance, voice is great for short back and forth pieces of information like “Is it going to rain today? There is a chance of few showers.”; and for streaming music, news, and so on. A visual display is great for collating large amount of data into a single display. For instance, in the Alexa voice app called 4mates for a device with screen, it shows the different types of empanadas available in a scrollable screen. This would be difficult to achieve using only voice.
Solutions need to be ready to work on one or more modes because Google Assistant and Alexa can run on different devices that support only voice, voice first and voice added modes. I coined this “Responsive Voice Design”.

Figure 2. The Google Nest Hub Max provides a multimodal experience combining voice-first, touchscreen, visual cues and buttons

Voice only devices

These devices such as the Google Home and Amazon Echo Dot only accept voice input and output. However, the user experience is enriched with visual cues such as listening and processing spinning light, and mute button and respective visual light.

Voice first devices

Google Home Hub or Alexa Echo Show are examples of devices that are activated via voice and enhance the user experience with screens, lights, buttons, etc. User selections can be made using touch screens or voice activation, creating endless opportunities for engaging solutions.

Voice added

Mobile phones and laptops are good examples of voice added devices. In this case, the main way of interfacing is either keyboard or touch screen; however voice is added to enhance the experience and provide accessibility.

Figure 3. Mobile phones are a good example of voice added functionalities

Responsive Voice Design

In the same way that Responsive Web Design is the approach in which the design and development responds to the user’s behaviour and environment based on screen size, platform and orientation.
A Responsive Voice Design is the approach in which the design and development responds to the user’s behaviour and device based on the mode, voice first, only voice and voice added.

Voice and ambient computing smart homes

Marisa Espinosa — Thu, 07 May 2020 08:14:11 +0000

Computing you can’t see or realise they are doing it.

What is ambient computing?

The key difference between “normal” and ambient computing is that you do not need to log in and execute a direct command in a specific device to get something done. Ambient refers to “in your environment”, and these devices integrate so well into your surroundings that you end up not noticing them.

If you compare ambient computing devices with mobile phones, laptops or desktops; you will notice the latter depend on active human input either by typing or touching. The goal of ambient computing is to eliminate friction between you and the device achieving it. For instance, with an ambient smart thermostat, the device judges the room and your interactions with it to adjust the temperature as needed.

Ambient computing uses a variety of technologies, including motion tracking, speech recognition, gestures, wearables, and artificial intelligence to achieve this goal.

Smart speakers and voice assistants

As of 2020, more than a third of US population has a smart speaker / voice assistant. This is the most ubiquitous form of ambient computing devices. Alexa and Google Assistant are not actively activated in the same as the other non-ambient computing devices.

Let’s say you want to eat empanadas, you would directly ask “Hey Google, send me 4mates empanadas menu”; or if you want to turn on the lights, you would ask to do so. What’s more, usually is part of a routine like good morning. Or turning them off included in good night.

There is no future that doesn’t have ambient computing or voice activation.
Mark Cuban, Entrepreneur

Voice and Ambient Computing are the Future of convenince

I see ambient computing getting more and more ubiquitous because people love convenience. For instance, each day when I wake up, I tell my assistant “Good morning”, and runs a routine including time, weather forecast, news. In the future, a room-sensor could detect your presence, first time in the morning and before 7am, and your speakers might kick-off the morning routine.

Other manufacturers are also integrating ambient computing within their product design. For example, Samsung calls it Smart Home and Philips Smart home automation. Google has also been keen to enter the ambient computing space. The company is likely to expand its hardware line to include more devices that further complement its current array of integrated services.

Is Alexa and Google Assistant always listening?

Marisa Espinosa — Thu, 30 Apr 2020 06:44:39 +0000

This article explains when smart assistants become active, when they start listening, what is recorded and how to manage it.

In the past, there have been some concerns that Alexa may have been listening to confidential conversations due to some voice applications waking up unintended. Another concern was raised about faulty activations making unintentional purchases. Privacy concerns are always valid and they have been addressed. Companies such as Amazon and Google are aware of it and working to enforce high quality on their products.

When do smart/voice assistants become active?

Smart/voice assistants are designed to detect only your chosen wake word (Alexa, Amazon, Hey Google or anything else you configure). The device detects the wake word by identifying acoustic patterns that match the wake word. No audio is stored or sent to the cloud unless the device detects the wake word (or the device is activated by pressing a button). Furthermore, anytime a smart assistant detects the wake word, a visual or audible indicator will signal it.

Let’s have a look to some examples of a device not listening and listening:

Illustration 1: This illustration shows examples of devices not listening

Illustration 2: This illustration shows 2 examples of voice assistants becoming active after a wake word is mentioned

Is the smart assistant recording my conversation?

By default, it is not recording anything. Once the wake word is detected then audio is sent to the cloud, and you can review and delete the voice recordings associated with your account (including any audio resulting from a false wake) in your Voice History available. No audio is stored or sent to the cloud unless the device detects the wake word (or Alexa is activated by pressing a button).

Audio storage

Amazon and Google say their devices are designed to record and store audio only after they detect a word to wake them up. For instance, Amazon Alexa also performs a “cloud verification” of the wake word using the more powerful processing capabilities of the cloud to double-check the audio to confirm detection of the wake word. If the cloud verification does not also detect the wake word, Alexa stops processing the audio and ends the audio stream to the cloud. If Alexa confirms that the wake word was spoken, Alexa will continually attempt to determine when your request has ended and then immediately end the audio stream.

Reviewing and deleting voice recordings

You can review voice recordings associated with your account and delete the voice recordings one by one, by date range, by Alexa-enabled device, or all at once by visiting Settings > Alexa Privacy in the Alexa app or https://www.amazon.com/alexaprivacysettings.

You can also enable the ability to delete your recordings by voice. Once enabled, you can delete the voice recording of your last request by saying “Alexa, delete what I just said” or you can delete all the voice recordings from your account for the day by saying “Alexa, delete everything I said today.” To enable deletion by voice go to Settings > Alexa Privacy > Review Voice History in the Alexa app or https://www.amazon.com/alexaprivacysettings. When enabled, anyone with access to your Alexa-enabled devices can ask Alexa to delete voice recordings from your account. For more details visit the following link.

You can achieve the same thing in Google Assistant by visiting your device settings application (Google > Manage your Google Account). At the top, tap Data & personalization. Under “Activity controls,” tap Web & App Activity Manage Activity. For more information visit the following link.

How relevant is an audio logo for your brand?

Marisa Espinosa — Mon, 27 Apr 2020 13:14:39 +0000

This article analyses the importance of building an audio logo to complete your brand strategy. Now that voice applications are increasing in popularity it is a good time to invest in an audio logo.

What is a brand?

A brand is the image and personality of a product or service that a business provides. A product’s features, such as logos or slogans, make it unique and different. Every interaction with a customer counts towards building a distinctive brand.

What is a sound branding?

Every interaction includes more than just what the customer can see. Valuable brands also include sounds. There is a powerful connection between sounds and your memory. I can still remember some of my childhood jingles. Jingles are a form of sound branding.

What is an audio logo?

If we pay close attention, audio logos are ubiquitous. Let’s listen to some examples that you may recognise.

What makes an audio logo great?

The best audio logos I can remember have something in common, they are short. Three to six seconds is fantastic to capture people’s attention.
Uniqueness will definitely make a difference ensuring it is impactful.

How relevant is an audio logo for your brand?

If we are trying to create a memorable experience, not just to remember your business; but to create deep connections with your brand, then an audio logo will make a difference. Audio branding gives an advantage when it comes to customer loyalty.

How does voice searches work on Alexa and Google Assistant? And how is your business going to be found?

Marisa Espinosa — Thu, 23 Apr 2020 01:22:28 +0000

In classic text searches we provide a query and it returns a list of potential matches ordered by relevance. Visually you are able to select the one you want. Voice search technologies work differently.

How does voice searches work?

When users engage with a voice application from a voice-first platform such as Google Assistant or Amazon Alexa, they state what they need and the platform figures out how to accomplish it.
It is not possible to pick the answer you are looking for from a list of results. The voice assistant will select the best answer for you.
This imposes some new challenges to businesses that want to stay relevant in these new platforms.

Different type of queries or invocations

A query can also be called invocation. At the end of the day, you are asking for a voice application that is going to resolve your need. There are 2 types of invocations:

Direct query or invocation
Indirect query or invocation

Direct invocations

In some cases the request is direct or explicit, this means that the user specifically calls out the name of the voice application that they want to engage with. In this instance, the platform will simply activate the voice application. Google Assistant calls it explicit invocations meanwhile Amazon Alexa refers to this as named invocations.
Let’s have a look to some examples:

Indirect invocations

The other scenario is when the user asks for something they need however, they do not specify which application will resolve their need. The voice-activated device may recommend to users what is the best available way to respond to their query. Google Assistant calls it implicit invocations meanwhile Amazon Alexa refers to this as name-free interactions.
Let’s have a look to some examples:

Why is this relevant to your business?

Your customers may not know your business name or how to get it directly. Similar to searching for a term on Google search. On voice, you want to be the first one so the indirect or implicit invocation favours your business. This opens up a brand-new era of search optimisation on voice-activated devices.

Driving traffic to your voice application

Firstly, you need to have clarity around your voice strategy and how you intend to drive your customer attention to your voice app. Secondly, finding the gaps or unattended terms that are still available to be leveraged. Generating engaging and meaningful content will make a difference.

Final remarks

The following table summarizes the different options when it comes to direct and indirect interactions with voice-activated platforms:

	Direct	Indirect
Google Assistant	Explicit invocation with or without specific request Example: “Hey Google, talk to Aussie Animals”	Implicit invocation Example: “Hey Google, tell me something scary about crocodiles”
Amazon Alexa	Named invocation with or without specific request Example: “Alexa, open Aussie Animals”	Name-free interactions Example: “Alexa, tell me something scary about crocodiles”

Consolidation of customer touchpoints

Marisa Espinosa — Tue, 17 Mar 2020 03:27:33 +0000

Having multiple isolated data sources for customer touchpoints (mobile app, website, voice app, etc.) creates a difficult situation for your organization. In order to extract insights from your data; it is required to aggregate, model and present it appealingly.
Furthermore, when data is stored in different databases, it is challenging consolidating different sources into one because:

Aggregating different technology databases is time-consuming and expensive
Unifying assorted and unrelated models is a challenging endeavor

The following diagram represents the different stages of the customer lifecycle and the databases that get updated along the way:

Consolidation of customer interactions at the customer touchpoint works as follows:

Each time an interaction or event is created, then data is centralized in one system that will take care of aggregate data, model and present it in a way that can be understood and analyzed.
This unified data model across departments allows generating meaningful promotions, campaigns and cross selling among many others as well as to follow the consumer journey and better understand their needs and what is their relationship with our brand.

Conclusion

Unifying channels across the value chain into a single customer oriented database has enormous benefits in the way your brand interacts with your customers. Loyalty programs such as Instant Wins, Event Invitations, Contests, Gamification, Surveys, among others are possible without having to go into complex exercises.

What is a Voice application?

Marisa Espinosa — Fri, 14 Feb 2020 11:35:18 +0000

Voice activated devices are becoming dominant in the market and usage trend is growing quickly; these devices range from smart displays, smart speakers, mobile phones, tablets to cars, wearables and even microwaves. This article is a good introduction to the topic covering the basics of voice applications.

Voice applications

Voice applications are software components that makes spoken human interaction with computers possible, using speech recognition to understand spoken commands and answer questions. A voice activated device is a device controlled with a voice user interface.
Usually, these applications can be static or dynamic depending on the level of complexity that is addressing. Static voice applications will respond the same every time is asked a given question or trigger; meanwhile a dynamic voice application will change according to some logic, for instance first time using the application respond “Welcome to Aussie Animals”, and the second time and subsequent conversations respond “Welcome back to Aussie Animals”.

Smart home or conversational experience?

Voice activated applications can be grouped into 2 major segments:
(1) Smart home activities
To manage IoT devices such as smart light bulbs, air conditioning, garage doors, coffee machines, etc. from your voice-activated device. For instance, Hey Google, turn the lights off.
(2) Conversational experiences
Provide voice applications based on interactive content, video, games, facts, etc. These type of applications rely on other services usually API based, databases for rich content persistence, images if using a smart display, pre-recorded audios and a wide range of techniques to deliver a high quality experience.

Name requirements

Each of these voice applications, or short “voice apps”, have a name, meaning a verbal domain name (e.g. Aussie Animals); the same way that websites have a domain name (e.g. www.pentatech.com.au). This name is used as invocation name, for instance, “Alexa open Aussie Animals”.
Relevant to notice from the legal angle, that invocation names must not infringe upon the intellectual property rights of an entity or person meaning the name is unique to your brand/intellectual property with proof of ownership established through legitimate documentation.

Availability of the voice application
Voice applications can be published public or privately. Public availability means that any user with a voice activated device can access the voice application and private means restricted to a business organization or range of users.

Purpose

Engaging voice apps often fall into these general categories:
(1) Things people can easily answer. Tasks that can be accomplished with familiar input like times or dates, like booking a flight.
(2) Quick, but compellingly useful activities. These usually give users immediate benefit for very little time spent, like finding out when their favorite sports team plays next.
(3) Tasks that are inherently better suited for voice. These are typically things you want to do hands-free, like receiving coaching during yoga or light exercise.

Conclusion

Voice applications are rapidly evolving so stay tuned for the next articles that will cover more interesting topics around voice.