Will the ease of voice control come at the cost of relinquishing real control over the products and services we use?

With 2017 already shaping up as the year of voice, it’s safe to assume innovations in this space will flow thick and fast. Amazon’s Alexa dominated this year’s CES, while most 2017 predictions have stated that voice — supported by AI-based services — will be a dominant influence in the tech sector this year.

Whether you back an Apple, Amazon, Google or Microsoft horse, it’s certainly looking like these services will infiltrate our homes and — eventually — our workplaces.

What does control really mean?

Voice control is like the proverbial iceberg: the simplicity of communicating with a device or service, using one’s voice, belies the complexity of the systems, algorithms and general smarts beneath the surface.

Such systems are clever and are relentlessly improving, however the net effect is they’re disintermediating us from the logic and decisions that are being made on our behalf.

By giving us the ability to control tasks using our voices, are we not, in fact, limiting the control we have over these tasks and how they’re performed?

Let me explain. Until comparatively recently, we’ve always had control over the technology we used. We click or tap and the device does what we want to do, driven by our actions and the choices we make.

Ten or twenty years ago, we may have even known how our computers even worked: what hardware and software they comprised of, often down to the component level. Even as computing as reached the mass market, most of us didn’t want — or need — to know how these things worked, but we had choices over the sites and services we used.

As the user interface increasingly becomes invisible and interactions are performed via voice — and, eventually, without any interaction at all — the consequence is that we’re further disintermediating ourselves from the decision-making process: our instructions go to servers in the cloud, and algorithm-based engines automatically determine the next best action or instruction.

By removing the complexity of interaction, we may be relinquishing the power of control.

Does control really matter?

Right now, there’s a novelty factor so this may not matter. And the current common use cases for using voice control are reasonably discrete and without significant impact: changing a music track, calling an Uber or requesting today’s weather forecast are all singular and specific tasks that are ideal for voice control. We don’t necessarily need to understand where the information we’re receiving is coming from, or how behind-the-scenes decisions are being made.

At the same time, voice control isn’t suitable for all tasks: Researching an article, looking for a news story, or selecting a new bank account require us — for the moment at least — to access and distil multiple pieces of information, tasks that are more efficiently performed using a screen than a speaker.

As voice services get embedded in more devices, as more third party services jump on the voice bandwagon and slipstream the hype, and as the services themselves become smarter and more widespread, we need to look at what control we want or need to have, and which interaction method gives us that control. This will be particularly the case in the enterprise, where a greater understanding of the methods and ethics of decision-making and option selection will need to be more transparent.

The arrival, en masse, of commoditized AI services will create new opportunities for new services that will attempt to make an increasing number of decisions on our behalf. This is undeniably exciting and the proverbial Holy Grail for many of us who have used technology for a long time.

As voice control becomes an increasingly important part of how we work and play, we need ways to understand how these decisions are being made on our behalf. If we don’t, then the empowerment we get from using our voices will be countered by the lack of control we have as soon as our words leave out lips.

There’s a new way to think about simplifying the complex and identifying opportunities for new products and services.

Millions of words have been written about how to identify market opportunities, test new ideas and launch new offerings.

But our world is increasingly complex. Knowing where to start, and understanding how things work, is becoming increasingly difficult. Just as we would break any problem down into its most basic components if we wanted to solve it — the same principle applies to identifying new technology opportunities: “atomization.”

Here’s how it works.

Just as you can start with a single Lego block and add more blocks to create something new, you can also disassemble an existing Lego model and see how the blocks worked together. The Lego block becomes the atom — the fundamental, most basic unit available — using which new things can be created.

Applying the Lego analogy, if we break down complex systems into their basic parts — just as a kid would do with a Lego building, or a mechanic with an engine — we can not only understand how something works, but we can reassemble those parts into something new, using these parts as proverbial building blocks.

How does atomization apply to technology?

We can apply “atomization” to the systems and services we use today to identify new opportunities. By breaking something complex and looking at its most basic form, we can derive “building blocks” that can be used to create new products and services of value.

Whether we’re emailing, tweeting or sending an invoice, each of these data forms — or atoms — contains valuable data and information. Here’s how:

• A tweet — We all think of a tweet as comprising 140 characters, but there’s a lot of additional metadata — comprising more than 30 additional elements — that are part of a single tweet’s payload. In addition to the basic message that’s sent, the data includes the identity of its creator, the location from which it originated (if location services have been available), the date and time it went out, and many other data points.
  By understanding the data that’s in a tweet, what happens when you group related tweets together based on any of those data points? What insights could be gleaned? This is why the Twitter fire hose — and Twitter — is so valuable: with more than 6,000 tweets tweeted every second, this corresponds to over 350,000 tweets sent per minute, 500 million tweets per day and around 200 billion tweets per year. This is also why there are myriad companies who look to mine Twitter’s data to identify patterns and perform analysis on this data, whether for marketing, sales or intelligence purposes.
• An email — Similar to a tweet, the format of the basic email message is also standardized. We’re all familiar with email and how it contains the details of the sender, recipient(s) and is timestamped, in addition to what the content in the email contains. Similar to a tweet, when taken in aggregate email can provide insights into the nature of a relationship a person has with another person, whether they have communicated with a person — or not — in addition to analyzing the content of a message. Again, by looking at email in an atomized form, new ideas on how to use this standardized, ubiquitous form of data can easily be generated. And despite rumors of email’s demise, it’s not going anywhere: Radicati Group estimates that by the end of 2019 the number of worldwide email users will increase to over 2.9 billion. That means that over one-third of the worldwide population will be using email by year-end 2019.
• Financial transactions. The same principles can be applied to the basic transactions in an accounting system, such as Xero. Take an invoice: the typical invoice contains a rich amount of information: what was purchased, when this happened and the value of a transaction. For any single organization, it can report on this information to help it understand its cashflow, or what it needs to buy or sell. When taken in aggregate — across multiple organizations — one can identify purchasing trends across whole industries or regions.

This is the principle of atomization at work: When something is broken down into its smallest parts new opportunities can emerge to take those parts and combine them in new ways to create new products and services.

Using Application Programming Interfaces (APIs), services such as data analytics and machine learning — which are becoming increasingly commoditized — now only are new services easier than ever to identify, they’re also easier than ever to deliver. This openness and use of standardization also allows combinations of data to be considered — what could be gleaned by understanding the flows of invoices sent by email, for example?

Atomization is not a new concept, but rather a new way of thinking about what’s happening and how one can identify new opportunities. Any such solution is reliant on data being structured, but this too is becoming less of an impediment: there are many initiatives out there focused on putting structure around unstructured data (such as text mining and analytics, or image and facial recognition). The upshot: the principles of atomization will continue to apply.

While the ultimate power may be held by the organization that holds the data — whether that is Twitter, Google, or Xero — the platforms that have been built deliver foundations upon which new products and services can be built. Atomization is one approach to looking at complex systems, breaking them down to their most fundamental atomic units, and then looking at how these units can be used — on their own or in combination with other atoms — to create new products and services.

Banner image adapted from one designed by Rocketpixel — Freepik.com.