Hey, Siri, what’s the future of enterprise voice services?

The future of enterprise voice services involves much more than asking Apple Watch to turn off your lights.

Apple, Siri, enterprise, voice, services
Gremlin / Getty Images

Apple’s focus on privacy is incredibly important when it comes to the development of Siri and voice services, as the potential goes way beyond asking your Apple Watch to turn off your lights.

Silent whisper

If you use HomePod, you already know that Siri will respond to commands only slightly louder than a whisper.

You can use it to carry out tasks, control home automation equipment, media playback controls, answer questions, and more. Apple’s AI teams are constantly developing ways to get Siri to do more.

Even then, it’s the thin end of the wedge.

I’ve explained before that Apple, Google, and Amazon all keep recordings of your voice interactions with their smart speakers.

I’ve also explained the fundamental differences in their approaches that make Siri the most secure voice assistant to use:

  1. Siri’s recordings cannot be traced back to you, as Apple does not link them to your account. Both Google and Amazon keep those recordings associated with your account.
  2. Apple deletes those recordings after a set period of time, while Google and Amazon do not (that’s my understanding).

A recent Washington Post story (summarized here) explained the dangers of these warrantless recordings. It’s why I’d never recommend transacting a business deal in the same room as any non-Apple voice device – even though around 50 million people probably do. What you say, the conversations you have, have value.

Unlocking the value of voice

Voice assistant services aren’t just about what happens at home.

Enterprises know the value of the voice – and they don’t even need to analyze what you are discussing in order to unlock that value.

If you’ve ever spoken with an automated customer support chatbot, you understand how voice provides value, in this case in terms of the provision of personalized services. Apple’s Business Chat service is its solution to support these machines.

It’s also important to consider how the nature of voice itself is changing now that conversation is digitized. This voice data can be far more easily analyzed. Data analytics of voice communications is real, and it is happening.

In a sense, it already takes place each time you speak with Siri, which (while mostly reliant on pattern matching) also boasts an increasingly accurate AI capable of figuring out contextual information to boost recognition of what is said.

So, now you have voice assistants that are capable of understanding what is said, figuring out the context around which an utterance is made, and matching patterns and analyzing the content of millions of conversations concurrently in real time.

What could the enterprise value of these systems be?

Hey, Sir, what’s RPA?”

There’s a way to go before you see voice widely used as a component in systems for business or workflow analytics or robotic process management – but there are already signs of the incoming change.

Think about voice as biometric ID.

Banks already use voice recognition biometrics as part of their fraud control systems — HSBC’s VoiceID has prevented $400 million value of fraudulent transactions.

Think also about how voice assistant technologies can act as some form of digital twin for real human chat.

Perhaps you are involved in a group discussion on some matter. Ordinarily, people take notes during the meeting. These days we see document sharing systems that work around enterprise collaboration suites, some even provide on-the-spot automated transcription. Voice assistants add automatic detection of key dates, future meetings, targets, and project outcomes – similar to the way iOS may spot a meeting date and time in an email. 

Similar technologies applied against customer service calls may identify similar trends:

  • Do particular product or service fault call requests tend to take place at particular times of day?
  • Automated systems may pick up signs (such as an unusually high number of customer support requests) that warn of a product or batch flaw. 

Mining the data

Back in the call center, data analytics of voice communications should yield useful insights around performance (was the resolution satisfactory?) to follow up to enabling and empowering support for personalized services: If a customer rings twice, can such systems flag what their previous request was and provide call center operatives with the current resolution status of the matter?

Properly mined, voice data analysis may also yield early insights around service faults, and at least one big telecommunications company plans to mine voice data in order to provide business analytics solutions based on voice for enterprise customers.

Underpinning all these transactions sits a need for privacy.

Indeed, Apple CEO Tim Cook’s demand for a bill of digital rights in which customers can control and manage all the different slices of their digital data held by third-party entities seems an appropriate response as enterprises large and small inevitably explore how to use this data for their own ends.

How many times a day does your smart lighting system send usage information back to its manufacturer? And what do they do with this data once they own it?

Learn more about voice services in the enterprise and share your thoughts: Join the @IDGTECHtalk Twitter chat on Thursday, May 9, using the hashtag #IDGTECHtalk

Please follow me on Twitter, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.

Copyright © 2019 IDG Communications, Inc.

It’s time to break the ChatGPT habit
Shop Tech Products at Amazon