The Drum Awards Festival - Official Deadline

-d -h -min -sec


Amazon’s ‘Polly Brand Voice’ heralds a new era of accessibility, but there's still a way to go

By John Campbell, Managing Director

TIPi Group


The Drum Network article

This content is produced by The Drum Network, a paid-for membership club for CEOs and their agencies who want to share their expertise and grow their business.

Find out more

February 25, 2020 | 6 min read

Amazon Polly is a text-to-speech (TTS) service which allows you to create speech from text to be used in your Alexa Skill, website or app.

Alberto Bigoni

Rabbit & Pork look into Polly Brand Voices and consider the likelihood that brands will adopt the feature.

As an addition to this existing service, Amazon recently announced Polly Brand Voices, a new functionality which gives brands the ability to create a custom voice specific to their brand rather than using one of the standard voices from Amazon.

Alongside the launch announcement, Amazon included two examples of brands that have already created Polly Brand Voices: KFC Canada and National Australia bank.

Brands using the standard Polly voice have a range of voices to pick from, for example in British English there is Brian, Amy and Emma. It is easy to switch between these voices to give some customisation to your voice experience. However, the amount of personalisation is limited, which is why Polly Brand Voices is so interesting.

KFC Canada

The KFC Canada use case is a perfect example of the power of having a custom text-to-speech voice. They recreated the voice of their founder Colonel Sanders to be used in a new Alexa Skill which allows you to reorder food.

Colonel Sanders has been the face and voice of the brand for years so it makes sense to bring that familiarity to a voice experience where there tends to be no visual connection.

KFC has used various actors to portray Colonel Sanders who passed away in 1980, being able to “digitise” and create a replica voice will now enable KFC Canada to produce dynamic spoken content across multiple platforms.

Consistency is key to drive recognition

As with traditional audio branding such as sonic logos, consistency will be key to getting the best out of a custom Polly voice. This can be achieved in two ways, firstly by using the Polly voice in all applicable locations, from voice app, to phone systems, to point of sale, to audio ads.

The second is by ensuring the same tone of voice is applied consistently. In the same way that any use of visual assets has to be signed off by the brand team, the use of the Polly voice should also be subject to the same checks. For example it might be the case with KFC that the new Polly Voice should always say “Kentucky Fried Chicken” rather than KFC. Getting this consistency ensures that it not only sounds consistent but it also “sounds like something the Colonel would say”.

Potential use cases for brand voices include

  • Alexa Skills - Allowing a custom voice to be used rather than a standard voice or having to pre-record all speech
  • Audio content on website (great for accessibility) - text can be sent to the AWS service and then read out to the user
  • Dynamic advertising creation (chaning price points, hyper locationsation of audio adverts) - Audio ads can be created in bulk with all variations or dynamic pricing. Much quicker than having to record in a studio with a voice over artist
  • Phone systems - automated customer service or ordering services - Simple functions can be completed with voice assistants deployed over phone lines, changing of address, checking balances etc.
  • POS - your order is read out to you - Checkout tills, vending machines could all use the custom created voice

What's the impact for voiceover artists?

Many brands currently use voiceover artists in their Alexa skill or other audio content. Brands may use a range of voice over artists or pick one particular artist to become the voice of their brand.

We’ve heard in the press about how automation and AI will take over people's jobs. This is an exact example of how someone's job could be taken by a voiceover artist.

For those brands who have an existing celebrity/voiceover artist they will have paid for that person's time to record audio and most importantly, for the brand to be associated with that person.

So what could happen in the future if the brand decides to stop paying said celebrity for their voiceover on adverts/skills and create a digital Polly version? Does the actor continue to be paid?

The film industry has already had questions posed with the digital recreation of actors, the most high profile case being Carrie Fisher in Star Wars.

How do Polly voices compare with voiceover artists?

Anyone with a sharp eye will have noticed that in Star Wars the digital Carrie Fisher wasn’t 100% the same as real Carrie Fisher despite all the advancements with modern animation technology.

The same can be said for Polly. It's good, but there is still something that's missing which means you can tell that it's not a real human.

In animation, they often say you can tell in two areas; the eyes and the mouth. With Polly, there are certain words which don't quite land which gives the illusion away.

In conclusion, this was a fairly unsurprising move by Amazon. It will be interesting to see if Google follows suit and which brands decide to invest in a custom Polly Brand Voice.

John Campbell, founder at Rabbit & Pork.


Content by The Drum Network member:

TIPi Group

TIPi Group is an award winning network of specialist digital agencies. Our agencies are built on performance and profitability and share an ambition to use the power...

Find out more

More from Technology

View all


Industry insights

View all
Add your own content +