Case Study - Jolly Roger Telephone

The Telemarketing Robot Vigilante: Jolly Roger Telephone

Developing the relationship between Jolly Roger Telephone and voice recognition.

The Telemarketing Robot Vigilante: Jolly Roger Telephone

The Jolly Roger Telephone Company, a valued Speechmatics customer, provides a unique insight into the many capabilities of voice recognition. You may think it’s mostly used for auto-captioning the news or in your home-controlled Alexa device, but as CEO and co-founder Roger Anderson explained to us, it’s uses go far beyond that.

We’ve dubbed Jolly Roger Telephone the ‘Telemarketing Robot Vigilante’ because that is essentially the AI’s purpose. Jolly Roger Telephone is, in Roger’s words, a robot whose job is to keep telemarketers and telephone scammers on the phone for as long as possible. Due to the small connection rate, telemarketing machines will call hundreds of channels at once. These machines are loaded with thousands of numbers, calling channels to feed a single telemarketer. Jolly Roger Telephone keeps them on the phone, by using phrases like “What’s your name and how are you today?” and “What’s on your screen right now?” This prevents telemarketers from potentially calling and scamming thousands of innocent people.

Roger’s story started with his love for telephone lines in the Caribbean. He said: “A long time ago I thought, how cool would it be to just be a telephone guy who was retired and floated around the Caribbean in his old boat?” His initial aim was to hook up old, unused telegraph wires to connect the islands together with a pirate telephone service that would bypass the local laws and regulations. It snowballed into the service he provides today.

A simple idea is often the most effective. Roger refers to his AI as artificial stupidity – the smarter the Jolly Roger Telephone robot gets, the more stupid it can be. The telemarketing vigilante's power grows, and so does Speechmatics’ importance.

Our Autonomous Speech Recognition (ASR) engine is what Alfred is to Batman. He doesn’t come out of the cave, but Batman wouldn’t be as efficient without him. Roger will use Speechmatics’ speech-to-text to strategically transcribe various parts of a telemarketing call. Our transcript enables the Jolly Roger Telephone robot to match up some of the keywords and determine what kind of scam is taking place. So, if the robot has recordings that are appropriate for that scam, it can be “intelligently stupid,” according to Roger.

How Jolly Roger Telephone Uses Speechmatics’ API

To use our application programming interface, Roger employs a combination of Python, PHP, and Perl. In his interview with us, he said he’s able to get it up and running with ease which, given the multitude of scams and telemarketing schemes popping up every day, is a vital component of Jolly Roger Telephone's effectiveness.

Our relationship is quite a simple one. Roger uses call cutbacks – he sends the recording his AI robot has gathered to Speechmatics, we automatically transcribe it and send it to Roger for analysis. From here, his AI picks out keywords and phrases to ensure the AI is constantly adapting to the variety of telemarketing schemes in place. As such, telemarketing vigilantism can prosper.

Developing the Relationship Between Jolly Roger Telephone and Voice Recognition

As they have been doing, the Jolly Roger Telephone robots will continue to get smarter which means they’ll sound dumber – the longer the call, the more gullible the robot sounds. Technically speaking, Roger hopes to take the telemarketing vigilante global by incorporating other languages into his AI. Luckily for him, Speechmatics’ ASR boasts a 33-strong repertoire of languages.

Furthermore, he aims to detect a potential scammer’s emotions, so the AI can adapt to anger, shouting, or whatever technique used. We recently tested our ASR’s accuracy levels against the iconic 90s sitcom Friends, where we found fear is the most challenging emotion for voice recognition to decipher. Roger hopes that as the technology grows, the combination of his AI and our ASR can detect a caller’s mood and adapt accordingly, potentially saving many people a lot of grief.

It’s cases like this that highlight just how vital our research can be and showcases the versatility of voice recognition technology.

Jolly Roger Telephone: 2022 and Beyond

Just as we’re constantly looking to evolve our ASR engine to become a more inclusive and powerful machine, Jolly Roger Telephone is looking to the future for improvements too. The company has two apps. ‘Captain’s Log’ lets you check your messages and manage your account, and ‘Pirate Maker’ enables you to make your own Jolly Roger Telephone robot.

For example, if you get a call and don’t know who it is, you can let it go to Jolly Roger Telephone. You can then pick up the phone and listen as they’re leaving a message. If you want to pick up the phone you can. If not, let the robot do the talking. It’s rather like a modern version of the classic answering machines of the 80s.

Ai-Media

How Ai-Media consistently deliver accurate video captioning in a range of languages.

3Play Media

How 3Play Media have improved its captioning service for accessible videos.

Liquid Voice

How Liquid Voice launched a ground-breaking solution, integrating ASR technology from Speechmatics.

Unlock the value of speech

Everything you need to deliver incredible voice-powered products and features, globally.