The rise in demand for speed and convenience from consumers has seen more and more businesses employ technology such as Artificial Intelligence (AI) and bots to help refine and take the pain out of time consuming processes.

There seems to be confusion and misconceptions, with some companies labelling what are actually automated systems and processes as AI or bots, but there’s actually a big difference between the two.

Put plainly, automation is software which follows programmed rules, it is their manual configuration which enables automation, such as sending out thousands of personalised emails, repetitive robotic processes (think of production line manufacturing) or data tracking. You have to programme the software to carry out the actions that you want to be made.

AI on the other hand is designed to simulate human thinking, explained by computer scientists John McCarthy and Marvin Minsky as tasks performed by a machine or programme which, if it were carried out by a human, would need intelligence applied in order to be completed.

AI often has the ability to plan, reason, learn, understand and problem solve. It’s intelligent technology that interacts with users through websites and apps, enabled to interact with humans, like a human would. This is much more complex and intelligent than programming software to do certain things at certain times.

AI bots can be used for a whole range of processes, helping to provide customer service support, answer consumer queries, push relevant new content or products and to help teach users new skills. They can be integrated across a number of channels, ensuring that experience is consistently maintained.

Although the employment of bots and AI has risen in recent years, we know that people still prefer the personal touch, which is why it’s important that even if responses and processes aren’t physically carried out by a human, we still perceive it to be an enjoyable and personable exchange.

When we developed the Cosmo bot for TechQuarters, we wanted to ensure that this was the case. We designed the bot to be humorous, creating a persona for it and setting out how it would respond, interact, sound and work.

This bot, like almost all others, didn’t learn on its own, it had to be taught. Just like the four-time winner of the Loebner Prize Turing Test, the world's best conversational chatbot, Mitsuku, it needs to be taught by a human how to answer. The more it’s used, the more unhandled questions it’s asked and the more a human can train it. So, Cosmo, just like Mitsuku, needs to be chatted with for it to learn.

Cosmo uses Luis, Microsoft’s Natural Language Parsing (NLP) platform and within Luis, it will highlight phrases that it hasn’t understood for you, the “trainer” to provide it with the understanding. Once trained, it can be tested, and the new understanding published.

If you have ever used an Amazon Alexa or a Google Home device when they say that they didn’t understand “but they are learning all the time”, they really mean it. There will be humans training them. I for one have noticed a marked improvement and this is shown in how much more they understand.

Clearly, this is very different from automation, but that’s not to say that both technologies don’t have huge enabling potential in their own rights, allowing businesses to make processes more effective and efficient.

Razor works closely with clients in order to see where AI can add value to business processes and human interactions. From the simplest automated query to complex conversations, it’s all about designing efficient systems that have just the right level of intelligence to get the job done.