WHAT WE DO | AI and Machine Learning
We listen to the needs of your business so that you can implement AI in ways that will benefit your efficiency or customer service.
What could you do with AI?
- Provide intelligent Chatbots
- Find trends in customer feedback
- Convert voice feedback (e.g. WhatsApp voice notes) into text for processing
- Flag and identify concerning data
What's all this about ChatGPT?
Even if you know little or nothing about AI, the odds are that you’ve heard about ChatGPT. This is only part of what people mean by AI, it’s just a specific area called a Large Language Model. Simplistically, an LLM is clever enough to take any input about anything and give some sort of sensible answer or output, even in the form of videos in more recent tools like DALL–E. In old-school computer science things are “deterministic”, meaning that a program will follow a path to get to an answer like 1+1=2 or how the bold button makes the text bold. For AI, or Machine Learning, outcomes are non-deterministic. This means that the path is based on all sorts of paths made between the information it has and the question you asked, which change all the time based on other questions and answers. This is similar to how the neural pathways in your brain work.
LLM technology like ChatGPT is incredibly powerful and can save huge amounts of time for many tasks, but it needs to be used in the right way. For example, a car dealership in the USA recently had someone convince its chatbot – powered by an LLM – to say that a car costs $1. Having expert guidance on the implementation of an LLM is vital, especially if you want to use it to automate features for your customers.
Sentiment Analysis
Humans are great at reading the subtext in written text or spoken words. For example, if I said “just exhausting” after telling someone about a really luxurious spa experience they would know I was being sarcastic and making the point that it was so amazing. It’s not enough to look for some keywords, you need to understand the sentiment (meaning) behind. That’s overwhelming for a business to do when there’s huge numbers of comments to read. It’s also hard to spot trends, how can you quantify and report on loads of unstructured text? We can assist you in solving these sorts of problems, whether it’s to process lots of text to find concerning sentiment or whether it’s because you want to see the number of people who talked about staff interaction vs. the number that spoke about the quality of your product.
Voice to Text and Text to Speech
Whether you have WhatsApp voice-notes you’d like to transcribe into text for easy processing, or lots of text you’d like to we can help. AI tools are great at solving this sort of problem, with OpenAI and all the cloud providers like AWS, Google, IBM and Microsoft Azure having great competing offerings. However, the problem is that these tools don’t just work magicially, you have to plug them into something and make them work. This is why you need development help and expertise to guide you and build the necessary integration. That being said, there’s likely simple solutions to your business problem that don’t have to be expensive.