Q&A: Dr Wallace on AI & having conversations with robots
Published on: | by Robin Block
Dr Catriona Wallace is an experienced entrepreneur and the Founder of Fifth Quadrant, ACA Research and The Ventura. As CEO of ASX-listed Cr8tek (ASX:CR8), her latest project is Artificial Intelligence FinTech, Flamingo.
“My vision is to create an ‘out-of-the-box’ virtual assistant. We see the big tech-giants democratising machine learning for developers — we want to do that for businesses. Our product will be usable and configurable by marketing people, sales teams and customer service representatives.
“We want to allow anyone to configure their own journey, train the machine and use the platform themselves. We are live with clients now and the results are hugely encouraging. So our next goal is to make the platform and product accessible more as a self-service product to businesses globally by the end of next year.”
Flamingo has built two AI programs, ROSIE and MAGGIE — Cognitive Virtual Assistants that “solve the problem of low online sales conversion rates and drive efficiencies for financial service firms.” The machine learning algorithms mimic human conversation and guide customers through previously opaque journeys.
It should be noted, however, that this is an early stage FinTech company and success is no guarantee. Investors should seek professional financial advice before making an investment.
MitchelLake’s Robin Block recently spoke with Catriona to understand the ingenuity of their product line and gain perspective on how AI is set to change business.
Robin: Can you describe what your products do and why they are important?
Catriona: If you look at where business is done best, it is when humans have conversations. The challenge for enterprise-level businesses in the digital age is to replicate that level of service at scale. My idea was to take that conversational element and automate it through machine learning.
Essentially, we have fused web-chat, dynamic web forms and machine learning to create a virtual assistant capable of guiding a customer through an online journey as effectively as a human in an office.
ROSIE does sales conversation and customer service and MAGGIE is focused on lead conversion. Try and imagine ROSIE as a 300-dimensional graph. She is a neural network program that is pre-trained to understand human language, natural language processing and language comprehension. She is capable of ‘learning’, integrating new concepts into her existing framework, in close to real time — as opposed to other types of AI that need to be batched and run separately.
AI as a concept can involve ‘rules-based decision trees’, ‘if and then’ problems or ‘machine learning’ — but does not necessarily involve all three. Our company uses machine learning — many of the companies operating in the artificial intelligence space do not.
Robin: What are the largest challenges you have faced in building Flamingo — what have you learned?
Catriona: Building a global company in a short time at the scale we have done has required a high degree of ingenuity — we have needed people who can think outside the box. To do that we have developed a company culture that is safe where people are able to make mistakes and learn quickly. Our largest challenge, however, is not selling the product or attracting talent — it is dealing with the procurement processes of our clients.
Large organisations still have difficulty moving fast when looking to engage with new high-tech companies. That being said, they are getting faster because they realise they will be left behind if they do not.
I have come to understand that with an early stage company, you have to expect everything to be on fire all the time. It is a completely different way of thinking than how I ran my other companies — but it has allowed us to grow faster, be more ambitious and accept the turmoil that brings.
One thing I have really learned through this journey is that because we are building never-before invented technology, it is all about learning. We don’t know where AI is going — no one does — but Flamingo has the opportunity to lead, which we are doing.
Robin: What is the vision for the company and, despite the uncertainty, where do you see the future of AI?
Catriona: Beyond expanding our product range and ramping up its scale, my interest is in machine-to-machine communications. We want to provide end-customers with similar AI tools to those we are selling to businesses so that the machines can transact with machines.
Currently we see the greatest tech minds globally equally split on the future and the dangers of AI, with Zuckerberg (positive) and Musk (negative) on either side of the chasm. I sit 50-50 — it could go very well or very badly. We already know that when machines talk to each other in conversational UI, they quickly create their own language — there are all sorts of things that could go wrong. This is why two vital points in Flamingo’s culture are ethics and security.
The regulations and guidelines around the general use of AI and machine learning are behind the technology. However, I think the right people are starting to think about these things. It is important for the public to know that those of us working in AI take these questions very seriously and are looking at ways to build robots safely.
The truth is that when we talk about the coming of AI, it is already here. The same level of impact that occurred when electricity was introduced is happening with the commodotisation of AI. AI will transform the way we work and how we live. People have to recognise that the transition is happening now. And it is hugely exciting.