Monday, January 18, 2016

Interview with Ben Taylor, CEO/Founder of Rainbird

Who are you?

Rainbird was co-founded by James Duez, a serial tech entrepreneur turned Angel Investor, and Ben Taylor, an ex-Adobe Computer Scientist.  James and Ben met in James’ last start-up which developed an AI platform which saved British Insurers £10s of millions each year by spotting manipulation, fraud and collusive behaviours.

What does Rainbird do?

Rainbird is an AI platform which enables you to build knowledge maps of human expertise that learn by being used. It has a visual interface that is accessible to business people, not just software developers.   

What caused you to start Rainbird?

We both had experiences building expert systems and other AI technology - but it’s always been very hard work. We recognised that AI was inaccessible to most businesses because meaningful systems typically require Knowledge Engineers, Data Scientists and Domain Experts to collaborate. Many AI projects fail because these groups do not easily work together well, especially when domain experts have to teach knowledge engineers about their domain. Rainbird is also probabilistic and can handle uncertainty. Unlike other systems, it is excellent at driving assertions in the absence of data.

What makes Rainbird an AI startup?
Tools built using the Rainbird platform are able to perform decision making tasks that would previously have been only possible by consulting a human expert. 

Why is there so much excitement about AI startups?

There are a good number of genuinely interesting new AI technologies out there although many commentators who confuse the different camps, and it is generally accepted to be one of the most disruptive technologies to impact the economy over the next 10 years (McKinsey Report on Disruptive Technologies). There is a lot of hype around AGI and ANI which captures people’s imaginations - despite the median expert’s view that these will take 40 and 60 years respectively to come to fruition. The real economic impact will come from platforms that can make ANI solutions today, which we think is a more useful. Of course there are a lot of AI consultancies and brokers jumping into the space with a good deal of tech companies who are actually recycling largely commoditised technology.  

Describe Rainbird's AI.

Rainbird starts as a semantic modelling tool - allowing domain experts to express what they know visually.  Rainbird’s model is based around capturing uncertain expert knowledge that would not necessarily be evident by analysing large amounts of data.  Once a model has been built others can enter a consultation with it to find answers to questions.  The model adapts and learns from these consultations, taking what it has experienced and discovered to strengthen future decision making.

How do you measure and communicate the quality of your AI?

AI is about improving efficiency, augmenting human knowledge workers and innovating new types of solution that would not be possible with the technologies that mainstream computing provides. We have seen the work of 1000 people being done by 50, and to a higher standard. The bottom line benefits of AI are demonstrated by knowledge workers having a reduced time to competence, delivering quicker and more effective customer resolutions, first time, reduction in risks, better business predictions and product recommendations. The ultimate measure is an improved bottom line for businesses, and improved customer loyalty. Of course there are some projects which aimed to provide an improved societal impact.

Can you share something awesome that your AI has done or that you have been able to do with your AI that would surprise most people?

Rainbird is working in the payments to reduce frictional cost surrounding disputes. We are modelling an holistic view of the customer to have the way for more useful financial service products. We are working to transform some of the busiest contact centres in the world to reduce costs and improve service, and modelling an expert resource on Ebola which could be accessible to millions of medial experts in the event of a pandemic.   

When you talk to investors, do they care about your AI or do they care that you are solving a business problem? How about customers? How about the media?

Investors tend to look at the underlying AI only in the context of differentiators and IP protection. For the most part they see it as a black box that can solve a genuine market need - best demonstrated by paying clients. The media are equally keen to hear about outputs, but are more fascinated by the underlying science,and how they might present it in a way that others can understand.

Where do you go to get information about AI startups?

O’Reilly have recently published a useful map of the Machine Learning space (which features Rainbird), and there are plenty of journalists compiling lists (see this one this week

Make a prediction about the future of AI startups.

We predict quite a lot more proliferation (particularly consultancies) followed by significant consolidation.  The tech unicorns will continue to bulk up by acquiring the most interesting technologies and the BPOs will acquire the most successful technology-agnostic AI consultancies. Many of the AI technologies will not live up to to the hype (quite a few are already falling short). I so think the ANI train is now unstoppable and will quickly become mainstream computer science. The AI Effect will still apply - As soon as AI successfully solves a problem, the problem is no longer a part of AI. 

No comments:

Post a Comment