AIBusiness recently spoke to Robert Dale, Arria’s Chief Technology Officer, about the potential he sees in applying natural language generation, and how he sees the benefits of AI in his industry. Arria is the leader in real-time data storytelling. Our core product is known as the Arria NLG Platform, a form of artificial intelligence software that specialises in extracting information from complex data sources and communicating that information in natural language. We configure the Platform for a wide range of client needs; and we also offer its technology as pre-packaged SaaS Products and as a Software Development Kits (SDK) via the NLG Cloud which has APIs that allow developers to add NLG functionality to their own applications.
Dr. Dale is Arria’s Chief Technology Officer, responsible for the development of software products in relation to customer demand and developments in NLG. After 17 years at Macquarie University, New South Wales, where he was Professor in the Department of Computing, Dr Dale joined Arria in 2012. He maintains an adjunct position at Macquarie University. He received his PhD from the University of Edinburgh in 1989, is author or editor of seven books and 160 papers on computational linguistics and is recognised as one of the world’s foremost experts in the NLG field. His books include Building Natural LanguageGeneration Systems, which he co-authored with Ehud Reiter, Arria’s Chief Scientist. Robert is based in Sydney, Australia.
What would you say is the key proposition of your company in enabling a business powered by artificial intelligence?
“Put simply, Arria lets your data tell its story. We constantly hear about how we are drowning in data – people were already saying that before Big Data became a thing, so now it’s only worse”, Robert Dale explains. “However, the problem isn’t the data itself – the data has tremendous value. The problem is making sense of the data, so that it leads to actionable insights”.
Dale explains how he has learnt that applying natural language is often the best way to communicate to someone what their data actually means, and what they should do about it. He says that despite graphs, charts and visualisations being great ways of delivering data-information, as far as they go they do all too often evoke a reaction along the lines of: “yes, but tell me what the graph means”.
“Arria’s Natural Language Generation technology is all about bridging what we call the last mile in information delivery: automatically turning the data and what it means into natural language reports, recommendations and calls to action at a scale and speed that you simply couldn’t contemplate using human experts and authors” Robert Dale.
Looking broadly at your products and solutions, could you outline a couple of their key enterprise use cases?
“We’re seeing a great demand for NLG in financial reporting tasks right now”, Dale says. He uses the example of how a natural language application can take a company’s sales data and produce, in natural language, a detailed revenue report providing all sorts of cross-factor analyses and explanations of variance that Dale considers much more meaningful than the graphs and charts alone.
“Reports like these read just as if they had been written by an expert accountant who is talented at explaining things in terms that make sense to mere mortals – except they are produced by the machine in seconds, instead of the hours it might take the accountant”.
“That democratises the access to information in a way that wasn’t possible before; now even small companies can have access to the kinds of reporting expertise that were previously only available to large companies” – Robert Dale.
What Dale considers most amazing about this technology is its appliccability in every enterprise. “Anywhere you have data that someone needs to have explained to them, you have a potential use case for Natural Language Generation” he says.
He outlines the different sectors where Arria have built their applicaitons, anywhere from financial reporting applications to applications used in the oil and gas sector that produces detailed engineering reports that analyses vast quantities of sensor data and talk about potential problems to relieve the engineer of having to spend hours compiling analyses of the data.
“We’ve built applications in the utilities sector, using grid availability data to provide detailed reports on performances and outages, with different reports tailored to the needs of different audiences”.
Elaborating on the wide range of applicability of NLG, Dale says: “That’s a benefit of Natural Language Generation that is perhaps less obvious: as well as tremendous scale and speed, the technology also offers mass personalisation, since it can be used to tell the same story based on the same data but presented in different ways for different needs or interests”.
AIBusiness were interested to hear what the enterprise strategy of Arria, both the nearest future, but also looking years ahead. Are there any particular industry verticals you are focusing on for your solutions going forward, or any particular applications on the horizon?
As the demand for the same use cases seems to appear frequently repeatedly in the finance area, Dale explains that ARRIA’s main focus at the moment is the finance industry. “To meet that demand, we’ve started developing a partner program, where we train other organisations on how to use our tools so that they can develop these bespoke applications”, he says. “At the same time, we see increasing
“At the same time, we see increasing scope for productising the kinds of applications we’ve already built for enterprise clients, so you can expect to see more from us in that direction”, Dale elaborates.
In terms of implementing your products and solutions in the enterprise, have you come across any challenges or difficulties on your journey?
“One of the biggest challenges can be getting a customer to understand the real potential of a true Natural Language Generation solution”, Dale begins. “Often they have just seen simpler template-based systems, and so you sometimes find a view that there isn’t much sophistication or complexity there”.
However, Robert has experienced that when showing people what can be achieved when pulling together diverse data sources and applying language as the common denominator for explaining that data, you often get the ‘aha!’-moment, and you have a convert, he says.
“Not surprisingly, people think generating language is something that ought to be easy, simply because we as human beings find it so easy to do; part of what we need to do is educate around what’s easy and what’s hard”.
Given that you have seen how applicable NLG is and how it really can transform how an enterprise interprets their data, what new products can we expect to see from Arria in the nearest future?
Dale reveals some exciting news that Arria has just released in beta a SaaS product called Recount, which works as an add-on to existing cloud-based financial accounting packages. “That came about because we realised small and medium businesses would love to have the kind of financial reporting we had been building in bespoke enterprise applications; so we transferred a lot of what we had learned from building those applications into a product that is available for a small monthly subscription, putting NLG in reach of a massive audience”, he explains.
In addition to this they are also well on their way to releasing their NLG-tools for third-party use, Dale tells us, saying: “Our Articulator Lite is a cloud-based authoring tool for building simple NLG systems; that’s in private beta right now, and interested parties can sign up to use it via our website. Then, in 2017, we’re releasing our much more sophisticated Articulator Pro toolset; these are the NLG tools we use ourselves to build enterprise applications”.