This month’s R U OK Day may be occasion for a check-in call from friends and family, but it won’t be long before that contact comes from artificial intelligence (AI) systems that monitor your everyday activities and raise a flag if something looks off – like if you’re about to quit your job.

Emotionally-aware systems are one of many emerging possibilities as software providers integrate AI capabilities into back-end business systems – putting sophisticated AI capabilities within reach of small businesses that could never hope to build them alone.

The promise of AI-enabling small businesses drove Elmo Software, which offers a cloud-based human resources (HR) and payroll system that has more than 1300 customers across Australia and New Zealand, to embark on a 12-month partnership with the University of Technology Sydney (UTS).

Working together, the two organisations will produce a predictive analytics tool that “will enable our system to use deep pools of aggregated, anonymised data across the whole spectrum of processes in the employee lifecycle,” Chief Commercial Officer Darryl Garber told Information Age.

“These data are quite unique to our offering because we focus on the ANZ region.”

Ongoing analysis of employee payroll data, for example, might identify employees that had become disengaged from the company and were working fewer shifts, or had rapidly taken large amounts of outstanding leave in preparation to leave the company.

It could also help small businesses recruit and choose staff by streamlining the vetting of staff and reviewing development programs over time.

“We think this will be a great opportunity to assist our clients by leveraging off their existing data pools,” Garber said.

AI everywhere

Building AI capabilities hasn’t traditionally been for the faint-hearted, with large enterprises the first adopters because of their deeper pockets and bigger potential returns.

The market for AI software platforms grew by 26.6 per cent last year to be worth $US2.6 billion ($A3.8b) and exploding demand is driving developers to seek competitive advantages by overhauling existing business applications with AI.

Smaller application developers are finding other ways to join the AI revolution.

Some are acquiring their AI expertise, as insurance consultancy Finity recently did, while Elmo was attracted to UTS’ AI-trained academics and facilities like its UTS Data Arena visualisation system.

Other software vendors are leveraging cloud-based AI offerings from IBM, SAS, Amazon Web Services, Microsoft Azure, Palantir, Google Cloud and myriad others.

For developers of applications that perform a specific business function or service a particular market, AI is a way to light a fire under industries where existing systems are often just digital versions of well-worn business processes.

Class-action powerhouse” law firm Shine Lawyers, for one, recently spun off an AI tool called Claimify that gathers details of a road accident and compares it against historical data to judge whether an insurance claim is likely to succeed.

Its ability to shorten the claims cycle makes it a boon for lawyer and insurance underwriter efficiency – “a key disruption in the legal services market,” as Shine puts it.

Local-government developer Civica, for its part, recently drew on Microsoft Azure Cognitive Services to bring AI capabilities to its Spydus Library Management System.

“Libraries are busy places that don’t have the time or staff resources to manually catalogue, index and manage all of these data resources,” Civica ANZ managing director for libraries and education solutions Simon Jones explained.

By integrating AI into Spydus, he said, “we have been able to automate the task – and enhance the quality of meta information making these data collections more accessible, and hence more valuable.”

Spydus is used in over 2,000 public libraries in Australia and abroad, and a pilot program with Melbourne’s City of Stonnington saw massive improvements in archiving and sorting important content.

The system was fed large volumes of digitised images – such as archival photos, maps, scans, and sketches – as well as long-form text documents such as meeting transcripts, flyers, newspapers, articles and community group newsletters.

“You cannot underestimate the value of local community information,” said Stonnington Library senior team leader for systems and resources, library and information services Natasha Tsui-Po, “but creating access to it has always presented challenges as each item has to be manually indexed, described and categorised.”

But can there be too much AI?

AI’s integration into progressively more niche-focused applications will see learning technologies working in the background to monitor and predict our likely behaviour across every aspect of our lives.

AI can be meaningfully applied to any data set with consistent, well-understood parameters, Gartner vice president and analyst Tracy Tsai told Information Age.

“If it’s repeated and regular, it’s likely that AI can do it,” she said, noting that AI toolkits were driving the “democratisation of AI” as developers rush to embrace AI-driven capabilities as “something new that they can do to enhance their applications.”

“As long as the content has a regular base and real basis, and you know the outcome you are looking for, you can train a machine to read the content and improve the quality and efficiency of the job.”

Effective AI implementations will learn better in environments where they have large volumes of data to be trained on – something that Elmo Software is counting on.

“For us, it’s about having a very relevant data pool where we can make meaningful quantitative assessments for our customers,” Garber said.

Yet even as current AI applications focus on known data sets to complement existing business activities, some developers are pushing AI into distinctly human realms.

‘Emotion AI’ was among 21 emerging technologies that Gartner recently added to its hype cycle for emerging technologies – promising to analyse a range of personal and behavioural traits to decide whether you are happy, sad, angry, or any other emotion.

This technology – tweaked, for example, to detect stress in a customer’s voice – is already being used to help call-centre operators detect an impending blow-up and change their support approach accordingly.

Over time, Gartner believes emotion AI will also provide benefits in areas such as detecting insurance fraud, improving dementia diagnosis, analysing the emotional state of drivers, and changing the presentation of educational materials to suit students’ emotional state.

Yet with a host of AI-powered systems allowing employers and service providers to maintain a constant vigil over our emotional state, the obligation to use AI ethically becomes even more pronounced.

A recent Australian Council of Learned Academies report weighed in on the inevitability of AI ubiquity – a positive force, working group co-chair Professor Toby Walsh said, “provided we ensure that the use of the technology does not compromise our human values.”