Democratizing AI for rational decision making as-a-service

Artificial Intelligence (AI) or its more palatable versions such as Machine Learning (ML) with its associated areas such as process/work flow automation is expected to be the silver bullet of tomorrow – both for information technology providers and business houses, however mature or immature they are in their digital journey. The increasing awareness and rapid rate of adoption of AI will make it all pervasive.

This will lead to democratizing AI for the benefit of everyone in an organization and not just a minority of leaders, domain experts or data scientists, without too much changes to existing systems and processes. The term ‘democratizing’ has been used often of late in the media. One of the dictionary definitions of the word is to “make something accessible to everyone”.

According to research firm Gartner, AI technologies will be embedded in almost every software product by 2020. The following are the 3 success factors, according to Gartner – i) differentiation among equivalents avoiding what it calls “AI washing”, ii) usage of proven, less complex methodologies, iii) addressing the need for skills, the biggest challenge organizations are supposedly facing for AI adoption.

"As AI accelerates up the Hype Cycle, many software providers are looking to stake their claim in the biggest gold rush in recent years," Jim Hare, Gartner’s Research Vice President

Business Process Management software vendor Pegasystems recently launched Intelligent Customer Decision making as a service, in collaboration with Accenture, for the telecom industry. AI-backed decision making tools will be used for effective marketing campaigns. Accenture will be the partner to execute the technology and achieve performance goals for targeted marketing, measured using metrics like customer churn and new customers attracted.

Software company Infor for example, known for its ERP and associated applications on cloud, recently announced Coleman, its enterprise grade, industry-specific AI platform for its suite of applications. Coleman is supposed to use Machine Learning in the background for improving conventional processes and helping faster, rational human decision making in areas such as inventory management and predictive maintenance, with automation where possible. Coleman uses natural language processing and image recognition for user interaction. To quote Infor’s CEO,

“We are now leveraging machine learning and our access to large amounts of data to assist users with less structured processes such as complex decisions, conversations, and predictions”, Charles Phillips

Other software applications must be having similar intentions to embed AI. Leonardo is SAP’s Machine Learning foundation. What will be important is not the technology or the algorithm itself but how well the usecases are defined reflecting reality and how well they fit into existing business processes for supporting fact-based decision making. AI is not limited to software applications. It is getting ubiquitous across the enterprise and beyond – from automobiles to smart phones and other consumer devices.

For consumers of technology, embedded AI can be a welcome development instead of understanding these complex algorithms, mapping them to their business and implementing them. It may be easier and faster to adopt and democratize if regular IT applications come packaged with Machine Learning. But then it is a black box. The tool or any human will not be able to explain the logic behind a decision it has suggested – the “why” part. “Explainable AI” is an upcoming area to translate the output of these algorithms to a human understandable form.

For technology service providers, this may not sound good. AI is a part of the ‘digital’ portfolio of tools, different from the traditional, legacy IT applications that have driven revenue so far. Commoditizing of AI and their embedding as part of every IT application will mean a lost opportunity for IT service providers. Like other IT waves we have seen, how much of AI will be out-of-the-box and how much will be customized for the unique needs of an organization needs to be seen and will address the question of employment it generates.

Image courtesy: By Adrian Baer (Provided by Photographer to Devanthro Society) [CC BY-SA 3.0 (https://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons

Note: All points of view expressed above are my personal opinions and do not represent that of any organization.