When Amazon’s Alexa in-home virtual assistant startled customers by emitting unprovoked snickers earlier this year, it appeared as comic confirmation of the online retail giant’s deployment of commercial use-cases leading the adoption of artificial intelligence technologies. Just like customer service chatbots and social media marketing, the descriptive and predictive consumer applications of AI remain in the headlines.
Providers of enterprise resource planning solutions have had a harder time demonstrating to buyers with more at risk from exercises in digital transformation that the application of AI in business is no joke.Opens a new window
Heady Growth of AI
According to a 2017 survey of more than 600 companies worldwide by Tata Consultancy ServicesOpens a new window , most are using AI to mine the huge amounts of business intelligence data they generate; for security, for technical support of system users, for automating production, and to track the use by employees of vendor systems. Having built the core systems and artificial intelligence software that run global business, ERP makers know there is more to be done, and are developing applications and interfaces with AI for the wider spectrum of operations they support.
Underpinned by market forecasts running as high as $400 billion in annual expenditure globally on AI over the next dozen years, the pace of change is heady, with product and platform launches, iteration and application releases and functionality upgrades from whole-of-enterprise providers and business-process specialists over the past 18 months.
Like the headline-grabbing commercial applications of artificial intelligence, these combinations of analytics, machine learning, sensors and robotics are aimed squarely at the user. Whether through automating repetitive tasks and initiating event-driven processes or mining data flows for more accurate decision-making, the goal is to engage employees in pursuit of organizational agility.
Intelligent User Interface
A primary feature of AI-enabled ERP systems is the intelligent user interface, which combines AI process technologies with speech and image recognition and natural-language processing. Products such as SAP CoPilotOpens a new window , introduced in October 2016 by the German ERP giant, free users from having to request data from a range of applications and key in commands as they interact with chatbots to gather information, initiate processes and collaborate with colleagues. Meanwhile, the company’s Leonardo platform enables users to develop applications using AI technologies. SAP also is partnering with Amazon, Apple, Google and Microsoft on enterprise application development and delivery.
Not to be outdone, whole-of-enterprise rival Oracle has unveiled an across-the-board upgradeOpens a new window of its cloud-based applications using AI technologies. Tools for commerce, customer service, finance, human resources and CRM have all been improvedOpens a new window , along with applications for manufacturing supply chain, marketing and salesOpens a new window – and all in ways that leverage insights from the San Francisco-based company’s data cloud, where 7.5 trillion data points are collected each month. Developers can customize Intelligent BotsOpens a new window , a feature of the Oracle Mobile Cloud that also came to market last year, to tailor the user interface for work and dialogue flows and to extract insights from unstructured data.
Cloud services provider Salesforce, also based in San Francisco, is applying AI in CRM, where its Einstein platformOpens a new window uses predictive analytics and natural-language processing to generate forecasts and identify opportunities for touchpoints that can lead to sales conversions without the need for separate tools and spreadsheets.
The company encourages developers to build AI applications, dedicating $50 million in seed money last year for start-ups working with technologies deployed on the Einstein platform. Introduced in 2016, Einstein’s latest upgrade came in March, when Salesforce added an automated query tool to its analytics engine.
Venture Capital Boom
The Salesforce Impact FundOpens a new window , endowed on the first anniversary of Einstein’s launch, is just part of $3 billion in venture capital attracted by US-based AI in 2017, a six-fold increase since the turn of the millennium, according to Stanford University’s AI IndexOpens a new window . First issued in November, the document is part of a 100-year survey of AI’s impact on all aspects of life since the phrase was coined in the mid-1950s.
Led by Eric Horovitz, the former head of the Association for the Advancement of Artificial Intelligence, the AI100 projectOpens a new window monitors financial, technical, academic and social indicators, with periodic evaluations and forecasts of AI’s long-term effects and implications on everything from automation to national security and from psychology to democracy.
Alongside the surge in venture capital spending, the Stanford index reveals that the number of start-ups in the US alone has increased by a factor of 14 since 2000, a trend expected to continue as AI consumes a larger portion of total technology expenditure by companies worldwide. With compound annual growth averaging 40% in recent years, $12 billion was spent in 2017 on AI, estimates Dutch technology publisher IDC, out of a total of $59 billion in enterprise investments worldwide.
Route to Profitability
The distributed nature of AI development is testament to the early-stage nature of the technologies at work in enterprise and reflected in the limited return on investment from AI projects. According to management consultancy McKinsey, which in January released findings from research into more than 600 companiesOpens a new window , the quickest route to financial returns comes from applying AI end-to-end in microverticals in nine areas of the technology stack it identifies as ripe for transformation.
More often than not, the consultancy adds, the use-cases best placed to produce immediate ROI are at the edge (device) level, where performance costs are lower and demand is high for customization rather than in the cloud, where AI investments aimed at scale tend to be made.
Examples in the workplace are varied, with specialist developers targeting specific processes with AI to boost employee productivity. They include WalkMeOpens a new window , a digital trainer that integrates with business software to determine user preferences for applications hosted on ERP platforms such as Workday. The Pleasanton, California-based company advocates a people-centric approach to AI in the financial and human capital management applications and last November unveiled a four-stage AI Maturity ModelOpens a new window to guide customers and developers.
Having the Last Laugh
Despite the relative paucity of immediate ROI from AI investment, industry experts predict a business boom will eventually emerge from rising investment in digital transformation underpinned by artificial intelligence deployed in the cloud. The rollout in 2020 of 5G communications technology, involving low-frequency, low-power edge devices that make up the Internet of Things, will boost ERP use cases and further impact the development of workforce efficiency.
Management consultancy Accenture released a strategic white paper Opens a new window in JanuaryOpens a new window that forecasts the average increase in profit for S&P500 companies from AI adoption at $880 million by 2022. Coupled with an expected 10% boost to employment to handle emerging types of human-machine interfacing, enterprise applications that foster smarter ways of working may well have the last laugh in terms of the fuller realization of AI’s potential in the enterprise.