Fifty-five percent of enterprises are set to make major investments in artificial intelligence (AI) by 2020, according to a new report.
A further 25 percent plan to make minor investments to determine AI’s value and possible ROI over the next two years, it says, suggesting that only 20 percent of organisations have no plans to explore the technology’s potential.
The findings come from spending management cloud provider, Ivalua, which commissioned a study of the plans of 400 finance, procurement, and supply chain business leaders. The research was carried out by Forrester Consulting in the US, UK, France, and Germany.
The research reveals that those leaders believe that one of the biggest challenges to adopting AI is the poor quality of enterprise data. Well over half of respondents (59 percent) said that low-grade data will make it impossible for AI to make accurate and informed decisions, undermining their ability to extract value from their investment.
Improving the calibre of that data is no easy task, however, with respondents saying that their inability to access data (44 percent), lack of normalisation between data sets (43 percent), and inaccurate data (41 percent) are the biggest challenges.
In addition, 36 percent cited information overload and lack of internal skills as further obstacles to making sense of their data.
Many also claim that they aren’t receiving the promised benefits from AI due to the immaturity of vendors’ applications, with 62 percent citing this as a problem – suggesting that many AI vendors’ marketing campaigns are more advanced than their technologies’ capabilities.
However, the research found that not all of the challenges are technology or data related. Forty-four percent of respondents said they do not have the support of C-level executives within the organisation to explore the opportunities of AI. There is also uncertainty about how to apply AI to certain use cases among nearly half of respondents (47 percent).
So what of those use cases?
Exploring the impacts
In terms of AI’s practical application, respondents believe it can have the most impact in alerting the enterprise and its partners to supply chain disruption (44 percent), recognising and flagging supplier compliance issues (39 percent), and quickly identifying instances of fraud (37 percent).
In addition, respondents think that AI adoption will lead to greater automation of menial tasks, making them actionable in minutes or seconds, instead of hours or days. Two of the biggest areas for this will be invoice processing (51 percent) and purchase approval (35 percent), found the survey.
“There is clearly a huge appetite for AI, and this will only increase as more relevant applications and success stories come to light,” said David Khuat-Duy, corporate CEO of Ivalua. “But when investing in AI, it’s important that organisations address challenges that will otherwise limit value.
“Driving accurate insights from AI is reliant on having a solid data foundation from which to work, and the findings show that this remains a significant obstacle for most organisations. Success requires organisations to simultaneously address enterprise data problems when investing in AI.”
He added, “Ultimately, if organisations can improve their data quality and address other challenges, they can tap into a wealth of AI benefits. Whether it’s automating low-value tasks or providing rich insights, AI can have a transformational impact on procurement and supply chain operations.
“For example, AI offers huge potential to enable smarter procurement, which can create efficiencies and enable better decision-making, offering a real competitive advantage to those that adopt.”
Internet of Business says
While the plans of every organisation can’t necessarily be inferred from a study of finance, procurement, and supply chain leaders – and it is hardly a revelation that such a survey base would see promising applications of AI among those business functions – this is a useful and timely report.
For one thing, it reveals a curious mix of willingness to deploy the technology, worries over data quality and the technology’s efficacy, and challenges over strategic support. Take a step back from the statistics, and this suggests a buy-side market that is being led primarily by hype and tactical manoeuvres, albeit fed by genuine hope and ambition.
This throws down the gauntlet to vendors to deliver against their promises, and to buyers to understand that AI is only as good as that data it is fed with – as explored in our recent report about MIT’s AI research programme.
Factor in data privacy, transparency, and informed consent – in the wake of GDPR’s introduction on 25 May – and many organisations must now be realising that their core priority is to put their ‘data house’ in order first. Common sense must precede the adoption of artificial intelligence.
The days of letting AI lose on a world of mass data without any repercussions or responsibilities were short lived – if they existed at all.