Monday , June 25 2018
Home / AI / The Enterprise AI Promise: first the machine, then the person

The Enterprise AI Promise: first the machine, then the person

Artificial intelligence (AI) is a buzzword promising an unbelievable future, but there are plenty of challenges facing early adopters so far. It is arguable that AI may not achieve mainstream traction unless organisations plan for its adoption.

The Enterprise AI Promise: Path to Value” study by SAS examines Enterprise Readiness for AI, through interviews with representatives from businesses across EMEA and from a wide range of industry sectors. Findings suggest that the vast majority of organisations have begun to talk about AI, and a few have begun to implement their own projects. There were some big differences in levels of practical organisational readiness, in terms of the data science expertise and suitable platforms. Most organisations had given some thought to the problem. Those who had already started work on AI often had inhouse teams and platforms in place, or arrangements to source them. Those who had not yet started, but felt that exploiting AI was essential, suggested that they were most likely to opt for cloud-based solutions and/or consultancy support. Many organisations are still experimenting with processes to approve AI projects.

Current status of AI in organisations

“Respondents confirmed our views that it is still very early days for AI. Most of our respondents had taken at least a few tentative steps down the road towards use of AI, but these were often pilots or proofs-of-concept, rather than anything more substantial. Projects mentioned included building general analytics capacity, setting up dedicated units, and very specific pilots. Most of the investments and initiatives were also at an early stage,” the researchers explain. The biggest challenges associated with AI are all about general awareness of, and acceptance of, technology and change. “What challenges do you expect to see as more AI is rolled out by business, government and perhaps even individuals? The responses to this question were probably the most detailed, suggesting that people are worried, but in a fairly broad manner. The general feeling was that AI would affect jobs, but it was hard to predict precisely how, and across which geographical areas,” the paper confirms.

It was likely that any changes would have a significant lag on the introduction of the technology, which might create problems, both practically and ethically. According to the report, business cases including ‘AI DNA’ were driven by the need to provide growth potential, keep up with the competition, or provide cost savings through efficiency. These early and enthusiastic adopters hope that they will see rapid benefits; their competitors hope to learn from their mistakes.

Among AI drivers and stakeholders central data science teams were identified as the most popular form of sponsorship for AI projects, although in some organisations, individual business units were driving their own initiatives. Platforms readiness Adoption of AI mostly depends on the readiness of organisations, which means having the tools and skilled staff in place to be able to exploit the technology effectively. “We wanted to know about two areas in particular: data science skills within the organisation and the availability of suitable platforms for analysis, as these give a very effective picture of the organisation’s approach to AI,” researchers clarify.

When reviewing the responses, it is clear that a minority of organisations felt that they and their teams were absolutely ready. Some organisations had a data science team, but recognised that they still might not have the necessary skills in-house. Teams were actively training their current analyst teams through attendance at workshops and conferences, as well as studying and keeping up to date with developments in the field. Other responses spotlighted the importance of expanding data science capacity to manage AI. Even if they already had data science teams, they wanted to recruit more data scientists, recognising that there would be additional need for data science in future.

Looking at the capability of various platforms to handle expected data management, the majority reported platforms not to be ready. “Creating an architecture to support AI is about creating a modern platform for advanced analytics, and means being able to support all steps of the analytics lifecycle. Alongside the technical capabilities, it is also important to support the process around the analytics,” the paper explains.

And, finally, the study focused on what respondents think on how soon AI would have an impact on everyday life. “This gives some idea of how seriously they think that AI developments should be taken, and whether the conversations need to happen now or later,” the study explains. The fact is around one fifth said that it was difficult to estimate a time frame, and slightly more said that they thought changes were decades away. These respondents were therefore likely to see little urgency in the situation, and have much less appetite for practical conversations about platforms or capacity, although they may be interested in longer-term, more theoretical discussions about ethics.


If to consider the key report findings – from self-driving or connected cars, through to virtual assistants, chatbots and decision making support, one thing was the core: AI would be used to augment human capacity, not replace it. “At its simplest, AI will take over some parts of some jobs, but will free up human time to concentrate on higher value work,” researchers conclude.

However, humans will need to take over when things get tough. Global banks, for example, are experimenting with chatbots in customer service operations. The chatbots can deal with simple queries like password resets, but when conversations get more difficult, or start to become circular, bots hand over quickly and efficiently to a human operator. So we might think of this as a sequential partnership: first the machine, then the person. In more complex cases, and as AI technology matures, the AI–human partnership is likely to be different. Instead of being sequential, it is likely to be a more traditional partnership, bringing together the skills and capabilities of both partners to create a new ‘whole’, greater than the sum of its parts.

Kate Goldfinch, Managing Editor of The Fintech Times

Check Also Partners with Exchanges and AI Trading Platform to Remove Existing Barriers to Cryptocurrency Investment

While developing its revolutionary blockchain-based KYC platform, has also been busy forming partnerships with other blockchain … | |