The buzz around Artificial Intelligence (AI) is huge, which can sometimes mask the real sense and value behind it.
IDG’s 2018 Artificial Intelligence Survey which surveyed 200 UK IT decision makers at the end of 2018, found that 65% of IT leaders were either considering or piloting AI projects, and that 80% agreed that AI would have a transformative impact on their organisation within the next three to five years.
Predictive analytics and deep learning stand out as the most exciting applications, but fraud and security detection and natural language processing (NLP) aren’t far behind.
IT leaders are starting to see how AI can drive their organisations forward.
What are the business objectives around AI?
Their objectives are equally clear. Over half – some 59% - believe AI can deliver enhanced customer insights, while 51% think these technologies hold the key to improving internal processes. On the one hand, AI can deliver improved efficiency and cost reductions. On the other, it can help enterprises attract and retain customers, empowering them to develop stronger relationships and offer more personalised services at scale. Indeed, approximately 31% said that AI would help them get a ‘better understanding’ of their customers.
But these benefits extend beyond better customer service, and greater service personalisation, to the opportunity to also empower employees with 54% of IT managers saying that AI enabled them to free up staff for higher value work.
All of this shows that we’re moving past the point where the promise of AI is theoretical. The US electronics manufacturer, Jabil, is using machine learning to inspect components flagged as defective, reducing false positives and the time operators spend in manual inspections.
Meanwhile, the property insurance teams at AXA XL in the US are using NLP to ingest information from surveys on commercial properties and prepare graded summary reports for the underwriting team - previously a tedious manual task.
Elsewhere and the user cases for improved customer experiences are growing. The Royal Bank of Scotland is now combining AI-powered bots with human representatives in its service teams, using cloud-based AI and NLP to answer customer questions where possible, then handing off conversations where the human touch is required. Customers get the exact services they need much faster – a win for both them and the bank.
The vast majority of companies have scenarios where AI could make a real, tangible impact – where the application of machine learning, predictive analytics or NLP technologies could enhance customer relationships or increase team productivity.
Yet IT leaders are also honest about the challenges their AI projects face. In IDG’s survey, market hype and cybersecurity come across as the biggest sources of concern, but IT complexity and data limitations are also major issues. Most seriously, finding talent with the right AI skills is a big ask. Approximately 34% of IT leaders say they’re held back by a lack of data science talent, while 21% lack the necessary data engineering skills. Some 17% don’t have enough developers in place.
Many IT leaders see no need to start from scratch; over one third are planning to buy an off-the-shelf AI solution, while 29% are looking to invest in an AI platform. Yet while companies are looking to work with existing partners and platforms, these must be the right partner and the right platform.
IDG’s survey also highlights concerns about lack of transparency and vendor lock-in, or about working with a platform that increases IT complexity.
The AI platform for everyone
The concerns and challenges around AI won’t go overnight, but businesses can speed their way to realising the opportunities by working with existing AI platforms. In the past these may have been restrictive, too expensive and too reliant on specialist talent. Yet there’s a new breed of AI platform emerging that makes AI more affordable and more accessible to companies who might not otherwise have the required IT budget or skills.
Salesforce Einstein, for example, can help you automate the creation of machine learning tuned to your specific data-sets and needs, so that you can build data-efficient models with minimal hand tuning and without hiring in an army of data scientists to do the job. This means training an effective model no longer takes weeks, but hours. What’s more, its point-and-click interface now enables more non-programmers to build AI-powered apps.
This kind of feature enables more people to take advantage of AI capabilities, even if they don’t have specialist expertise. For instance, Salesforce’s Trailhead online learning services include courses on AI basics, working with natural language processing, voice recognition, bot development and more. Skills shortages are a perennial IT problem, particularly in AI, but by adopting a more accessible AI platform, you can hit the ground running and develop the necessary skills within your business. In doing so, the old barriers to AI could become a thing of the past.