If data is the new oil, artificial intelligence (AI) can arguably be its best drill, able to uncover insights and mine real business value from the huge and complex data sets that typify modern organisations. Enterprises are not blind to the massive opportunities that can be extracted: according to the latest Gartner data, enterprise adoption of AI has grown 270% over the past four years. In the last year alone, AI adoption has essentially tripled within enterprises of all sizes.
That’s not surprising considering 85% of global CEOs believe AI will fundamentally change the way they conduct business within the next five years. Until recently, the majority of business decision-making was predominantly driven by human centric capabilities. Much of it involved intuition, experience, judgement and a reliance on skills and expertise. Business leaders were more likely to rely on a mix of these elements and a healthy dose of ‘gut feeling’ to make decisions over certain strategic objectives, key business outcomes, product or service capabilities, and more. But the human brain is error-prone and subject to certain inescapable biases.
Data collected by computers over the past half a century has enabled more evidence-based reasoning and decision-making. But it has – until recently – still required human processing. That was until the commercialisation of AI. What AI offers is an opportunity to process vast amounts of structured and unstructured data accurately and without any of the cognitive biases from which the human brain suffers. Deployed effectively, AI gives organisational decision makers the ability to greatly improve their decision-making and creates opportunities to automate processes, boost productivity, uncover meaningful insights from large data sets and improve overall innovation efforts.
Where is AI being used? According to the IDC, IT operations is the leading business area for AI deployment, followed closely by customer service and fraud and risk management. Top industries include healthcare, which is predicted to see a 49.7% compound annual growth rate between now and 2026 to reach over $8-billion in value.
Integrating AI into your business is not as easy as 1-2-3. A quarter of global organisations that are already using AI report a failure rate of up to 50%. These organisations cite lack of skills, unclear business value and unrealistic expectations as the main reasons for this failure.
There are ways to improve the success rate of enterprise AI incorporation. Integrating your AI deployment into a broader Intelligent Enterprise strategy that rapidly transforms data into insight and supports process automation, innovation efforts and great customer experiences can accelerate time-to-value and increase the opportunities for AI to deliver value.
For enterprises of any size, it all starts with data. At an enterprise level, a successful incorporation of AI into the ecosystem largely depends on the quality of the data that the algorithm has access to, and the strength of the platforms that provide that data. To establish a strong data foundation that can continuously utilise to deliver AI-imbued business value, enterprises need to look at five key aspects:
1. Defining an integration strategy that allows the organisation to embed AI into the end-to-end business processes of the organisation; 2. Establishing a holistic data platform that eases the process of data management across large and complex organisational structures – ensuring the one version of the truth everyone agrees on inside the organisation; 3. Developing a full understanding of the organisation’s data, including where it comes from, which business processes it represents, quality levels, and the Five Vs of big data: volume, velocity, veracity, variety and value; 4. Defining appropriate governance and compliance policies for internal and external requirements, and building governance controls into data management operations; and 5. Ensuring a positive customer experience by simplifying the use of analytics and visualization tools, encouraging self-service to speed up adoption among internal end-users.
Once the data foundation is set, organisations can move on to planning their AI deployment. It is important to get a clear view of the business outcomes you’re trying to achieve using AI. As the IDC research illustrate, many companies who report AI failures cite unrealistic expectations and a clear lack of business value as primary causes. Be clear about what business problems you’re trying to solve - for example poor customer service, inefficiencies in the supply chain, or lengthy financial reporting timelines – and ensure the planning and deployment are focused on solving these problems.
Once you know which business outcomes you want to drive, using AI as a means to achieve this end, determine the business impact. Value needs to be measured and effectively articulated for it to have true impact on the organisation. Having a clear view of how value will be realised and how it will impact the business along the metrics that matter to the business will ensure higher levels of organisation support and buy-in. This, in turn, will help accelerate success.
Be realistic about the resources you have at your disposal to make the project a success. These could be talent resources – having experienced data scientists, systems integrators and business analysts available to implement the project – or financial resources. Depending on where the enterprise is with its digital transformation strategy, there may also be some technology constraints that need to be taken into account.
Supplement internal resources with appropriate partners and assign a team to run a pilot project to test how the team would perform and where additional resources may need to be deployed. Developing a proof of value and live demo of how the AI implementation would play out can help the enterprise avoid costly mistakes and wasted effort. Facilities such as SAP’s Co-Innovation Lab give enterprises and their technology partners an opportunity to see how the interplay of AI, data and business processes could deliver the intended value before full development and implementation starts.
Once things are running, test, measure, analyse and refine. There are no secrets to success; the organisations that see value from their AI implementations are ones that are unafraid to test new ideas, try new approaches and reimagine how they do business – all with the support of AI, of course.
In all things, don’t forget the customer experience. No deployment of technology can be considered truly successful unless there is strong end-user buy-in and participation. The key is to put the customer and their experience truly at the center of what you do. Then you build your value proposition around that center and work your way back to the technology that enables that differentiated experience! And with new measurement tools such as Qualtrics, which can help organisations marry experience (X) and operational (O) data: Empowered with data about the quality of end users’ experience with the AI deployment, the implementation team can more easily make changes to encourage buy-in
Lastly, AI adoption is ultimately predicated on user trust. Trust in how the technology is developed, trust in how the technology is used, trust in the value the technology can offer in the moments that matter to users. This trust is foundationally based on a governing set of principles in the areas of ethics, privacy and security. Whatever AI solutions you design and deliver, must be grounded by the principles that promote responsible use to ensure a greater chance of success.