Published on the 28/01/2019 | Written by Dion Williams
Will the wait for large amounts of ‘clean’ data stymy your AI plans?...
Many organisations believe they need to get their data in better order before embarking on the AI path. But the reality is that most never really do get their data fully structured – and that’s just fine when it comes to AI, because the technology lends itself to consuming large quantities of unstructured data to deliver insights.
While research shows that most organisations are looking towards AI to drive automation efficiencies in their service organisations, it can be a little daunting as it is quite a step change.
“By taking the approach of having ‘clean data’ first, you’ll never get started on the AI path.”
To get started in the process and avoid suffering from paralysis because of the apparent enormity of the task, businesses can start by taking a small, pragmatic approach to change at the outset and make the shift to AI incremental. For example, a good starting point is a service desk with a large volume of calls. AI can leverage data to resolve calls more quickly, particularly those involving common requests.
By taking the approach of having ‘clean data’ first, you’ll never get started on the AI path.
Data is data – you need to take what you have and get started, and understand what AI can do to get the most value from it.
Organisations might start off with historical data from the service desk, then over time, want to provide contextual understanding of where the person calling fits into the organisation and what their profile is so that the service agent is fully equipped with the knowledge required to meet the caller’s needs.
AI can also be used to unlock organisational knowledge and make it searchable and accessible when and where it’s needed.
Some CIOs have taken a data lake approach (a repository that can hold a vast amount of raw data in its native format until it is required) but this is an expensive way to get started. By starting small you can gradually increase value without having to dig yourself out of a huge financial hole.
One of the benefits of AI is the ability to process and consume large volumes of data that appear not to be related.
Trying to predict customer churn typically involves more than one data set. For example, was it due to a poor service experience by a service agent, an issue with the functionality of the website, or both? The more different and disparate data sources you can bring together the better conclusions you can arrive at and, subsequently, make better decisions about how you interact going forward to reduce that churn.
AI makes Big Data more accessible because it can analyse huge amounts of data to generate valuable, actionable insights. AI is the bridge between harvesting massive amounts of data and using that data to the benefit of your business.
We won’t see a wholesale shift of traditional BI analytics to AI analytics, but rather a combination of both. The smart AI and machine learning-driven prescriptive applications that are now emerging will have a transformative effect not only on sales and support services but also HR, logistics, supply chain management, and a range of other areas.
It is important to do measurable projects with AI in your initial foray, as the presentable outcomes can provide the decision makers within an organisation with the justification and the impetus to fund more AI projects. For example, if it can be demonstrated that AI has delivered greater customer insight, enabled your organisation to predict customer behaviour and improved customer service while reducing costs, it stands to reason that further investment in AI across the business will be looked upon favourably.
Lastly, you don’t need to wait for the purpose of data to become apparent before retaining it. Intelligent AI technology will be able to analyse your data and generate actionable insights.
The key takeaway is you don’t need massive volumes of data to start – that is a disconnect from what most people assume to get AI working. AI simply makes the process of analysing massive amounts of data much more straightforward, but it does not require huge amounts of data to transform a business in a positive way.
Dion Williams is the CEO of Soapbox.ai