Almost any business will tell you, with very good reason, that its most important asset is its people. A services provider without people is just an empty office, a manufacturer just idle heavy machinery, and a health system just a collection of empty beds.
This truth is not going to change, but in recent years the picture of an organization’s value beyond its people has been fundamentally transforming. Where different sectors might once have given very different accounts of where their value lies, today the second greatest asset is becoming, in the majority of areas, data.
More digitalized tools, infrastructure, and processes which are more highly connected have delivered an explosion of data which offers deeper insights into customers, more efficient operations, higher quality innovation, faster decision-making, and more besides. Data, in short, is promised to enable all the agility and responsiveness that businesses need to drive growth.
The promise of artificial intelligence
Of course, as any CTO or CIO tasked with leveraging data into a competitive advantage knows, raw data itself is only a promised, hypothetical benefit. The exponential growth in data does not arrive in an easily-usable form, precisely because it is coming from everywhere. Sources like connected devices and physical sensors, financial transactions and website user journeys, social media chatter and market trends all collectively emit a vast, chaotic stream of information for businesses to manage.
That chaos leaves technology leaders juggling two significant pressures. First is the spiraling cost of storage. Knowing that value may be embedded in all of this data incentivizes keeping as much of it as possible around, which – especially with heavier formats like video and audio – can lead to storage bills running into the millions per year.
The growing weight of IT costs itself exacerbates the second major pressure: rising expectations around finding value in the noise of business data. In particular, businesses are increasingly seeking to respond to the overwhelming nature of modern business data by deploying AI, aiming to find workflows with it that can (unlike human-driven processes) scale seamlessly with data volume by adding additional computational resources.
However, this leads to something of a paradox: while emerging AI tools certainly have the ability to deal with data at scale and make it valuable, those tools are as a rule only as good as the data they are given. Bad, irrelevant, or incorrect data, stored in unhelpful or contradictory formats or locations, will not deliver the AI-driven value that businesses expect. The AI answer to business data challenges relies on first solving some of the challenges that businesses experience around organising and managing their data in the first place.
Understanding what matters
The idea that the quality of an AI solution’s output is limited by the quality of the data you feed it has been a common statement since the latest boom in AI technology kicked off a few years ago. ‘Garbage in, garbage out’ has been a saying in the IT industry for an extremely long time, and it remains true today.
However, that leaves an open question about what ‘good’ data actually looks like for businesses seeking to use AI to find those growth-enabling advantages that it promises. One way to look at this is to go back to that fundamental fact that the most valuable thing a business has is its people, and so the data has to work for the people who need it.
We know from experience of working with large organizations to transform their data strategies that, typically, half of all data on record is noise. That might be duplicated information, outdated information, or information that never needed to be stored in the first place. Around another quarter of an organization’s information will be necessary for paper trails rather than actual application uses: knowing the journey that a piece of information has been on is important for an audit, but not for the end user.
And then, not all of the remaining data which is relevant and necessary is of equal status. Details of a customer’s unfulfilled order need to be instantly available, for instance, but invoices from a decade ago can probably be placed in slower-to-access cold storage in order to save on expense and energy.
We’ve likely all been in situations where getting the information we need at work means combing through layers of unhelpful or unnecessary files and records. If a business wants to deploy an AI solution like a chatbot to empower employees or customers to access that data more effectively, that underlying challenge doesn’t go away: whether it is operating on a traditional database or a lake of unstructured media data, the AI solution needs access only to what actually matters.
Step one, therefore, is making sure that the data is relevant and well-managed from a human perspective.
The useful data a business generates might be in the minority, but it is by no means small – and it will continue to grow at pace. Scalable AI solutions will be vital for getting value out of that data, but the process starts with a transformational approach to the organization’s data strategy that puts the foundations for success in place.
We list the best employee management software.
This article was produced as part of TechRadarPro’s Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro