Information governance is boring. There, I said it. And though everyone should care, they typically don’t.
I remember the boom in records management after 9/11. One year later, the interest had faded back into obscurity. Just like few people want to do housework, few want to take on the tasks involved in information governance — you clean things up only for people to make a mess again. So it would be dishonest to say that we can ever make governance sexy, but we can at least make it much less boring and far more effective than it is today.
Information Governance Defined
Let’s start by getting a grip on what governance means. Information governance is applying rules, processes, procedures and bringing order to the management and storage of information. Historically it has been reliant on employees and dedicated staff defining, applying and monitoring governance activities. But technology is changing fast, and the volume, variety and velocity of enterprise information is also rapidly growing.
As governance, by definition, involved creating and enforcing rules, doesn’t it make sense to make more use of technologies like machine learning (ML) to help here? Don’t get me wrong, such ML-driven systems exist, but their use in this area is still relatively rare. ML can interpret and apply rules and adapt to complex rule changes at a speed and volume impossible for humans. So while ML-driven information governance systems are available, information governance professionals don’t appear too keen to use them.
Related Article: How Digital Workplace Governance Supports Agility
Learning From a Machine Learning Governance Case Study
I recently spoke with a company that has made wide use of ML in its records management system. Though the technology itself was fascinating from a nerdy standpoint, their battlefield experiences of deploying it were more fascinating still.
The team originally pitched the project as a records-related project. This elicited zero interest. One year later, after dropping the term “record” from all documentation and replacing it with the word “file,” the result was immediate buy-in from senior management concerned with risk. It’s a somewhat simplistic semantic interchange, but the results speak for themselves.
Similarly, when the project was pitched to the records management team, the pushback was enormous as the team saw it as the automation of their jobs. Pitched again a year later, with the immediate disclaimer that machine learning (they had dropped AI, the original term used) was useless without their expertise, the project was enthusiastically embraced as something they could grasp onto to save their jobs and raise their status.
To recap, a governance project was pitched twice. The first time it fell flat. The second time the exact same project was pitched with some slight tweaking of the messaging, it was embraced by all. A governance project that nobody wanted to touch became something exciting and empowering.
Here’s the thing: this is not a one-off situation. What happened at this company has happened with slight variations multiple times over in the last few years, and there are a couple of key lessons for anyone trying to promote or sell governance within their organization to grasp.
Related Article: Measuring Information Governance Success
What Records Managers Can Learn
The first lesson is that nobody cares about records management. Let the hate begin, but it’s true. Even so, it is sad. Records managers are under-appreciated and often misunderstood. They provide an extremely critical service that seldom gets the credit it deserves. But the reality is that as the sheer volume, variety and velocity of data and information has grown at exponential levels, traditional RM practices have failed to adapt to meet the challenge.
As an industry analyst, I talk with countless buyers and sellers of RM technology. Without exception, they tell me it’s an increasingly uphill struggle to gain attention, budget and executive sponsorship. Yet that’s not the case when it comes to managing risk and compliance. Or to be more accurate mitigating risk and non-compliance, as these topics do get executive attention. And if that risk and non-compliance mitigation can be automated with ML, all the better! Hence just tweaking the terminology from records to files (or data) can, in turn, flip the attention switch on. That’s hardly surprising as regulations such as HIPAA and GDPR are now emerging as just the tip of a growing regulatory iceberg that is set to be enforced far more rigorously than in the past.
Similarly, telling record managers that AI can and will do their job, as well as if not better than they can, will not go over very well. Trust me, I know. I gave the keynote in Nashville to ARMA (American Records Management Association) members a couple of years back, and let’s say the response to my message was mixed …. Though the facts were accurate enough — that the principles of RM are rules-based, and ML can apply rules faster and more accurately than humans — nobody wants to hear someone telling them their jobs will be automated. In my defense, though that may have been what people heard, the message I was giving was more nuanced. ML will automate much of the work of RM, but without genuinely expert guidance, the ML tools can do nothing. Ultimately, the world of AI and ML is nothing more than many tools and a mountain of data that predict probabilities. Where it excels is when it is augmented with human expertise to guide and work with it. Record managers can never even hope to manage the information deluge swamping their organizations. Simply sifting the wheat from the chaff is itself near impossible. But by adapting and working with ML, they can do it better than they ever have before.
Information Governance Is Never Going to Be Sexy. But It Will Always Be Needed
Information governance may never be the sexiest of topics, but modernizing practices, procedures and tools will help meet today’s real-world challenges. Mitigating risk or seizing opportunities to advance your career creates a moment in time to grow and develop. And such opportunities don’t come around all that often, so grab them while you can. Reframe and update the conversation within your organization, and you may well be pleasantly surprised at the response.
Alan Pelz-Sharpe is the founder of Deep Analysis an advisory firm focused solely on disruption and innovation in Information Management. Deep Analysis provides research and guidance to firms looking to leverage new technologies and to make a digital transformation.