Close-up of a server with blue lighting in a data centre

Protecting the data you cannot see: what CDOs can learn from Rubrik's latest customer story

Most data leaders can tell you where their most important systems are. Fewer can say with confidence where all of their sensitive data sits. Files get copied into shared folders, passwords end up in documents, and old records stay put long after anyone needs them. Spread that across several clouds and on-premises systems, and you end up with a lot of data that nobody is really watching.

A customer story published on 19 October by Rubrik, which describes itself as the Security and AI Operations Company and is a regular partner of our summits, shows how one organisation set about closing that gap. The organisation is a leading European financial institution and banking group. It is not named, but its starting point will be familiar to any CDO working across a mixed estate.

The visibility gap

The bank's data is spread across Microsoft Azure, AWS, Microsoft 365 and Oracle Cloud Infrastructure, as well as on-premises systems. The story, based on insights from its Head of Core Infrastructure, puts the problem in one line: how do you protect what you cannot see?

Before the change, finding and monitoring sensitive data was a slow, manual job. Overshared files and misconfigured permissions were hard to spot across SharePoint folders and multi-cloud storage. There was no proactive way to fix problems or adjust labels, which left the organisation exposed to data leakage. Manual audits could not keep pace with the amount of data, and the team could not automatically enforce clean-up policies, such as purging aged data. According to Rubrik, that had become a growing compliance risk.

Starting with credentials

The bank decided it needed a dedicated data security posture management (DSPM) tool. It wanted to monitor its data continuously and shrink its attack surface, so that risks could be found and fixed before anyone exploited them.

Its first priority was credential sprawl. Using Rubrik, the team scanned its estate for passwords and other sensitive credentials stored in the wrong places, and fixed those gaps before they could lead to unauthorised access.

"Rubrik is the most elegant way to introduce DSPM," said the Head of Core Infrastructure.

Working with what was already there

Rubrik was brought in to give visibility and control over the estate as it stood, rather than to replace it. By connecting to Microsoft Information Protection, the team automated the labelling of both older and newly created documents, based on its own sensitive data classification policies.

"Rubrik connects flawlessly with our existing infrastructure: no additional product, no new authorizations," the Head of Core Infrastructure said.

The results

The bank now reports 95% accuracy in classifying sensitive data. According to the story, this has directly improved its regulatory standing, particularly against Chapters II, III, V and VI of DORA, the EU's Digital Operational Resilience Act, while reducing reputational and financial risk.

Day to day, the team can automatically find and fix sensitive data, including IBANs and passwords, that has been shared across the whole organisation. Automated risk scoring has allowed it to move to semi-automatic and fully automatic remediation, through Jira tickets and the Rubrik interface. That also closes gaps where high-risk data previously had no proper label.

From the customer's side, the story says Rubrik provided the "complete visibility and control over data risk in the cloud" that manual audits could not. It also let the bank deal with its highest-risk data first.

The same thinking, applied to the lakehouse

Rubrik is applying a similar approach to the platforms that analytics and AI teams rely on. In September it launched Rubrik Apache Iceberg Protection for AWS. The company points out that native snapshots are metadata pointers rather than backups. If a table is deleted, corrupted by ransomware or overwritten by an errant AI agent, those pointers can disappear along with the data. The new capability backs up both the data and the metadata that defines each table, so that recovered tables come back complete and ready to query.

"Data lakes have grown beyond analytics and AI to now run the entire business, and Apache Iceberg is one of the most critical systems built on them," said Anneka Gupta, Chief Product Officer at Rubrik.

Inspired's take

The 95% figure will catch the eye, but we think the more useful lesson for CDOs is the order in which the bank did things. It did not begin with a large data programme. It began by finding out what it actually had, dealt with the riskiest items first, and plugged the fixes into tools and workflows the organisation already used.

That is a pattern any data leader can borrow. A few questions are worth asking of your own estate:

  • Do you know where your sensitive data really lives? Not where policy says it should be, but where it is today, across every cloud, shared drive and collaboration tool.
  • Who applies the labels? If classification depends on people remembering to tag documents, it will drift. Automating it against agreed policies keeps it consistent.
  • Are your retention rules enforced, or just written down? Aged data that should have been deleted is still data you have to protect and account for.
  • Where do the fixes go? An alert only helps if someone acts on it. Sending remediation into the ticketing tools your teams already use makes that far more likely.
  • Could you show a regulator? For financial services firms in the EU, DORA means being able to evidence control. A clear, auditable view of sensitive data makes that conversation much easier.

This story is told from the infrastructure side, which is a useful reminder that knowing your data is a shared job. CDOs who build a common view of sensitive data with their security and infrastructure colleagues will find every one of these questions easier to answer.

Join the conversation

How confident are you that you know where all of your organisation's sensitive data sits? Share your experience in the comments. Questions like these are exactly what our data leadership community gets together to discuss at our CDO Inspired events.

Sources

Image: panumas nikhomkhai on Pexels (Pexels License).