Our experts answer questions raised in the chat during our recent webinar Keep or Delete? Why you need an effective data lifecycle management strategy for Microsoft 365 and how to create one.
What roles does data quality monitoring play in the data lifecycle policy for maintaining the effectiveness of Gen-AI tools in Microsoft Copilot?
Data quality monitoring is a critical component in the data lifecycle policy, especially for maintaining the effectiveness of Generative AI (Gen-AI) tools like Microsoft Copilot. It involves the continuous assessment and management of data to ensure its accuracy, consistency, reliability, and overall quality.
It also supports Gen-AI tools, by ensuring accurate and reliable data- by regularly profiling and validating data, data quality monitoring ensures that the information used by Gen-AI tools is accurate and reliable, which is essential for the tools to generate meaningful and correct outputs.
Why is it important to remove old content* in the context of Microsoft Copilot and AI?
*Data from legacy systems that have been replaced or upgraded, which may no longer be relevant to current operations.
When using M365 Copilot, responses are formulated with consideration to your existing content. Should you have redundant and outdated content, it will result in this content being incorporated into Copilot responses. Content which was retained with the intent of remaining dormant could now unintentionally become active again.
How does Microsoft Copilot leverage advanced natural language processing and machine learning algorithms to mitigate the impact of ‘garbage in, garbage out’ on predictive accuracy and user output quality, and what specific data sanitisation techniques are employed within the data lifecycle management strategy to ensure optimal performance of its Gen-AI tools?
Microsoft Copilot and similar Gen AI tools cannot distinguish between redundant, outdated and trivial (ROT) information, which is why we strongly advocate for working on retention schedules and programmes to balance regulatory and information management requirements, starting with the most valuable important and risky information first. This will improve accuracy value and reduce risk on the Gen AI’s abilities and the sooner you can get going with the Governance and Risk teams, who will then bring in stakeholders/content owners from the various areas in the business to ensure the information value is available and promoted. These could be stakeholders from HR, Finance, Contracting etc. and you may even want to deploy the Gen AI tools to departments who have participated in the process first.
In reality it’s hard to take action. An organisation is a living thing, with people leaving and joining – file ownership changing. To dispose of a file – ownership is very hard to find out- how does Purview help with this?
There are two main items to consider with the data ownership challenge most organisations face in their data governance journeys:
- We always start with Role Based Access Control (RBAC) best practice & discipline, keeping users in groups wherever possible and then controlling access to resources and files assigned to these groups. In this way you limit individual ownership of content as far as possible. Ideally OneDrive and email being the only exceptions to this.
- The second strategy, is to use Microsoft Purview to focus on the sensitive information types (SITs) in the content itself. Defining and developing your organisations own SITs is critical. In this way you can protect and retain information based on what it contains and overcome the limitations of location or file ownership. Sensitivity and Retention Labels in particular transcend these challenges by applying the control directly to the document based on these SITs (i.e. it lives with the document regardless of where it goes or who has access to it or even “owns” the content.
When a Team gets deleted, what happens to the underlying sites, Team site, channel site etc, in SPO? Are these deleted as well?
With no additional retention policies or retention labels in mind, when a Microsoft Team is deleted or underlying structures, like the mailbox, SharePoint Site and channel are also deleted. It initially is “Soft Deleted” which means the Teams and all content, can be restored within 30 days of deletion. If there are any Purview Retention Policies or Labels applied to and files or items in the Team, when it is deleted, the Teams and all content will still be deleted from and end-user perspective, but the files will be retained until the retention period has lapsed and would be discoverable in the eDiscovery functionality in Purview.
Instead of deleting a Team, you also have an option to Archive a Team natively, but it means this inactive Teams will still consume SharePoint storage. Third party solutions like AvePoint’s Cloud Backup and Archiving products, offer better control & cheaper storage cost for retaining your legacy content for longer. It is especially useful, if you need to be able to recover and access content from deleted Teams past the native 30 day limitation.
