Chris McNulty from the Microsoft yesterday blogged about a whole range of exciting updates that have been released, or will be released soon, for SharePoint Online. This is all pretty awesome stuff – I can’t remember a time when so many great features were announced all at once!
• Flow for Cognitive Analysis – Process any text for sentiment, key phrases, translation or moderation
• Image Analysis – Create immersive formatting for any list or library with scripting
• AI for Images – Teach the cloud to recognize new objects for auto-tagging
Available soon (Summer 2018):
• Build Microsoft Flow workflows with Visio – Model a new process in Visio and export it to Microsoft Flow to activate and run custom processes
• Add file upload to Microsoft Forms – Add a custom question to allow users to supply a file to upload to SharePoint
Available later in the year:
• New ways to create lists – Create lists based on Excel, templates, or other lists
• Quicker list editing – Edit list content in place, and paste data from other sources
• Link list items – Connect list items to Planner, Outlook calendars, locations, and more
• Realtime list updates – See updates to lists and libraries instantly without refresh
• Analyze lists with Power BI – Connect list items to Planner, Outlook calendars, locations, and more
• Row Formatting – Create immersive formatting for any list or library with scripting
Chris’ blog gives an overview of all of these features (and I would recommend checking that blog post out), but in particular I’d like to take a quick look at the ‘AI for Images’ functionality.
The ‘AI for Images’ functionality
This feature is all about automatically extracting data from scanned images, so you can do useful things with it. The obvious example is making scanned documents findable in search, but an even more exciting use case is the automatic categorisation of scans (e.g. receipts), and the processing of them using a Workflow engine such as Flow.
First off, it is important to mention that a flavour of this functionality was released in 2017 as part of OneDrive for Business. However, this newly announced feature appears to apply for actual SharePoint lists – which massively increases its usefulness, as the feature can be used by teams in a collaborative context.
Finding information about this feature is difficult, but I did find a post that indicates that it applies to ‘Asset Libraries’.
I decided to test the feature out, and so created an ‘asset library’ and uploaded some example receipts I found from Google images. Looking at the document properties, there did not initially appear to be anything related to this new functionality (nor in the view settings). To view the new fields, I performed the actions below:
The fields relating to this new feature appear to have the naming convention MediaService[XYZ] e.g MediaServiceOCR (for the extracted text) and MediaServiceAutoTags (for the clever “AI” categorisation of the image). I suspect the new fields are not available until such time as SharePoint has successfully processed an item with OCR content (the new fields did appear in the traditional ‘view settings’ page after a while). As below, the OCR had patchy success for the receipts I uploaded (the categorisation was good, but the OCR didn’t work for receipt number 1).
Despite the fact that this feature is not perfect, I am looking forward to seeing what our development team can achieve with it. The potential applications are numerous, so I’ll report back once we’ve used this as part of a client solution.