Understanding the new Microsoft Syntex announcements at BUILD
Thank you to our episode sponsor Specops. For a Specops free security audit, click here: https://password.specopssoft.com/unplugit
Transcript
Stephen Rose
This episode of UnplugIT is brought to you by Specops Software. Specops is an easy to use password management and authentication solution for your active Active Directory. Their breach password protection software actually blocks your end users from choosing any one of over 3 billion known compromised passwords. And that’s a list that’s updated daily. With Specops, you’re creating a password policy with custom parameters and extending the functionality of Group Policy all while simplifying password management. Finally, Spec Ops provides dynamic and user feedback at password change. This means your users are educated on the company’s password best policies without them having to call help desk. And hey, one less call to help desk makes every IT pro happy.
Specops is offering our viewers of UnplugIT a free password vulnerability audit. All you need to do is visit the URL right there and you’ll be able to pick it up. So hey, Specops, check them out today… That’s good password management.
Stephen Rose
Hey everyone, welcome to another episode of UnplugIT. I’m your host Stephen Rose, thanks for joining. I was at Build a few weeks ago and amongst all of the great announcements, there was one on Microsoft Syntex and that Copilot was coming. And I thought to myself, What the hell is Syntex? I really didn’t understand it and how Copilot would make this even better.
So I sat down and did some research. And what was interesting is that I found out that every day, 2 billion documents get added to Microsoft 365. That’s a lot ! And if you think about all of that data going in, how do you find and process and organize and understand all this so it’s not a big chunk of content ? How do you get it to scan and categorize and organize all this into a smart repository? And that’s where Microsoft Syntex comes in.
And I thought, if I’m going to talk about Syntex, I need an expert and I just happen to know one. Mr. Chris McNulty, who is the Director of product marketing for Microsoft 365 and specifically works very closely with Viva and Syntex. So, Chris, welcome.
Chris McNulty
Oh, nice to be on the show. As always, good to catch up, Stephen.
Stephen Rose
So, before we get into any of this, how much does Syntex cost us? So, before we talk about all the cool stuff I can do, because I know people are going to go to spend all this time and tell me how it needs a new license or something like that… So, as a base, what if folks need to get started with Syntex?
Chris McNulty
So, for the most part, you don’t need to do anything up front. So, Syntex predominantly for us really represents establishing a new way of doing business with our customers for modern work. So rather than need to buy seat licenses for Syntex, it’s important to remember that Syntex is a family of about 22 interconnected products. So Syntex does things to help you enrich content with some AI processing and tagging, help you manage content with things like backup and archive, and helps you build applications using repository services. And all of these are needed services. So one of the things that we heard when Syntex was in preview from customers is it’s great for the thousand people in my company who are claims processors or financial analysts who do the same thing over and over every day, and they need this privilege. But for every one of those users, there’s 20 more who might need to do those things once a month or maybe once a quarter. And so, the philosophy that we have is to use consumption pricing, so that you pay a… If you have today any Microsoft online plan that includes SharePoint, you have access to Syntex. You need to activate it inside the admin center, and then it’s just a couple of cents per transaction based on what you’re doing.
Stephen Rose
Got it, okay. So, who is going to benefit from this? If we take a look at different verticals, and this also goes not only to folks who are storing documents in SharePoint, but also the ones who are doing lots of scanning, like when you get into medical and legal, where they’re scanning in lots of forms, this is going to be a tool that’s going to help you to grab all that data, kind of like an advanced OCR, but you’re able to do so much more with it. Is that a correct interpretation?
Chris McNulty
Yeah, I think that’s that’s fair. Now, we see Syntex taking off the most in the industries that tend to have the most complex information requirements. So, think of areas like financial services or healthcare or manufacturing.
One organization kind of intersects a few of those: TailorMade, the golf manufacturer. They looked around the firm and realized that they had lots of people doing things that shouldn’t be their job. So, their attorneys were spending countless hours reading and tagging and filing documents relating to intellectual property or their contracts or lawsuits. They were becoming document managers instead of being attorneys. And so, they were able to use Syntex to build these no-code AI models that can recognize each of those different kinds of content that attorneys work with, whether it’s a patent or a contract or or a motion paper for court. Being able to read those, tag those, secure those, and get them into the right places automatically so that the attorneys can do their job. And that’s a pattern we’ve seen across other industries like financial services. TailorMade also uses Syntex to optimize all of their accounts payable processing, looking at things like invoices and POs and statements of work, and that kind of transits all the way over to a group like the London Stock Exchange.
The London Stock Exchange has a group of about 200 financial analysts whose workload had grown to where each of them was spending more than half of their week just reading these lengthy financial disclosure documents so they could scan them, tag them, file them to be able to build a knowledge base and trigger workflows about when they needed to go get more information.
And with Syntex, they were able to reduce that amount of document processing time for each person down to less than an hour a week. Again, letting them go back to being financial analysts instead of readers and tag writers.
And so we’re seeing that, you know, a third area where we see this is kind of across multiple industries. But for large organizations, you have HR firms, and you may be taking on lots of people in the course of a week : 50, 100, 500. And there’s a major north american consumer and manufacturer who brings in about 500 new employees a week, and they use Syntex, content assembly automation to be able to build out all of the key documents that they need to onboard a new employee. You may be eligible for a stock plan, so you need to get that included in there. You may be a factory worker that requires a different health plan. All of that conditionality, being able to build templates, sort of the reverse of a document modeling process. You take unstructured documents and be able to put structure on them. If you already have data structures and you need to generate one, ten, a thousand, ten thousand similar documents with flexibility, that’s what they’re using to onboard all their employees every week.
Stephen Rose
Okay. So let me see if I can break this down. So, we have site template solution, so this is great, this allows you to sort of track business documents, things like accounts payable. You can generate routine documents, so this can be a template that has to be done on a regular basis like timesheets or maintenance reports. You can generate a large amount of documents in bulk if you need to if you’re going to be changing some processes and want to make folks aware, take in incoming documents and then process all that to make it easier to find and even find specific details.
Now, is some of this needed, let’s say, for compliance where you have to make sure that certain things are in there or documents that have this information have to be managed or stored differently. Can it figure that out, leveraging kind of the intelligence within it?
Chris McNulty
Right, absolutely. So when you are structuring these models, you can also use them to invoke the labeling system from inside Purview. And if you’re not familiar with that, what that means labels are pieces of metadata with, as my good friend Dan Holme likes to say, they have superpowers.
There are two kinds of labels inside Purview: There’s sensitivity labels which relate to security, and retention labels which relate to compliance. When you attach these tags to a piece of content, on the sensitivity side, you may say that this document is confidential, which may mean for your organization that it should always be stored and downloaded in encrypted format, it’s shareable, it should be watermarked, whatever those handling rules are around that. And that’s what a label can do for sensitivity. Retention relates to records management. How do you build rules to say this kind of information? It may be a contract, it may be a proposal, it needs to be kept for three years, five years, ten years, and what happens at the end of that?
So, you can apply those things interactively. You can apply them with default settings, but Syntex, when it’s modeling those documents, can automatically add those tags for you. Most information that we find when we think about data breaches today, an awful… a growing number of data breaches are happening because of misclassified information. It’s put in the wrong place, It’s too hard… You know, if you were taking in your share of those 2 billion documents a day, it’s too hard to expect you’re going to read and manually classify each of those. Any tools that we can bring we think are helpful to help people reduce their exposure and be able to set the right compliance and security on their content.
Stephen Rose
All right. Let’s take a break and let’s take a look at a demo, because I think taking a look at this is going to help folks to better understand it. So let’s go to a demo.
Chris McNulty
Thanks, Steven. So let me just set up some of the things you’re about to see. So Syntex, as I mentioned before, introduced last fall to a broad family of AI powered capabilities for your content. Some of those help you get more value from your content by using AI to be able to pull information out of content, generate new content or drive business processes.
And so we’ve talked about processing and assembly and some of the other capabilities for working with images, OCR, even e-signature. Secondly, with more information, you need more tools to help you manage information at scale, and that’s where Syntex is supplying new capabilities to help me manage information throughout its lifecycle and governance access and administrative controls to make sure information is properly protected.
And lastly, all of that enriched information is even more valuable if I can bring it to more applications and before, especially ones that you may need to build, go outside of M365. And so that’s where we’ve introduced new capabilities that are currently in private preview to help you build new shapes of applications to extend your content for M365 beyond the traditional boundaries of the tenant.
Stephen Rose
When we talk about AI, we’re not talking about, you know, ChatGPT and stuff like that… Are we talking about the graph as that AI layer?
Chris McNulty
Well, the graph is part of that AI layer, but there’s even more that we’re doing, which I’ll show you here in just a second. So let’s start by looking at some of the things that Syntex can do for you. Syntex provides access to powerful AI directly in the flow of work in the apps you use every day.
So here I’m inside a SharePoint environment. I have multiple documents that I’ve received that are in other languages. I can dynamically translate those with Syntex to be able to build new content in English. Syntex can also invoke summarization to be able to extract key points and an index of where to find that in the information and be able to extend that generated content into environments like e-signature.
Stephen Rose
All right, Now I get it. It’s able to do this. So, Copilot was announced at Build that this is going to be coming and we’ll talk about availability in a minute. How does that fit in with all of this? Because Copilot seems like one of those things like you’re in the app and can you do this for me? And then it does it and does that for you. Is this sort of the same idea with Syntex, except are you teaching it? Give me the overview and then I want to take a look at excuse me, at a demo.
Chris McNulty
Absolutely. So if you’ve seen Copilot before, Copilot is using, you know, the power of your own words to help you interact with information, build new information and the rest. And so with that rich range of services that we’ve already created for Syntex, it just makes sense for us to extend those into Copilot.
Here. I’m inside of a document library in Teams. I invoke Copilot and I have a list of sales quotes, but I want to be able to add more content to it. So, I can add a new column dynamically, and Syntex is then able to dynamically model that information, propagate values down the column for all the information. That enriched information makes it easier to run searches to find relevant information.
Here, I want to build a new quote based on information that I know from a sales quote from last year. This is where I can invoke Syntex’s content assembly to be able to use structured templates that I merge with my own data to be able to dynamically build a new sales quote using precise inputs to it. And in a few seconds I have my due document. I can open it up and be able to use Syntex to find the key places inside the document I may want to review. And again, as I showed you before, when I’m ready to send this on to get someone to accept it, I can invoke Syntex’s e-signature from the convenience of the Copilot interface. So we really see Copilot as being yet another set of ways that you can tap into the power of Syntex here.
Stephen Rose
Got it. And we’re doing all this in SharePoint but through Teams as part of that. So you have the application and the Microsoft 365 experience with those sort of services lying underneath it, and then Copilot just adding that kind of a new interface or a better, smarter way to access it, would that be correct?
Chris McNulty
That is… only one second here. It’s doing something we don’t like to do, but we’re going to do. Let’s see if I can just drag this up here.
Stephen Rose
Did you want a demo or something else?
Chris McNulty
I’ve got to share something else.
Stephen Rose
Okay, that’s fine.
Chris McNulty
It is between you and me. It’s something that hasn’t moved beyond MS IT yet. It isn’t even in our demo environment. But I can show it to you.
Stephen Rose
I think your package arrived.
Chris McNulty
Well, okay. So let me just show you one more example of how this works interactively. So here I’m inside of a SharePoint library, and I can invoke Copilot directly on a file inside of this. So, in the Copilot preview, we’re able to generate an automatic summary automatically of what’s there. But as well as get that summary, I can see propose questions that Copilot already knows how to answer, or I can add my own.
(Chris types) “What are the risks in this agreement?”
And the Syntex plug-in will invoke those and show me things that it knows just from being able to read the document and to apply advanced AI to it.
(Chris types) « Are there any personal details in this document? »
And it’ll think about it. And yes, it’s showing some of the examples, and I can see…
Stephen Rose
Some data, so you may want to lock this down or change that. Got it.
Chris McNulty
And I can see the prebuilt questions, and it shows me the next two questions. And so I can see this is a flat fee contract for $4,000 without having to ever open the document up. And so that’s another way that we’re bringing the power of Syntex into the Copilot experience. This is a capability that we are just starting to roll out in limited preview for current customers, and there’ll be more availability coming throughout the rest of 2023.
Stephen Rose
Okay, so how does this all fit in then with Project Archimedes, which is something else that was mentioned and I couldn’t find a lot of data on it. So, I think it’d be great if you could sort of fill us in here.
Chris McNulty
Yeah, Archimedes is really interesting. So if you think about how the Graph works… so if you have a billion documents in M365 and a growing number of our customers have, you know, information at that scale. If you need to find something out about a billion documents, you can write a Graph query and you can set it to iterate through, you know, every container that lives in M365 and build it into a record set, and it will come back in a few months time and you will have your billion docs record set with the right metadata. If you’re looking to find how is this shared, is information over shared, Archimedes is leveraging what we’ve previously described as Microsoft Graph Data Connect services to be able to interrogate and pull out these very large data sets about your M365 content to be able to bring it into your reporting tools, to analysis tools, to experiences like Azure Synapse, as well as being able to write your own applications against it.
So it’s a powerful way to be able to extend information about your content as you may need Whether you are a data analyst, whether you’re working in a line of business situation or developer, ot is a much faster way of getting to that information than writing a traditional Graph query and waiting for all throttles and get your billion items back.
Stephen Rose
Got it. Now, this is leveraging the new Microsoft Fabric. So let’s explain what that is, because that was another announcement made at Build and there seems to be a lot of excitement about it, but I think Copilot tends to overshadow things. So take a moment to talk about Microsoft Fabric.
Chris McNulty
You’re asking the wrong person.
Stephen Rose
Okay, that’s fine. So I can talk about Fabric. So, all right, so let’s go back and we’ll do that…
All right. So a lot of this now is based on Microsoft Fabric, which is this new unified analytics platform so that you can manage Microsoft 365 alongside your other data sources, but do that all in one place. Fabric will bring the data that you’re getting from Syntex and be able to bring in other types of data and into other apps and allow them all to sort of work together. But that’s really sort of the key of how do we get that together. But Archimedes is also about, you know, governance, and hey, are we sharing sensitive information with external users or how many sites have over a terabyte of storage and what’s in it? And that’s part of this as well, correct?
Chris McNulty
Right. So when you think about something like let’s talk about some of that management tool in SAM, otherwise known as SharePoint Advanced Management. So with SAM for Syntex, I can deploy very fine grained access control policies, do things like prevent download, to prevent external owners from further sharing content or from restricting permissions in the first place to prevent sharing at all.
But if you’ve got a large information architecture, how are you going to find, you know, across hundreds of thousands of sites and teams and groups, where do I need to focus my attention? And so, Archimedes comes in by being able to help provide that interrogation layer, to find out where’s my greatest level of exposure, the places that will most benefit from those kinds of policies.
It also extends to information like archiving. Today, we have a growing number of customers who are using M365 to the fullest, kind of beyond the limits of what their storage entitlement is. And so some companies are managing that by, frankly, deleting content or moving it into unintelligent repositories, whether it’s…
Stephen Rose
Like Azure Cold storage or things like that.
Chris McNulty
Right. And so when you do that, you strip out the versioning, the security, the metadata, all the richness, all the collaboration and capabilities. What we’re doing with archiving allows you to keep that information exactly in place and just move it to a lower cost storage tier so it doesn’t go into hot storage anymore. It’s still searchable by admins. It’s still subject to all the same security and compliance rules. If it’s under a retention policy and it reaches the end of its lifespan, it will get purged or whatever needs to happen. So in all cases it’s still live. It just is cut off from user interaction and it’s removed from hot user storage. And again helping find that information, our communities can be a valuable tool to help find large sections of your information architecture where you may be paying for storage at a rate you don’t need to keep paying.
Stephen Rose
Sure, yeah. No, I think that that makes good sense. All right. So let’s go through a little bit of this again. So, requirements : First of all, if you have SharePoint, you have Syntex and it’s that SharePoint on-prem or SharePoint Online, or…
Chris McNulty
Online, it’s not the legacy on-premises product.
Stephen Rose
Got it. And then you’re just using Azure usage for the processing of that. Now what about Syntex Copilot? When are people going to be able to get their hands on that?
Chris McNulty
So, the Syntex plug-in for Copilot is kind of tied to the mainline Copilot experiences. So we are rolling those out to customers right now. We expect to be able to broaden that preview availability later this year, and our fingers crossed should be a general availability for Copilot before the end of the calendar year.
We’re taking a pretty cautious approach on this. If you look at it kind of from an IP architecture perspective, all of this AI runs on GPUs. GPUs, as you may know, are fairly expensive and they’re not easily burstable. And so what we don’t want is for users first experience with Copilot to be something where they log in, they get all excited and they get told there’s no capacity, they should try again later.
Stephen Rose
Right.
Chris McNulty
So we’re carefully expanding this while we are building in the compute capacity so that people will have a good first run experience the first time they put their hands on Copilot. We haven’t even deployed Copilot to every person inside Microsoft, as far as I know… although I have it, so I’m happy.
Stephen Rose
Sure. Now, we’re also looking at a public preview program, which I know you only have like 200 users and your goal is to increase that and have it in maybe a public preview by Ignite time frame. Is is that still kind of the case? Is that what you’re thinking?
Chris McNulty
We’re shooting for that. But, I mean, you know, stay tuned to us, that’s something we’re going to definitely be continuing to refine throughout the summer as we calibrate the feedback we have from our first wave of 200 preview customers.
Stephen Rose
Got it. Final thoughts on Syntex for folks who are watching and where should they go to learn more?
Chris McNulty
So one of the best places to go to learn more is if you just go to aka.ms/syntex, our website will direct you to our Adoption Resource Center, which has tons of demos and things to read about as well as to our blog, which tends to be where we make all of our announcements. We made some announcements over the past couple of months about new services like OCR and image tagging that are coming into general availability here this month. Most
of those interactives services should be at GA by July. We’ll have more on what we’re doing with things like back up and archive next month, in July as well.
Stephen Rose
I went to the site and it was really interesting because there are prebuilt models there, which I love. There’s one for contract processing, there’s one for invoice, even for receipts. So that somebody could literally on their phone scan a receipt or take a picture of a receipt, drop it into SharePoint, it will grab the information from that person and drop it into an expense report, things like that, which is great. That’s something you can’t really do on a Power app and this is going to move that forward.
Chris McNulty
You’re touching on something that’s really important for people to think about with Syntex. There’s two design philosophies that we have. The first is there’s all of those powerful capabilities and you’ve mentioned just three of them. So Azure Cognitive Services have these pre-built models. There’s other services from Power and from Form Recognizer. One of the things we’re doing with Syntex is taking these things and wiring them directly into the M365 experience so that you don’t need a developer to go access them, right? What makes these capabilities things that anyone in the no-code environment can use? And as the Azure team adds models for things like 1099, W2, ID cards, tax forms, all of those prebuilt model types will surface inside Syntex.
The other thing is to maximize the value of your M365 content by keeping it in M365. So rather than take a document that’s ready to be signed and it goes out through email or some other cloud where you lose control of it, the best copy of it now lives somewhere else. Keep that in m365, keep your version and keep your archives. You get the most value from your content from an experience like Copilot If there’s more content for it to work with. And so making sure that we have the right management, the right economics, as well as the right integrations to let you do that is another one of our core design philosophies.
Stephen Rose
I love it and I can see a lot of value both for a small company with, you know, implementing some simple processes that help to make it easier and not require to hire more people to do very basic things to a large company that’s just managing mountains of data and trying to manage it. This is awesome. Chris, I appreciate you taking time out. I know how busy things are right now. I’m assuming we’ll be sharing more at Ignite on Syntex and we should watch for that, you’ve given us the blog, which is great. So thanks so much for spending time with us today.
Chris McNulty
Take care. Nice to meet you and your audience.
Stephen Rose
Absolutely. Well, meet my new audience!
Chris McNulty
Yeah, we’ve met before.
Stephen Rose
Yes, we have. Many times. All right everybody, for UnplugIT I’m Stephen Rose. Thanks so much for joining us. We’ll see you on our next episode.
Chris McNulty
Thanks. Bye bye.
Stephen Rose
Bye bye.