Speaker 1: Hey what's going on everyone? Today we're going to talk about the intelligent speaker category for certified Microsoft Teams devices. What is an intelligent speaker? It is a device that can be placed on a meeting room table and a Microsoft Teams rooms setup, but only Microsoft Teams rooms on Windows. That's the first caveat. So we've got an MTR on Windows setup. We want a device in the middle of the table that'll be our speakers and microphone. These devices that are all certified for Teams have a seven microphone beam forming array and the device in the table will, in a room of no more than 20 people and with proper placement, it will hear the people speaking, reference data stored in Office 365, and identify these individuals in a recorded transcription, assuming transcription is enabled, so that someone who isn't in the meeting or wants to come back and reference the spoken transcript later will be able to know who said what in the meeting. Even if you've got a room full of people and you're not able to be in the meeting or see the faces of who is speaking, you can still see, because of the voice profiles that are set up, who is said what. Now this process relies on a number of things being set up and done. The IT admins in the organization need to make sure the proper settings are in place in your Microsoft Teams meeting policy. The end users themselves in the tenant need to set up a voice profile and there are a number of caveats for getting in during a meeting, correcting who was attributed to what if it's incorrect. Maybe the transcript isn't identifying a speaker and you want to place someone who is invited to the meeting as be that speaker because they didn't have a voice profile set up. You can do all that and then the users who have been identified on the fly can get notified after the meeting that they were identified as a speaker in a transcript and they can either accept that or reject that action. And then of course the transcript, the real-time transcript is available for download after the meeting. So we're unpacking a lot here but the idea is meeting equality, inclusion in the meeting room experience for all involved, people that need to go back and reference what was said in a conference room and don't have the visual frame of reference. This enables an organization to leverage voice data of individual users to identify them in a meeting as having said certain things. So there's some flexibility here, there's a lot of controls placed in the hands of the IT admins, but to set this up there's a few key steps. Now what we're not going to do in this video is actually demonstrate a meeting where an intelligent speaker is being used. However, there is a handy little link in the upper corner of the video to a separate video for the EPOS Xpand Capture 5 intelligent speaker and there you can see information about that device, you can see the device in action actually identifying myself as the speaker based off the voice profile that we'll set up in this video. So check that out for the actual demonstration. In this video let's take a look at how to set this up for your organization and manage those transcripts after the fact. Okay so first thing we've got this very handy little article here gives us all the steps about what an intelligent speaker is, how it works, how to set up your profile, how to go in and make corrections and identify a speaker that wasn't properly identified or remove an identification and then of course editing the identifications in a saved transcript and finally updating or removing your voice profile. I'm not going to cover all of these but I am going to provide this link here so that you can come in and read through this in depth. But the things we're going to take a look at is what is needed to set up the settings for your tenant and then for the individual user. Now first at the tenant level we come into our meeting policies. Here we are in the Microsoft Teams admin center. I've got my global, if you look up here we got our global meeting policy open and we come down to recording and transcription. We need to have transcription turned on because it is turned off by default. That's the whole idea here is we're getting a transcription with people being identified as speakers for certain parts of that transcription. Now if the people can't be identified they will simply be identified as speaker one, speaker two, speaker three. So that's the first part of the meeting policy. The second part is that we have to set a couple key settings via PowerShell for this same policy. It doesn't have to be global policy, you could set up a separate policy and apply that to the users that you want them to apply to, especially if you're doing this in testing. We're going to pull up our Teams PowerShell window here and first thing we've done is we've imported our Microsoft Teams module and connected to it. Once we're connected we have gone and done the set CS Teams meeting policy commandlet. We identified the policy that we want to modify here, which is the global in our case, but it can be anything in your case. Then we've got this enroll user override field. We have set that to enabled. This is going to allow the users to actually set up their voice profiles and allow them to override policies and attributions there. The next thing we want to do is the room attribute user override setting. There are a couple different things here. We can disable this and not allow any of this to work at all, or we can set it to attribute, or I should say attribute, which will then have individual people attributed to spoken portions of the transcript per their voice profiles, and there is one other setting in here which is a little more generic. It'll just say speaker one, speaker two, and speaker three. We don't want that to be the case, so we're going to leave it to attribute. We ran that. It was good to go. We said get CS Teams meeting policy just to validate our settings, and you'll see as we come down here, we already saw that we set our allow transcription to true, so it is what we need it to be. Coming on down, we see our two settings that we want to look at here. We set our enroll user override to enabled and our room attribute user override to attribute. Once we've set that, we need to go and make sure that the users actually have the policies assigned to them, so coming over to our user profile, we click on the policies, and we can see our meeting policy right here is global, so this is an easy one because I'm modifying the global policy. If I were modifying a separate policy, we're going to need to come in and apply that meeting policy to all the individuals that need to have these settings applied to them so that when they're in meetings, they can be identified and of course override settings where they need to, so we're good to go on our policy establishment there. We've got transcription enabled for the tenant, and we've set the meeting settings to what they need to be. With that setting applied to the user profile, the user can then pull up their Microsoft Teams client. Now keep in mind, it may take up to 48 hours for this to show up for them, but if they sign out of their client and sign back in, it's likely this setting will pop in for them. Okay, to record our voice profile, we want to have a microphone and a good quiet space, as little noise as possible. We want to isolate our voice the best we can, something high quality, so you know, find something that works good for you. I'm going to use this, this is actually, it was a Plantronics headset, now it would be a Poly headset, but avoid your focus, and I'm going to click on the three dots on my Microsoft Teams client and click on settings, and just a second here, once it's finished reading the policies that are associated with me, it populates this recognition tab. Now our captions and transcripts, these should be turned on. Okay, we automatically identify me in meeting captions and transcripts, yes, we want that to happen, so we leave that turned on, and once it finishes spinning, it would let us know that we can turn that on. I'm going to click on recognition, and I'm going to go to Teams voice recognition and say get started. Now you have to choose from the microphone up here, I'm going to choose my Plantronics Focus headset here, and I'm going to read this whole box once the start voice capture has been clicked. When we're done, we'll click end voice capture, and it will process our voice. Let's go ahead and give that a shot. Hey there Teams, I'll talk for a few seconds to help you get to know me. It takes more than just a sentence or two, because there are lots of details about my voice that you need to learn to be able to recognize me when I speak up in a meeting, and it's okay if I don't read this perfectly. I can screw up like I did earlier in this thing, you can let me know if I need to start over. You're learning my voice, so I'm saying plenty of things so that you can learn my voice and learn it well. Thank you Teams, hopefully we have a successful voice profile. And we click end voice capture, and it goes through the process of learning my voice. Okay, we had to stop the recording because something I think with Camtasia may have been throwing the audio input into the Teams app for this purpose. I'm not quite sure, but once we ran through it a couple different times and it prompted us to try again, it actually captured the voice, processed it, gave us the little check mark, and said we are good to go. So now we can close the window or we can say view and manage your Teams voice information. Bringing us back to this recognition tab, we can remove our voice profile if we want to. We can update it if we want to update our voice. Maybe we feel like we're not getting properly attributed in our various meetings. We want to try to have the voice captured a little more true or with a clearer microphone, we can come in here and do that. And if we want to let our teammates know about Teams voice and face recognition, we can say share link. For now this is obviously a voice recognition component. So there we go, that's setting up the voice profile. Now at this stage with the voice profile in the system and our tenant settings as they should be, I as that user should be able to join a Teams meeting which you would create from Outlook and create your meeting, invite the room to the meeting, and then I will start the meeting from my laptop, say start transcription or somebody in that meeting needs to start transcription, and then with me sitting in that room that intelligent speaker will pick up my voice and it will attribute things that I say in the meeting in the transcript to my name. Now a couple situations that may occur. Maybe we have a situation in which the identify speaker x during live transcription needs to happen. So we've set up our voice profile. Now we get to the point where the person in the meeting is identified as speaker one and we know who that person is. We can click identify speaker next to that person and the search box it'll give us all the names of the people who are invited to the meeting and we can choose from one of them. So if they weren't invited to the meeting, we're not going to be able to attribute the speech to them. If there's someone outside of the company, we can't attribute the speech to them. They'll just show up as speaker one, speaker two, you get the point. Once we have identified who we want to attribute that speech to, we say do we just attribute it to this one thing, the one sentence that is spoken, or do we attribute it to all sentences that are spoken by that person who is identified as speaker one or speaker two or speaker three. Finally, a pencil icon will show up next to a name where we have had to manually identify that person. So after that, the person that we identified is going to receive a notification in their team's activity feed with the option to reject that identification. So there is control given to the participants in the meeting. If they don't want to be involved, if they don't want to be identified, they can opt out and reject your suggestion that they were the speaker. Now if we need to correct an identification, let's say someone has been identified as a speaker, we can go and say edit speaker and then we get the same list of people, we can make the same choices, we can choose whether to update just that one sentence or all of them, and of course after the meeting the person's going to get that notification and they can then again accept or reject the attribution that you have updated or corrected in the system. Finally, we can go in and remove an identification. When you remove the identification that the intelligent speaker chose for you, then the identification is going to go back to something like speaker one, speaker two, speaker three. We can also edit identifications in a saved transcript. So we download that transcript after a meeting, any speaker edits that were made during the meeting will be listed next to the applicable entry in the downloaded transcript. So to identify an unidentified speaker or correct the misidentification or remove an identification in the transcript, just follow the previous instructions, the process is the same. So there we go, there's that. We already talked about where you can go to update or remove your voice profile that's in the Microsoft Teams client, that's how we would handle that. Now which languages are supported? Of course as this video goes on this may change. Currently we've got English in a number of locales that is supported. Maybe in the future that changes, but for now that is the supported language and locales. So there you go, that's what an intelligent speaker is, that is how you set it up as an IT admin, how you set up your voice profile as a user in the organization, some of the ways that you can manage those attributions in meeting and after meeting. So there's a lot of things you need to be considering when you're setting this up as an organization, a lot of planning that needs to be done, users need to be trained, placing the intelligent speakers is also important. You don't want them in meetings with more than 20 people, they need to be properly placed a certain distance from a wall and speakers should be a certain distance from the actual intelligent speaker itself. So all these things need to be carefully considered, carefully planned for, but if done properly the intelligent speaker can be a very powerful tool for inclusivity and meeting room equity in your organization. If you want to see this in action, check out the card that's listed in the upper corner there, you can watch my overview of the EposXpand Capture 5 device and you can see it in action capturing my voice and identifying me in the transcript per the profile that we just set up in this video. Thanks for tuning in, hope you found this helpful and I hope we'll see you back here for the next tutorial or device overview video.