The new Microsoft Copilot: what MSPs need to know

Welcome everyone. This is the first of a series of webinars that we're doing with Microsoft, particularly around Copilot and with the broader Microsoft Copilot team. This one specifically is going to be around the new announcements that Microsoft has had over the last few weeks, with some new Copilot experiences, some new commercial updates and business model changes, and some great new ways that you can lean in to build more services and revenues as an MSP. 

I'll start by introducing myself. I'm Jason Ulacio, SVP of Business Development here at inforcer. I joined the company about a year and a quarter ago and, prior to that, I actually worked at Microsoft. So, I have actually come across a few of our guests before. With us today we have Sneha Daguvaati, who's a Principal Partner Solution Architect, and you'll hear more from her in a little bit, as well as Jason Thompson, who leads the AI Workforce SMB sales team, which is driving a lot of the go-to-market efforts around Copilot here in the Americas. 

We're going to be covering a breadth of things. First, an introduction around the new Copilot experiences and the updates there, unpacking what the commercial updates are and what usage-based billing looks like, and what all of that means for you. Then a couple of demos and a product walkthrough of what that looks like, as well as a deep dive into the admin experiences that Microsoft has built for you to help you manage and track usage, ongoing expenses and so on around Copilot. 

We have the chat open, as you all can see. So, as we go through the call today, feel free to drop questions in the chat. We have a few people from the inforcer team who will help answer any inforcer-related questions. We will also, as this is part of a series, be taking in all of your questions around Copilot and the Microsoft partnership in general. We'll track all of those and make sure they become part of our follow-up webinar sessions. So, feel free to add anything you have questions on, anything you're interested in us covering, and we will make sure to get back to those, if we can, during this call itself or in our follow-up sessions. 

Before we dive into the agenda itself, I do know from our registrations and what the team has shared with us that about a third of you who are on the call today are completely new to inforcer. So, I want to share a bit of background on inforcer as well and very quickly cover a bit of what we do. Then I will hand it over to our Microsoft partners here, who will be sharing more of what you see here today in the agenda. 

First, to set the stage a little bit around the conversation we're having today, we're seeing the MSP market go through a bit of a transition. This is not something new. You all have been here before. If we go back two decades ago, we had this big shift that took a while, and some would argue is still happening, from break-fix to managed services. 

Then, with the rise of ransomware and threat attacks, we saw a transition into not only managed services but a big focus around managed security services. Then came COVID and a push to the cloud with remote work and so on. During that time, we saw real standardization of the market around Microsoft and Microsoft 365 services, with many of you platforming or leading with Business Premium and building a lot of your services on products like Entra and Intune. 

Following that, in the times we are in now, we have the rise of AI and this rush around building AI services and delivering AI services. This is a mixture of both getting your environments, or your customers' environments, ready and protected to be able to onboard and use all of this AI safely, as well as transforming your own businesses to become AI-forward, AI-focused MSPs and, as some might coin it in the industry, a managed intelligence provider. 

When it comes to inforcer specifically, if you haven't heard of us before, inforcer was founded in 2022 by a group of people in the MSP channel. It started with an MSP owner and a CTO within an MSP, and a couple of other people who had been in the MSP channel for a while, realizing or seeing a trend where many of themselves and many of their colleagues in the space were having a hard time managing and scaling their Microsoft businesses and practices. 

They went on the path of building multi-tenant capabilities to help deploy and manage Microsoft at scale. We launched the first version of the product, the first MVP, in late 2023. We have since onboarded over 1,500 MSP partners, a good amount of you on the call today. So, thank you so much for your trust, service and business. 

In that time, we've also onboarded over 100,000 SMB customers and businesses that those MSPs are now supporting with inforcer. Those businesses have over seven million users or endpoints under their management. As I was sharing, our whole mission as a company is to help you as MSPs go on this Microsoft journey and take your customers on this Microsoft maturity journey, building more profitable services and businesses on top of Microsoft. 

We're a global company, founded and headquartered out of the UK, as many of you already know. We've got offices in the US in LA and Tampa. We have several locations across EMEA and also in Melbourne, where we cover the APAC region. 

One last slide here on inforcer and what we do. We're a unified platform with broad coverage of the Microsoft 365 stack, and we straddle both what we would call left of boom and right of boom, as you see here on the slide. 

We help MSPs with everything from automating the deployment, configuration and ongoing management of Microsoft products, policies and features across their estate. We help MSPs get visibility into their tenants at any point in time and see where those tenants stand, any drift that might have happened and so on through our assessments. 

Over time, we've also built in a lot of the administration tasks that MSPs have to do and helped bring that into the platform in a multi-tenant fashion, whether that's user management, group management, application management, device management and so on. 

Importantly for this call, as our MSPs have moved into AI services and the Copilot world, we've also started developing some really neat features to help you deliver Copilot services, Copilot readiness assessments, manage shadow AI and more. More recently, in this quarter, we'll also be shipping a SharePoint management solution that is geared towards helping you clean up those environments and deliver secure AI services packaged with data governance to your customers. You'll hear more from us on that in this quarter. 

The other thing that we launched this summer was our right-of-boom solution, or Threat Detection and Response solution. This helps you catch any threats whenever they do get through into your customers' environments and automatically contain and remediate those with support from our human SOC, as well as several agentic flows that we have built. 

As threats and the threat landscape continuously evolve and become more agentic-driven, we see a lot of those attacks now being led by AI and therefore having to be defended with AI. I'm going to transition now to our friends at Microsoft, but I just wanted to give you all a bit of an overview of what we're offering here at inforcer and how we can help you from our end. Jason, are you ready to take it over? I'm going to pass it to another Jason for you all. 

Awesome. Good morning, good evening everyone, and welcome. First off, thank you to Jason and the inforcer team for inviting us to this community call. There's been a lot of Copilot news recently, so we thought we'd take the time to make it practical. We're going to cover what's changing, how usage-based billing works and, more importantly, what it means for your MSP business and your SMB customers. 

I'm Jason Thompson, Sales Director for Microsoft's SMB customer segment. I'm joined by Sneha Daguvaati. She is our Partner Solution Architect. Just a quick recap of the agenda that Jason highlighted: we're going to walk through the latest Copilot announcements, I'll explain what's changed, we're going to look at the evolving commercial model that includes usage-based billing and what it means for partners and customers, and then I'll hand it over to Sneha and we'll bring it to life with some live demos. 

Over the next 30 to 40 minutes, we want to share both the technology and the business opportunity that's evolving around Copilot. Let's start with the product announcements. It's good to get grounded in the four ways that customers are putting AI to work today. 

First, there's assist: help me do a task. Think of a customer that's drafting a basic email or getting an answer on a quick topic. Next, there's delegate. Think of multi-step tasks that need information from several apps or sources. Then there's build. This is where customers want help to create or code new applications. Lastly, there's automate. This is your digital assistant that you set up to do work on your behalf. 

For you as a partner, the starting question here is: what do your customers need help with across these four areas? Are they more on the left side? Are they more on the right side? We wanted to get into how Copilot brings all of these together. 

How does Copilot support these four ways of working? For assist, that's where Copilot Chat sits. It helps customers with everyday tasks right inside their Office apps. For delegate, there's Co-work. This is where you give it a goal and it works through various steps across apps, skills and connected data. A good example is preparing for a customer briefing using various emails and files. 

Then there's build. That's where Code sits. Customers can describe the application they want in plain language, even if coding is not their day job. Finally, we have automate. That's where Autopilot sits. Previously called Scout, if you're familiar with that product, it's your digital agent or teammate that you can personalize and have work on your behalf. 

Your customers will have these needs, and we're now bringing these capabilities together in one place with the latest Copilot announcements. 

Let's start with the new Home button that you're going to see in a moment. If customers don't know what mode they should use, this makes it really easy as the starting point for most work. Customers don't have to worry about, 'Am I choosing the right model? I hear Sonnet is good or I hear Opus is good. Which one should I be using?' They can just focus on what they want to get done rather than choosing between various models. 

Within that Home button, you've got Copilot Chat. Customers will see Chat. This is their everyday assistant that will likely cover 80% of their end-user needs. It helps customers work securely with context through Work IQ. Customers can choose a model or use the auto feature to balance accuracy, speed and cost. This really helps both customers and partners simplify where to go for everyday tasks. 

On the Co-work side, this is also now going to sit in the Home button that you'll see in a moment. You give it a goal instead of a prompt. It breaks down that goal into steps. It picks the right skills, the right tools and the right model to complete the task. Just like Chat, Co-work has Work IQ to keep the task grounded in all of your email, files and data that are relevant for the user. This is where customers can also create skills to help Co-work follow a more consistent process. 

Code, if you're not technical, customers can describe the app they need in plain language even if they don't write code. The apps can run locally or in the cloud, and they can be published as a managed app that IT can govern. 

Last one here is Autopilot. Previously called Scout, that was a separate desktop application. Now we're bringing it into the Copilot app directly. This is where you see some extremely advanced capabilities within Copilot. This is your personal digital agent. It's your digital teammate that you can have take on work on your behalf even when you're not working. 

Depending on the controls you set up, it has persistent memory. It starts to learn how you work. It understands the priorities that are important to you. You can personalize it, name it, give it an identity and give it specific skills that you want it to run. It actually has its own ID and mailbox. You can tag it in a Word document that you're working on because you want its input on a certain topic, or you can bring it into an email thread and ask it to go run a certain task. You're really speaking to the agent just like a real teammate. 

This is how your customers are working. This is why we're evolving the product. This is why we want customers to be able to match their needs to the right capability using Copilot. 

For all partners, their customers are asking about AI. We know a lot of customers are using shadow AI and free tools already. One of the harder parts for a partner can be having customers take that first step to get started with Copilot. 

We wanted to share today our Copilot trials, which we're calling Copilot in 30. It's available now as a 25-user, 30-day Copilot Business trial through CSP. It's focused on first-time Copilot customers. We want to help customers try premium AI and then build a business case before buying. 

On the partner side, we want to help you create pipeline. We want you to prove the value with Copilot and turn successful trials into paid deployments. This is a partner-led motion. We'll provide the data on customer targeting, the success planner for relevant use cases for your customer scenarios, and short micro-skilling videos that sit on the Partner Skilling Hub that you can use. 

We really want this to be more than just a free trial. We want to help partners build a repeatable path of moving customer interest to paid seats and then expansion. 

Shifting gears a little bit, the commercial model update is really important for partners to take note of. In this section, we'll cover how partners pay for Copilot moving forward and how partners can monetize it. 

This slide illustrates one of the best pieces of advice you can give a customer, and that is that the most powerful model isn't always the right one for the task. On this chart, you can see purple shows what the model can do and blue shows what the task actually needs. At some point, a more powerful model stops adding value to the response. 

I'll say that again in a different way. For every task, there comes a point where a more capable model no longer provides a meaningfully better response. Think of drafting a simple email. You don't need advanced models like Opus or GPT, but a complex multi-step job in Co-work or Autopilot can benefit from that extra capability. 

For partners, that can be a really good customer conversation. Where does advanced AI improve the outcome, and where is everyday AI good enough? The logic behind the commercial model with Copilot is to get customers licensed on everyday AI, plus add in the new usage-based AI that we're going to go through. 

This is the industry conversation right now. It's not unique to Microsoft. User subscription licenses cover everyday AI. Usage-based billing covers advanced workloads. Everyday AI covers the large majority of what users need in their work. It's included in the user subscription license, the USL, and that's usually priced per month, per user. 

On the advanced AI side, it's work that you hand off to Co-work or agents like Autopilot. These jobs can run for hours. They have multiple steps and multiple data sources, and they're going to be running on usage-based billing using Copilot credits. Usage-based billing really complements the USL by giving options to users when they need more advanced AI. 

A useful analogy for usage-based billing versus the user subscription license is a plug-in hybrid car. The user subscription license is the battery. It covers most of the driving needs at a predictable cost. Usage-based billing is your gas tank. It provides more range when the task needs capabilities that are beyond what the license can accomplish. 

The power of this is where the auto feature is relevant. Copilot looks at the outcome that you want and packs as much license value into that task as possible, and then uses usage-based capability when that task requires it. 

This is the customer story partners can share. Copilot provides predictable everyday AI with flexible advanced AI when the workload demands more range. 

What does this mean for partners? If the industry is moving to this hybrid model, it creates a ton of new revenue opportunities and requires partners to build new capabilities. Partners will need to have deep capabilities in FinOps. That's what we're going to look at today, as customers will look to you to help them plan their consumption and optimize their spend. 

Here's some major news we wanted to reiterate, if you didn't see it, and make sure that all CSP partners are aware. Starting on December 1, all Copilot Business seats are going to have usage-based billing on by default for all new seats. 

What that means is when you, as the partner, sell a Copilot Business license, that license will automatically set up an Azure subscription with a pay-as-you-go limit of up to 4,000 Copilot credits, or $40 per user per month. 

Just to break that down again, you sell the Copilot license, it automatically comes with an Azure subscription, usage-based billing is on and it has a $40 cap. Partners do have the ability to change that limit up or down. It's also enabling a great conversation with your customers about what they want to get out of Copilot. 

The value to the partner is that the Azure subscription is already set up in the back end, and now you have this foundation for usage-based services. On the customer side, they get access to powerful products like Co-work and understand what's possible with advanced AI. 

To reiterate, you do have the option to adjust the limits as the partner. When you're meeting customers, you might decide Co-work's not a great solution for this customer at the time, and you can disable that at the tenant level. 

One of the biggest questions we get is, 'My customer doesn't know how to estimate what they should spend and what users are going to drive consumption.' We do have the Copilot Credit Estimator. It's designed to help partners have that conversation. 

It can help you identify how many new users will use Co-work and whether their expected usage will be light, medium or heavy. The estimator then provides a recommended output of Copilot credits to budget for those users. It's a really good starting point. You can also always check your own credit usage in Co-work by typing '/cost' and review your spend over time, and that's live today. 

To wrap up here, from a partner lens, the opportunity around Copilot has three revenue streams now. The license revenue, think Copilot Business seats, standard margin and incentives, grows as the user base grows. 

On the services revenue side, that's AI strategy, security, compliance, adoption and governing that spend. Now you have the usage-based revenue that includes Co-work, Work IQ APIs and the frontier products such as Autopilot and Code that we covered. There's a lot of new opportunity to build services and capability around usage. 

We've covered what's changing in Copilot and what it means for your business as a partner. Sneha is going to walk through the new Copilot Home experience that we covered, along with Autopilot as your digital teammate, and then she's going to wrap up with the FinOps tools that help manage consumption. 

Thank you, Jason. Good morning, good afternoon. I'm Sneha Daguvaati, Principal Partner Solution Architect at Microsoft. I'm going to walk you through some of the latest innovations around Copilot that have landed recently that Jason just talked about. 

I'll start with the Home tab. As you can see, Copilot has a new Home here. We are calling it this super app. Copilot is now termed as a super app. It is the new OS for work, which is going to be that AI suite for you going forward, and we'll keep adding more capabilities. 

Today, we are going to be talking about Home, Code, Autopilot and Co-work. Starting with the Home tab here, we see Chat and Co-work are brought together in one unified experience under Home. 

Chat is always our go-to when we have something for an instant or conversational type of interaction. Co-work is something you use when you want to delegate a task. If you give it a task, Co-work runs it end to end and delivers the result. That's where you go for Co-work. 

We also introduced Code capabilities inside Copilot to bring solution building closer to business users. You don't have to understand coding. You just have to understand what you want as an outcome and describe it in natural language. Copilot does all of the figuring out and planning, and then builds out that solution for you. That solution could be a dashboard, an automation, a tracker or a workflow. Copilot figures out all of it and then delivers you the outcome. 

We also have Autopilot, which I don't have in my tenant, which I'll show you in a separate one live. 

Let's run through a quick scenario here. Let's say I'm a small business owner and I want to launch a new coffee shop. I need a thinking partner to brainstorm with me on some ideas about how I launch this within Austin. I go to Copilot, I give it a prompt, and it thinks through it. 

Here we have the response. Copilot comes up with three launch ideas for me on how I can go ahead and build my business there. What is the target audience and what's the value proposition for each idea? Who are those potential customers? In addition to that, it also gives me how I can differentiate from Starbucks. It also gives me some nice launch slogans. 

These ideas are great, but I want to make sure that I can get them executed within my budget. So, I ask it, 'What is the idea that I can run through with my $15,000 budget?' Then it gives me a complete breakdown of how to use my investment and how to break it down. 

If you observe what I have been doing here, Copilot is helping me think, but I am driving the conversation. I am asking it questions. I'm guiding it on where to go next. This is good for brainstorming, but now I have a final idea. I zeroed in on Northstar Coffee and using it as a place for others to gather and work together. 

Now I need to go to the planning. I have an idea. I need to plan for the launch. This planning involves looking through multiple sources, researching the customer base and what the scene is in Austin, how I could position it, how I could market it and a lot more. 

I move to Co-work. Co-work would be my execution partner, where I can delegate the task and then rely on it to get it to completion. That's where I give Co-work a prompt and then it keeps running. 

I also wanted to make sure that, while Co-work is running, you do not have to wait for it to complete to bring in additional asks. I forgot to mention that it should also build me some marketing one-pagers or campaign handouts for my marketing team as well, so I go ahead and ask it to also build some. 

It queues up the request and, as you see here, it is planning step by step based on the instructions that I gave Co-work in the initial prompt. It also queues up this ask that I've shared while it's executing. 

You don't have to wait for it to complete. You can keep adding asks to it so that it takes all of that into consideration, realigns its plan and moves it towards completion. 

For the interest of time, I'll move over to a similar execution that I've run earlier. As you can see, Co-work ran through the entire prompt. It gave me a nice and complete launch workbook, which detailed how my week-by-week work or spend phasing should look and what KPIs I should target while executing my launch. 

It gives me a nice breakdown and, in addition to that, it also gave me a document or a launch plan that I can share with my team to ideate and iterate over. 

We also launched Office in Copilot capabilities, which will help us edit these documents or any Microsoft 365 Excel, Word or PowerPoint files that Co-work has generated. 

I can ask one of my co-workers to review this, @mention them, and add somebody to come and review this plan, share ideas or ask them for iteration. This would be similar to how you collaborate on an existing Word document, Excel file or PowerPoint that exists in your OneDrive. 

You can directly collaborate with your colleagues, as well as with Copilot, in the same experience. That is the Office in Copilot capability that we also announced recently. This is a big update and a key differentiator for the product. No other product in the market has this capability here. 

Let's say I delegated this task to Co-work and then I stepped away. While driving to pick up my son, I remembered that I needed to ask one more thing of this task. I open up Co-work on my mobile and I can simply type the ask to it. It executes it while I come back home, and then I can get back to it. 

We have the same experience for Co-work available in your mobile apps as well. Wherever you go, you can continue to work on it and have the tasks execute. 

Now I'll move on to Code. Copilot helped me think. Co-work gave me a launch plan. Now I have to execute the launch, but I don't want to manually manage this launch campaign every week. 

So, I go to Code. I ask it to simply build me a launch planner. As you can see, I have not given it any specific language requirements or any architecture or design requirements. I just gave it a natural-language prompt, my budget, my scenario, what I want to build and what I want the outcome to look like. 

I gave this prompt and it will run. For the interest of time, I will open the one that's already running. Copilot does all of the reasoning and planning. It thinks through the scenario and how it should break it down, and then it builds me this app. 

It's not just a shell app. It also asks me if I want to persist the data or if it's simply short-lived data that should live in the app itself. I said I should save the data about how I'm spending my budget. So, I asked it to persist the data in SharePoint. 

It will also create your SharePoint site in the back end, or if you have an existing one that you want to connect to, you can also connect your existing SharePoint sites. 

Now, as you can see, there is an output. I have this app, the Launch Command Center, that Code has built for me. I'll open it on the big screen here. It has done all of this in a matter of minutes. 

As you think back to what I have done here, I have been using Chat, Co-work and Code, but I made the call as a user to switch between these modes. 

What we are adding towards Ignite or after Ignite is that, when you give a prompt to Copilot, it will automatically recommend the right mode and automatically route you to the right mode. Right now, as a user, I made a call that for interaction I go to Chat, for Co-work I'm delegating and for Code I want to build something. 

After Ignite, you will see capabilities like smart routing that will automatically switch to the mode that is right for the task at hand. Copilot will automatically make that switch for you. The user doesn't have to understand what the nuances of each of these apps are or when to use what. 

This is a major question that a lot of partners and customers come to us with. After Ignite, when those smart routing capabilities land, you'll no longer have to educate users on that because Copilot does that heavy lifting and all that thinking for you already. 

Now we have the app. I want to share and distribute it to my team so that we can all collaborate and update our activities for the launch campaign in the app. 

I click on publish in the app, and then it gets published. Once it gets published, it gets hosted on something called Microsoft Copilot Managed Runtime. This is the hosting infrastructure that we also announced during the September 25 announcements, along with all of these capabilities. 

This hosting infrastructure will give your apps identity, security and governance so that all of your information and the app would be within the Microsoft 365 trust boundary. 

Using Copilot Managed Runtime, you can host any of the apps that you build, not necessarily only those that you build using Co-work, not necessarily those that you build using Code in Copilot, and not necessarily only those that you build using Copilot Studio. 

We have an SDK available for you that you can use to bring apps from third-party services like Lovable or any others onto this and host them on Copilot Managed Runtime. 

Not just that, because it's backed by GitHub Copilot, you can first build the app using low-code or no-code experiences like Code or Co-work, and then you can hand it off to your developer. They can open it up in their familiar GitHub Copilot experience or through the CLI. They can version-control it, add new capabilities and edit the app from their familiar interface as well. 

That is really valuable for teams that have different personas and different requirements. 

Now I'll switch over to Autopilot to talk about that. I do not have it in my tenant, so I opened it in a different environment here. As you can see, Home, Code and Autopilot show up as separate tabs inside the Copilot application. 

The first time you launch Autopilot, it will appear like this because Autopilot is your digital teammate. You can define its appearance to start with. I go ahead and click on Create Autopilot. I can define its shape, face, color and expression, and then I create an avatar for it and name it. 

I chose purple here, I chose a face and then I named it Evo. As you can see on the right, it is working for me. Let's say I'm Kevin in this scenario. It sits in my org chart and, like Jason mentioned earlier, it has its own identity, including its own mailbox, storage and access to tools. It also reports to me in the org chart. 

Once Autopilot is launched here, it is something that we call a personal, proactive and persistent Autopilot agent. That means it doesn't wait for your prompt. It acts on its own. 

For it to act on its own, you need to define certain routines: what it needs to do, what it needs to learn from you and what it needs to help you with. That's where you define some routines for it. 

Here are some things that it comes with out of the box, and we'll have more additional scenarios that you can explore after Ignite once it officially lands in your Copilot environments. 

To start with today, you can use any of these things. I set up a routine for it to go through my email drafts and help me respond to them. Do not send them, but create draft emails in my inbox. That's what I asked it to do. 

I set up that routine and it runs that task every day at 4 p.m. That is the routine for it. 

In case you want to build a routine with Autopilot that interacts across multiple tools, that's where these plugins come in. Plugins are a combination or collection of skills and connectors that you can build, wrap into one and then upload as a plugin to Autopilot. 

That will help your Autopilot extend beyond the Microsoft 365 boundary. You can connect to external systems and get tasks done in those systems as well, but everything operates within the scope of the identity that Autopilot has. 

Though it reports to you, you can further define the scope of permissions that the specific Autopilot identity can have. It can access everything that you can do if you allow it, but you can also limit what it can do. You have that level of control. You and your organization will be able to govern the identity and scope of access that the Autopilot identity has. 

Let's bring Autopilot into the real world. I'm part of a procurement team and there is a crisis here, and we need to make decisions.

Instead of me reading through every message, I bring my assistant, Autopilot, into the chat and ask it to create a decision brief based on all the interactions. It reads through the messages that my colleagues have sent and all of the email conversations and contacts that I have in my Microsoft 365 ecosystem.

I asked it to create a decision brief and put it in a team folder. Immediately, as you see, Evo gets into action. It responds to me that it's working on it, and it also delivers the decision brief.

Now that I have the brief, if I open it up, I see that it's created a complete decision brief. Like with the Office in Copilot capabilities, I can bring my colleagues into this document, @mention them, call them in and collaborate on editing this one.

Now that I have the brief ready, I would like to meet with one of my colleagues to go over it and finalize the decisions. I want to meet with Elia here, so I ask Evo, my Autopilot, to schedule a meeting with her.

It went and looked through our calendars, identified a time slot and asked for my confirmation to go ahead and book it. I say, 'Go ahead and book it,' and then it schedules the meeting. Here is the meeting draft.

It does everything like my personal assistant does and helps me proactively. So far, you have been seeing that I instructed it and it executes the things. But, as we have established some routines already, it keeps running continuously in the background.

As you see here, here are some routines that I have preset for my Evo Autopilot, and it keeps running them in the background. I can go and see what those routines are and what's enabled. I go to the run history here. I can see how many runs it has done and all of that.

Because Autopilot reports to you in the org chart, there are colleagues who can actually reach out to your Autopilot to meet with you, schedule time with you or ask questions that your Autopilot could help answer.

One of my colleagues, Mona, has reached out and she's asking for time to meet with me, and Autopilot is coordinating that with her.

If you only want certain colleagues to reach out, or you always want to be in the loop when Autopilot is interacting with any of your colleagues, you have the flexibility to set those settings here.

You go to the settings and select who should be involved: everyone, just you, or you and others. In addition to that, you also want to limit who the Autopilot can reach out to. You can click on the settings here. In addition to defining the appearance of Evo, you can also define who in your organization can reach out to Evo. That's where you define that.

Now let's switch over to talk about usage-based billing and AI FinOps. Jason has talked about a set of capabilities inside Copilot that will require usage-based billing.

Whether it is Co-work, Autopilot, Code or Copilot Managed Runtime, all of these consume Copilot credits based on the amount of usage and number of tokens that your users consume.

In a FinOps scenario, let's say your customer has usage-based billing enabled and they want to understand: 'Now that I have usage-based billing enabled, a lot of consumption is going on. How do I control and manage it? How do I know what business value the tokens I'm paying for are actually delivering for my organization, and how can I improve upon it?'

That discipline is your partner opportunity, and AI FinOps is the discipline through which you can help your customers answer those questions.

That's where the FinOps capabilities are improving or evolving in this maturity model. As you see here, we have a set of capabilities that will allow you to manage and control the spend.

We also have a set of capabilities that will help you understand the value that it's generating, such as tokens to business value. What is the correlation there? How much value, time improvement, productivity improvement or cost savings are these tokens generating for you?

When you want to optimize, you need to understand the spend and also be able to forecast the spend and predict the patterns.

Today, I will talk about these capabilities, and you will hear more about the capabilities around optimizing the value and optimizing the spend around Ignite. You can tune into the webinar that we have in December through inforcer to learn more about them. Today, I will cover the first two pillars.

First of all, when we want to control and manage the spend, we need to understand what capabilities we have as an admin to limit that spend.

In the Microsoft 365 Admin Center, you have cost management capabilities under Copilot that will allow you to define or configure policies to manage spend, control spend and allocate tokens to your organization.

As you can see, these are all the capabilities in Copilot that will consume those credits. The second thing that you would need is visibility. That's where visibility for users and visibility for the organization about how much spend they are consuming is important.

Starting with visibility for users, as you see here inside Copilot, you go into the three dots and then click on Usage. You will see what spend is allocated to you, how much is remaining and what the spend pattern is.

End users who use Co-work will be able to understand how much they have left. Do I have to save credits for an upcoming presentation or an upcoming business review that I need to prepare for, or do I have more leeway to use it today for other asks as well?

They can plan accordingly by having visibility into the amount of credits that they have left or how much they have used already. As they near certain thresholds that your admin may have configured, such as only having 20% remaining, end users also get alerts like this while they are using the app.

If we go to the admin experience here, this is slightly different from the screen that I showed earlier, but you'll have Code and Copilot Managed Runtime on the screen when you configure cost management.

The first thing that you will need to do is set up the default spend policy for your organization. You have to go into the cost management dashboard, first select the Azure subscription that's tied to it, then create the billing policy and create the spend policies as well.

You have to decide how many credits you want to limit, what your monthly spending for your organization would be and what the per-user limit would be. This will be the broader policy for all users who have Copilot licenses that you are allocating here.

Once that is done, you can decide who should be notified when a threshold is met. You can set alerts when users spend a certain number of credits or when they reach a percentage of their allocated spend.

Once that spending policy is created, there are many scenarios where you'll have to create custom spending policies.

Let's say you have an AI Center of Excellence team within a customer organization, or a marketing team would like to use more spend this month. Maybe in general they want to use more spend because they would be generating more artifacts compared to a finance team or some other engineering team.

In that case, you will want to create a custom policy for your marketing team so that they can get more credits allocated.

These spend policies can be scoped by users, and you can also scope them by resources. Identity could be helpful when you're creating that spend policy for your marketing team.

With resources, you can also allocate or limit the spend by the environments that are tied to it. Let's say your engineering team is heavily involved in building agents and they have certain environments allocated. That's where you go and set the limits for the resources as well.

You can allocate the spend by the user scope, as well as the resource scope. That's really helpful. This is the unification. Previously, we had these capabilities in Copilot Studio, which is a different platform. We have unified all of this into one FinOps experience so that it will be easier for our customers.

After I configure the scope of the policy, I go and select which services they should be allowed to use. As we can see, usage-based billing is spread across different modes of work: Co-work, Autopilot and Code.

Of these, which would you want to allocate the credits from this spend policy for? You can select, for example, that you only want the marketing team to be using Co-work, not Code and not Copilot Managed Runtime. Then I can limit the scope of this policy to only Co-work. That's how you can scope it out as well.

I'm almost running out of time, so this is the experience that we have in terms of cost management and how you can limit and control the spend.

In terms of visibility, we have capabilities around how the cost management dashboard also exists in the same Microsoft 365 Admin Center experience, where you can understand the usage patterns. Based on these insights, you can make decisions on how to optimize the spend and how to reallocate the licenses.

This could be another managed service opportunity for you to have cost optimization reviews or FinOps review conversations with your customers, to optimize their spend and drive up their business value by limiting the users who are not using it for the right reasons.

All such decision-making could happen based on these insights. In addition to this, your leaders will also have a custom dashboard, which will launch around Ignite. We'll learn more about it later. Now I'll stop and hand it over to Jason to close us out.

Awesome. Thank you, Sneha. I know we've got two minutes, so we'll be super quick. I'll close on this: we see the partner opportunity as much bigger than just selling the license. It's really about helping customers put AI to work and building those ongoing services around the value.

Pick a customer, identify the need and start that conversation with Copilot. We've highlighted four things to help you with that journey.

First, get familiar with the Copilot in 30 trials. Review the announcements. There's a great partner FAQ that's constantly updated on Copilot.

Second, bookmark the usage-based billing partner opportunity deck and the FAQ. That will evolve as we get closer to the December 1 timeline.

Third, as Sneha went through, build that FinOps capability. There's great skilling content within the Partner Skilling Hub on usage-based billing.

Fourth, register for or follow all the announcements that will come from Microsoft Ignite, which is in mid-November.

From our side, thank you to inforcer. Thanks for the partnership and bringing this community together. Thank you to everyone who joined today.

Awesome. Thank you, Jason. A couple of last quick things here before we wrap up for the day.

As the Microsoft team shared, building AI services is a journey that we're all on, including you all as MSPs. It starts from helping get your clients ready, to driving that ongoing adoption, as well as building custom apps and agents and helping them manage and adopt those across their environment.

At inforcer, we're already starting to build experiences to help you with that. We have things like our Copilot readiness assessments that many of you are using already to identify who the top candidates are for Copilot.

We automatically suggest different actions and things that they could leverage Copilot for based on their personas and use cases. We then help you build what their security and data governance posture should look like and provide a step-by-step guide on what they should be doing in order to prepare them for that AI rollout.

We also have shadow AI detection in the platform, as most of you know. People are using AI at work whether it is sanctioned or not.

Within the platform, we give you the ability to see what AI is being used in your customers' environments that is not Copilot today, report that out to your customers and empower you to have the conversation around whether they knew that all of this was being used.

Are they comfortable with all of that data being transferred to these third-party AI apps? Would they rather you help them platform and standardize on Copilot, where you can manage that for them?

Lastly, we also help you track Copilot adoption over time. Beyond just having this be a single license, you can deliver this as an ongoing service and continue to provide enablement.

It will be more and more necessary as AI services expand beyond a simple license and chat experience to something that really evolves with these usage-based frontier models and more advanced workloads.

Lastly, one thing I wanted to show you, which I just think is so cool, is that we see so many of our partners already starting to lean in here.

When we first shipped our initial Copilot experiences, it was around Ignite last year. In around January 2026, we saw among our customers and inforcer MSPs a total Copilot usage of about 16,000 licenses. That has gone up to 409,000 licenses over the last three quarters, up until September 30.

We see so many of you, so many of our MSPs, already starting to take your customers on this journey and really leaning into this.

If you haven't started already, speak to your PSMs and speak to your account teams on our side. We're ready and happy to help you along this journey.

This wasn't all. As the team has shared, we'll be back as part of this series in early December, on December 3, to share more of the updates from Ignite.

Thank you so much, everyone, for joining us. Thank you to the Microsoft team for coming along, and everyone have a good day. Bye-bye. See you soon.