ECI Pulse
At ECI, it is our mission to be the most transformative business partner you will ever engage. We thrive in a state of constant progress and pushing the boundaries of what’s possible. Over the past two decades, ECI has emerged as the premier provider of managed services and technology solutions, across cloud, digital, and cybersecurity, to the investment management industry. To date, we have helped more than 1000 global clients, from financial hedge funds and private equity entities to asset management companies, to activate their full potential through technology, a consultative approach, and a relentlessly innovative spirit.
Join us as we explore the latest trends, innovations, and strategies that are shaping the industry. Each episode features insightful conversations with industry leaders and experts who share their experiences, challenges, and visions for the future. Tune in and stay ahead of the curve with ECI.
ECI Pulse
Microsoft 365 E7 & Agent 365: The Governance-First Path to Enterprise AI
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
AI agents are multiplying fast, but who's governing them? In this episode, Julius Damato, Director of Microsoft Modern Work & AI Solutions at ECI, unpacks two new Microsoft licenses built specifically for the agentic AI era: Microsoft 365 E7 and Agent 365.
Julius breaks down what's in each SKU, why the bundled pricing makes sense as Microsoft price increases loom, and how Agent 365 treats AI agents like employees, complete with identities, audit trails, and conditional access. If your organization is starting to build agents, this is a must-listen before your next renewal cycle.
Hello, everybody. My name is Julius Damato. I'm Director of Microsoft Modern Work and AI Solutions here at ECI. And importantly, today we're going to talk about some of the new licensing that Microsoft has made available that is really meant to bring forward agents in agentic AI solutions in 2026 and kind of on the go forward in the future. Those two licenses are Microsoft 365 E7 and Agent 365. And really the way to look at these two licenses is it's really a governance first approach. to agentic AI and AI agents in general. Just to kind of paint a picture on the landscape behind this, uh there's currently projected to be 1.3 billion AI agents by the end of 2028. These are according to Microsoft. And Microsoft internally themselves have over 500,000 agents that live and breathe in their enterprise infrastructure that users and employees of Microsoft use on a day-to-day basis. They generate a little over 65,000 responses on a daily basis. And there's tens of millions of agents that were registered to the Agent 365 preview that happened over the last couple of months. And importantly, organizations today face agent sprawl with no centralized visibility into who's created them, what access they have, or really a way to enforce what agents should or shouldn't be doing. And that's really the goal of both Microsoft 365 E7 as well as Agent 365. So let's talk about what... Microsoft 365 V7 is and kind of why it exists. So it's the first licensing bundle that Microsoft's created as well as in the marketplace itself around agentic AI and treating AI agents as though they were employees. That is a key differentiator of what kind of agent 365 and the E7 skew brings. So when we look at your current licensing that most clients have today, you're buying them in multiple buckets. You probably have, you know, Microsoft 365 V5 today, or you have Office 365, E3, E5 with EMS E3, E5 plus, you know, the preview add-ons skews there. And when you start bundling it all together, it does get quite expensive, especially when you tack on Copilot and everything else as well. So that's really what E7 is kind of designed to do is consolidating your E5 licensing, Microsoft 365 Copilot premium, Entra Suite, as well as Agent 365 into a single bundle. That one generally available on May 1st. It's priced at $99 US per user per month. That is an annual commitment, annual payment. And when you look at multiple SKUs versus the single E7 SKU, it's roughly a 15 % savings increase. Important piece on that savings increase is there will be a price increase on all other Microsoft licensing besides E7 here in the coming weeks. So importantly, it's something to keep in mind, this percentage will be a higher percentage of savings if you decide to switch to over to E7. But importantly, what the license is designed to do is bring around a governance layer into a number of pieces. So when we talk about what's in the E7 SKU, as I mentioned, it's Microsoft 365 E5, the Microsoft 365 Copalite Premium license, Entra Suite, and Agent 365. And when you break down the costs of those individually, that's where you're seeing kind of the biggest cost savings. So when we talk about what Entra Suite brings you, in addition to what E5 has, if you don't know what that license is, or if you don't have it today, That is Microsoft's Zero Trust platform. It's called GSA or Global Secure Access. And then there's an internet version of it as well as a private access version of that. As the name implies, it gives you access to different endpoints. That's now included with the E7 SKU and allowing you to have a greater control over AI tooling that's being used in your environments, log in, log out, things of that nature, as well as apps, agents, and other governance controls. And then agent 365, we'll talk about a bit more in a little bit. comes with additional features around shadow AI usage, shadow agent usage, and things of that nature that build on top of the other two license SKUs that we've mentioned before. And then just to break down that math a little bit more, when you look at, again, those individual components, you're at a little over $117 or $120 for all of those. So if you have E5 with Copilot and you're not looking at doing a ZTNA-style solution or you're not looking at agent governance yet or anything like that, certainly stick with E5 plus Copilot. But once you add on any other skew on top of E5, that's really where the cost savings really starts to come into play from a month-to-month perspective. So let's talk a little bit about what agent 365 does and what's kind of included in there. So if you want to purchase it separately, it is something you can do. It is $15 a user a month. Again, went generally available on May 1st, but importantly agent 365 treats agents like they were air quote human employees. So it gives you ability into identity, a life cycle, access control, audit ability, gives agents the ability to follow conditional access. You're able to govern them just like you would any other person in the organization. And importantly, with Agent 365, there's the idea of an agent registry that can look across the board. So it's not just something that exists inside of Microsoft 365 ecosystem. So Copilot Studio, the agent builder experiences that exist in OneDrive, SharePoint, as well as Agent Builder and Copilot, Microsoft 365 Copilot. But you can also connect to the Foundry. It also connect to a number of supported vendors that exist today. So currently there's 170 different platforms that are signed up and committed to what Microsoft is building with Agent 365. So you'll see Anthropic in there soon. OpenAI will have their agent building experiences be able to be governed and controllable by Agent 365. Again, permission and governance access is really important there. But also a number of the other platforms that exist, Cursor is on board and a few of the others. So it does give you kind of access control, see what those agents are doing, what data they're touching. and really giving you those deep insights that a lot of clients that I'm working with regularly on building agents and workflows for the enterprise really want to see and it really brings that home. So why it matters now, uh importantly, everybody's on the bandwagon to start building agents, start creating dashboards, start doing all the fun GBT work, starting to look at platforms that have specialization. And importantly, when you start doing that, a lot of what we're talking about in that scenario, auto-refer to as declarative. So it's acting on behalf of, or it's being initiated as the end user. ah That's good and bad, because when you start looking at logs, it will say, Julius did X or Julius with agent, depending on the platform. perform the action. But giving agents an identity really does enhance what is possible. You can permission it to data sets. You can allow it to access content over certain connectors. You can set up a zone all off authentication. And importantly, you want to see what the agent is doing in its own thought process, depending on which model it's picking and depending on which platform. So it does establish a clear ownership, identity, lifecycle management, because you should be treating any agents or any AI workflows you're building. similar to deploying any other application in your environment where it requires regular maintenance, requires regular patching, requires regular review. That's how you should be looking at your agents. shouldn't just be at one time. We've built it. It's running great. Let's just leave it. That's not how it works, unfortunately. And important with regulated industries, like all of our clients are in, is really the governance, auditability, and kind of compliance controls around it. So maybe able to wrap that up into every single agent that you are building, whether it be centralized or... and users being able to create, that's how it functions. Agent 365 as well as E7 is really designed around licensing it per user. So you're licensing a user to be able to act the way it works as any user with Agent 365 that is going to access a agent regardless of platform should be licensed for Agent 365. So you can get that telemetry. So that's really where that goes there. And it's really designed around broader AI adoption, broader agent adoption. without creating enhanced risk that we're seeing today across the client base. And then finally here, really you should be looking at Microsoft 365v7 and Agent 365 is really the governance first for AI adoption. So it's registry first. So you want to be able to deploy and govern these agents at scale. Again, make sure that you're able to provide identities, let those agents use accounts of their own that you can audit, review and perform actions on. You really want to make sure that you're being safer with the different AI platforms and make sure that you have really centralized visibility and policy control. That is the core function of what these two SKUs bring in the marketplace. And today there are discounts that are available. If you are uh one of our CSP clients or if you have a current CSP, you can reach out to them to see what discounts are available to you that is subject to tenant availability. But you can reach out to the service teams, your account managers, your sales reps, or myself, if you're interested. And we can certainly look through what that would be. A big thing I want to highlight here is organizations should really be starting to evaluate E7s as you start to come up on the renewal cycle for copilot and or E7. you, since most of the clients are set up for annual commits, we can also upgrade E5 into E7. But if you have copilot licensing, you're kind of stuck with those. So they're made of the term, but your account managers can do that. And really a big piece I want to highlight on here is not following into the FOMO of platform chasing. That is really what Agent 365, as well as the E7 license, is really designed to bring where you can start evaluating those platforms more holistically and not kind of start adding Copilot plus ChatGBT plus Cursor plus Cloud Enterprise plus et cetera, really start to hone in on the tools that you're trying to bring. And then Microsoft's vision, as well as kind of what the market is starting to follow, is the vision should be human led. agent operated work that you want to have governed. You want to have reporting access reviews on that since more and more workload, especially in the agentic AI era that was kicked off this year, is letting agents and AI tools perform actions. So you want to be able to report on that, make sure that the action is human uh led slash has human visibility and not just accepting default, go delete these 200 records for me because that's the best decision. So that is a cool part of the governance is being able to put wrappers around controls on if an agent is deciding that I need to delete two terabytes of data before I can commit that action or if I can complete the request that you have something to stop it from doing that. And that's a core purpose of this function. That being said, that's end of my time here. If you have any questions, again, reach out to myself or your account management team. uh Thanks for joining.