Microsoft has often highlighted the importance that their partner ecosystem plays in the software business. For every one USD spent on MS product licenses, partners are said to make roughly 10x:

These types of big numbers are usually designed to support a narrative. It’s hard to validate that for the annual MS revenue of $282B there would be 3 trillion USD worth of business created for partners directly. Yet we can still conclude that there’s a sizable economic impact from deploying, supporting, and extending software that originates from the Redmond giant.

One could therefore assume that support for building business software products on top of the application platform from Microsoft would be a critical focus area. But is it really? While doing software consulting business related to BizApps has been fairly lucrative since the early Dynamics days, trying to build ISV products to extend and integrate with Dynamics has been a rocky road. With Power Platform and the low-code story, it seems to have only become harder.

In this week’s newsletter I’ll look at the app store ambitions of Microsoft and explore what options exist for partners that want to create products with license enforcement on top of Power Platform. Then I’ll share a bit about what I’ve been working on in terms of productizing low-code solutions recently. Finally, a few takes on what AI means to the business of apps.

AppSource is where customers are directed to go and browse for options when looking for products or partners that build on top of Microsoft’s offering. To say that AppSource hasn’t lived up to the expectations set by Microsoft’s pitch to its BizApps partners about the business opportunity would be a massive understatement. If anyone expected it to be an app store that directly brings in subscribers to software products, it’s more like a train wreck that keeps on happening, year after year.

You don’t have to take my word for it. You can check out the writings of another CRM baldie, Steve Mordue, who has been both the most vocal advocate as well as critic of AppSource ever since MS launched it. From initial optimism on the business potential (“Will AppSource Save the Day for SMB?”) to frustration on lack of execution (“AppSource – 19th time’s the Charm”), Steve’s brutally honest blog posts have documented exactly what could have been and yet never was. Including sharp observations about what has stopped Microsoft from succeeding, over and over again.

The reasons aren’t technical. Salesforce launched their AppExchange back in 2005, allegedly based on advice Marc Benioff got from Steve Jobs. Starting late should have offered Microsoft a clear template of the best bits to copy from the leader. The reasons this hasn't happened have to be primarily about MS internal organization, politics, and incentives, rather than a technical challenge.

We haven’t come a very long way since the initial January 2011 launch of Dynamics Marketplace, which then later got replaced by Microsoft AppSource. Some of the new features that were developed, such as Test Drive environments, have since then been rolled back. I struggle to remember any positive piece of news on AppSource - if we only talk about what has actually been shipped rather than vaguely promised to partners at events like Inspire.

Whenever I end up browsing content on AppSource, the experience ranges from annoying to enraging. Partners must have come to terms with this reality by now and accept that “it is what it is”. You can’t ignore being listed on the catalogs of the mothership, yet it’s all too obvious that most of Microsoft’s portals have been mainly built to support Microsoft themselves.

The above is referred to as Conway’s law. It is often expressed with the cautionary phrase “don’t ship your org chart”. But orgs just can’t help doing that.

Life would certainly be easier for Microsoft partners and customers if there was a useful app store. But there isn’t, so life finds a way. Partners use whatever marketing channels and techniques that help them capture the attention of potential customers. The community of Power Platform developers create and maintain catalogs like XrmToolBox Plugins Store or PCF Gallery. When no default listing of products and solutions exists, the value of networks and connections only increases.

Once a customer learns about a commercial solution and wants to buy it, we get to the transactional part: how to manage the licenses? There hasn’t traditionally been a great answer to this for MS business apps partners, as was evident from these 2021 ISV survey answers about how companies building Dynamics 365 and Power Platform products handle their licensing management and enforcement:

This chart was taken from the May 2021 product team blog post “Introducing a new way to manage and enforce licenses for your products”. At that moment, Microsoft did in fact bring brand new capabilities for AppSource listed apps. After almost a decade of just talking about transactions being on the roadmap, the ISV app license management feature made it real. So, despite the many challenges, there have been some concrete steps forward.

Today, when purchasing products like RapidStart apps via AppSource, the licenses from that transaction will show up in Microsoft 365 Admin Center. The management experience is just like how you’d work with native products like Microsoft 365 Copilot.

What’s the catch? This license management capability applies to “Dynamics 365 apps on Dataverse and Power Apps” offer type. More specifically the documentation says this:

What about other types of solutions, like Power Apps canvas apps? Sorry, no licensing enforcement for you. At least not from Microsoft, not as part of the services they offer to software-led partners wanting a piece of that 10x revenue.

If you’re creating products specifically for Dynamics 365 or Power Platform customers, the idea of rolling your own licensing management system might sound daunting. Recently, a partner solution for this specific partner need was launched. Ianus Guard is a cloud service that focuses specifically on creating enforceable licensing for Power Platform artifacts. As you can see from the admin portal, you could configure products that cover PCF components, plugins, canvas apps and flows:

When developing the canvas app, you could then drop in the PCF code component from Ianus Guard. You save the correct identifiers into the control properties. The PCF will render as a notification dialog on the page where you drop it, offering the app end user the ability to enter a license key. In addition, everything else inside your canvas app can check whether this PCF returns IsValid = true. If not, you can lock any features you want to protect an unlicensed user from accessing.

It looks deceptively simple. Can this really guarantee that users who paid money for the solution can always access it? Here’s the key point from the product’s documentation:

Rather than being a platform agnostic license management service, this one is specifically built for Power Platform based products. That significantly reduces the hurdles for adding a reliable licensing mechanism inside your canvas apps, flows, and other elements of your product. Offering a premium version of your app won’t theoretically require more than choosing a suitable plan for Ianus Guard, collecting money from your customers and then generating the license keys from the workspace UI.

AppSource and these 3rd-party licensing management options revolve around the idea that a solution package will get deployed to the customer’s Microsoft 365 tenant. Meaning, there’s a piece of software that is installed inside an environment that the creator of the product does not control. While that’s certainly how the usual CRM related add-ons and extensions will operate, what about if you build a product that doesn’t require such integration to existing Microsoft products?

That’s what we did when launching FinModeler. It’s a SaaS app built on Power Platform that helps startup founders, SMBs, and financial consultants turn structured business planning data into dynamic, formula-based Excel financial models. Instead of struggling with templates or error-prone spreadsheets, users enter their assumptions through a guided web interface and instantly receive a professional model as an editable workbook they can trust for forecasting, fundraising, and decision-making.

This is what the experience looks like when the user has gone through our business plan creation wizard and FinModeler has generated a financial model with key metrics based on it:

“Wait, that’s not a Power Pages site — it’s a canvas app?!?” Yes indeed. When you sign up for a trial, we’ve got Power Automate cloud flows in place to provision an Entra ID guest account plus a Dataverse user record for you. Everyone gets their isolated business unit in Dataverse and can only ever see their own business plan data in this environment. (Until we build team collaboration features in future product versions.)

As the use case for this app is primarily based on entering parameters about the target business to be modeled (revenue streams, fixed costs, personnel, growth assumptions, interest rates etc.), we don’t really need to deploy this solution anywhere else. That would be like providing a Windows desktop installer package for an app that will have a single user. No one does that today unless you build games or something heavily dependent on local PC resources.

In a way, we do still support working locally. Because the end result of our formulas and business logic applied to the business plan parameters is a full-fidelity Excel workbook. All the formulas we use for calculating the metrics are going to be present not just inside Power Platform but also inside Excel. If you pay for the Pro subscription, you’ll be able to modify the figures either in our web app or in the Excel workbook. It’s all there — because financial folks love Excel.

So, let’s recap a bit. We have a wizard style UI in a canvas app that guides you through the creation of a business plan. Then, we use Power Automate to do the magic that otherwise would require talking with human financial advisors (who aren’t necessarily cheap nor always available): the financial model configuration in Excel. Both the inputs as well as the resulting model data are stored securely in Dataverse, in your own private BU. You can iterate on the metrics like revenue sources, costs, growth projections, and then generate a new version of the financial model. The machine does all the hard work for you.

In the end, this isn’t all that different from the Microsoft suggested model of serving external parties via a Power Pages website. Both require a license, so we purchase Per App licenses for our customers instead of Power Pages authenticated users capacity. The difference is that you couldn’t achieve a nice UX with Power Pages for working with financial data without a ton of custom code. At which point one could rightly ask where the benefit of a low-code platform would then be in such an approach?

That’s why we chose the quickest path to getting a working product out and built a canvas app. It’s not the traditional route how you’d build SaaS products, but both me and Nuno are highly experienced with Microsoft’s business tools. This is the stack we know best, and we’ve made a product with it that doesn’t require any solution deployments. If you want to see what’s possible with this approach, go and sign up for a 14-day free trial (no credit card needed).

Recently, Microsoft leadership team has been busy setting their own platform on fire. The investments made to the post-app era of AI and agents seem to be the only factors that matter in terms of market cap for the US big tech. Everyone is talking about the hundreds of millions that get announced one tech giant A invests in giant B, or simply says that it will buy datacenter capacity from C. What’s not necessarily obvious from all these individual press releases is that mostly these corporations are just passing piles of money between each other:

We’re yet to see any of these companies make AI revenue that would signal there being demand for all that gigantic AI capacity being built up today. Yet just like a dollar from MS is said to turn into $10 in the partner ecosystem, the big tech corporations appear to multiply their money by just passing it around and watching their valuation climb as a result. I’m not saying it’s a bubble but… Oh, sorry, I seem to have already done so:

Anyway, all this focus from MS on building the thing that supposedly comes after apps has left the partner story kinda hollow. Sure, selling AI consulting is a big business for many parties right now. But if you need to actually sell licenses to software that does more than wave its hands, where is the revenue model in the age of Copilot hiding?

My personal perspective on this situation is that the software-led MS partners are less likely to bet the farm on GenAI than the principal vendor of the ecosystem. Most will likely include new AI capabilities in their software products (FinModeler does that, too). How many are going to throw away the concept of an app and start offering purely AI agents, though? I don’t see MS discontinuing their current Dynamics 365 products either, rather they are just bringing contextual agents inside the good ol’ web app UI.

Once there eventually is proof of money being made via products that don’t resemble traditional Power Apps in their user interaction paradigm, I’m sure partners will build new offerings for such use cases and platforms. Until then, it may be a positive thing for the end customers if not everyone in the ecosystem follows the agentic agenda that Satya and his crew have set their targets on. Because if all you get from MS is more Copilot agents, then at least MS partners can maintain an offering of business apps to meet the operational everyday needs of the customers.