Keeping up with all the product updates and announcements in Power Platform is hard. Arguably one of the hardest challenges an individual must navigate alone. When I wrote about my tips for how to keep up, it became quite a popular post:

When it comes to tracking the official release plans for products, the native experience provided by Microsoft isn’t always the most helpful one. That’s why I created the releaseplans.net site to visualize the RSS feed data from the MS Release Planner in a Power BI report:

One perspective that’s worth looking at is where do the feature ideas come from - and what happens to them in the process. That’s what I’ll be discussing today, using one example that illustrates the big challenge: how to connect user submitted ideas and MS developed features together?

The story of one feature requested and delivered

Many community members care deeply about the products they work with and willingly invest time into thinking about improvements. Some even go through the effort of writing about their ideas.

One day, I saw one such LinkedIn post from a Power Automate advocate who often shares his insightful takes on real-life issues and opportunities. It was about the need to enable and set a limitation for Dataverse storage consumption on an environment level. Indeed, as the post stated, “this would easily be a dream come true for every single PP/tenant admin out there”.

Believe it or not, sometimes Power Platform dreams do come true! A few months ago, this “manage capacity” flyout menu became available for Dataverse environments in PPAC:

I of course jumped right into using it. This means I’ve been receiving this kind of alerts about one environment in my own tenant that I’ve intentionally set a tight storage allocation limit for:

Sure, the email message is a bit rough. Environment display name vs GUID is used inconsistently. Rounding up the gigabytes would be a wise decision here. But still: it works! Despite being a soft limit instead of a hard capacity blocker as of now, there’s huge value for admins in Microsoft customer orgs that want to stop unintended storage consumption that eager citizen developers may cause, when using ungovernable MS tools like Power Platform Dataflows.

Where was this eagerly awaited feature then communicated upon its release? Well, that’s the neat part: it wasn’t. The community stumbled upon it by accident when the PPAC UI revealed the option four months ago. It was shared on LinkedIn, where I first heard about it and of course went to explore and further share my thoughts around it.

“Jukka, are you sure you didn’t just miss this release item?” I am pretty sure, yes. Being the pedantic low-code geek that I am, I had to check if there was any mention of storage allocation in the data available from MS release plans. Using my releaseplans.net site, I was only able to find the item “Get insights into currency and storage consumption” that talks about visibility to the numbers - not the actual allocation actions.

Two months after the environment level capacity management feature had landed, there was a blog post called “Seamless Capacity Management: PayGo for Overage” published in the Power Platform product team blog. As the title suggests, the focus here was promoting how it was now possible to link Azure subscriptions and use them to pay for capacity beyond the default tenant level entitlement. Given how this is a way to transfer more money to Microsoft, I guess it makes business sense to frame the message into this context.

It’s not supposed to go quite like this, though. The idea of having release waves and their plans published in advance is to give customers visibility into what’s coming into the product. Right? In practice, we’ve seen the role of release waves diminish quite a lot over the years. Big hype features aren’t ever added into the original plans, rather they are saved for special product marketing moments like Ignite, Build, and so on. Similarly, release items get rescheduled and cancelled all the time, meaning you don’t really know what’s arriving until it does.

The problem is that existing customers of the products would surely be interested in knowing when the features they’ve been requesting get implemented. Rather than hearing about the wonderful new opportunities unlocked by investments into AI technology and such. The latter is not insignificant in the long term - I’m not claiming that. It’s just that customers care about what is real today and addresses an existing need. Couldn’t you tell those things first, only then move to the envisioning a brighter future part. Pretty please, Microsoft?

The missing feedback loop

Long before the age of Copilot arrived, we entered the age of overwhelming feedback collection through digital channels. Once companies learned that it costs them hardly anything to present feedback dialogs, rating buttons and survey invitations to customers, they jumped at every such opportunity.

You know those “would you recommend Microsoft 365 Admin center to your friends” kind of popup dialogs that keep interrupting the users everywhere in MS apps and sites? Well of course you do - because there’s simply no escape from them.

Acting like you’re highly data driven is still trendy among corporations, which means the poor users end up drowning in feedback dialogs and requests to review their experiences with mundane products. While those NPS style ratings may not give the feeling that your ideas posted in the free text boxes would ever be heard, there are luckily other formats of digital feedback collection.

Idea portals have been a regular element of most Microsoft products for many years now. The concept of posting your feature ideas and voting on the ideas of others has remained surprisingly valid, ever since UserVoice was launched in 2008. I still remember setting up an internal idea voting site with it for the Dynamics CRM 3.0 users I was supporting around that time. Microsoft was running their own site called Connect for user ideas for a long time. Newer products in O365 adopted UserVoice, before Microsoft decided to migrate them to 1st party idea portals in 2021.

In a perfect world, the ideas that customers post to Power Platform product related feedback forums like this would have a clear linkage to what gets funded as product features developed by MS. While I do know that the PMs are indeed leveraging this data as part of their justifications on getting items into release plans, it’s not a mechanism that I see working very well. You could say the “idea” is great, the execution is lacking.

Using the example of environment level Dataverse storage allocation feature, I went to see what we the user community had posted about it on the Power Apps Ideas site. The first hurdle came from the broken/limited search features for locating ideas. You can go and test the search with any commonly used feature name, and you’ll always get exactly 50 results. So, there’s either an intentional or unintentional cap that Microsoft has placed on browsing the feature requests. (Given how the sites are based on Power Pages, I suspect it’s not an intentional limit.)

I spent a few minutes going through storage related ideas that were about this capability of controlling resource consumption on a more detailed level than just the entire tenant. Below is a table of ideas that I discovered:

All the items have a status “new”. Not a single comment from Microsoft’s side was found for these. Many have been posted 6 years ago. Users seem to have found and commented on these ideas long after they were originally posted - but what about Microsoft?

I can understand that with “big data” comes big challenges in managing it. The archive of ideas that gets inherited by new PMs that take over roles and areas previously handled by another person every few years - I bet it doesn’t feel that motivating. You can hardly blame the individuals in the product team.

If you can’t blame anyone, though, then no one owns it. No one manages it. You end up with a forum that’s primarily accessed by users with either passion, needs, experience, frustration - or all of them at once, towards your product. They are looking for a way to be heard and acknowledged. Idea forums at first sound like a great way to address this audience’s needs. But there’s a lot more than just the website needed for it to deliver positive outcomes.

What do you do when such an idea site isn’t quite working? You set up a new site, of course! While it isn’t exactly the way things have happened here, the end result is similar. Because of how Power Platform and its many software products have evolved over time, the platform level needs of users are today spread out to more than one site. In addition to Power Apps, the storage management ideas may well be posted under the Dynamics 365 Ideas site.

This is also a Power Pages based site, but the UX for idea management is different. I didn’t bother collecting a similar list of ideas from there because you probably get the point by now.

Making ideas count

When pivoting into being The Copilot Company, Microsoft had to reallocate a significant share of their product development resources into building the GenAI capabilities no one had on their product roadmap. Only the future will tell if it was a smart business move or not - but that’s not really the point of this post today.

What I am hoping to illustrate here is the impact that these decisions had on the as-is business. The existing products, their existing customer base and the needs that are not met currently - those will require attention again at some point. Because in the real world of business apps, the people who should get excited about your shiny new features are often the same folks who are dealing with the gaps that exist in the old features of your products.

I’m an outlier in the user community since I keep writing about the issue, week after week. The more dangerous segment in the community is those who simply tune out when all they see is talk about Copilots and agents. The ones who used to engage with the content and invest their time in learning about how to solve existing business problems with Power Platform. If they feel like these needs have now become irrelevant for Microsoft if it doesn’t directly drive AI service consumption and license sales - you’ll just not hear from them anymore. They’ll stop sharing their ideas.

What could be done about this then? Instead of approaching this as an either/or type of a question, a battle between existing products and new things to be sold - what if it wasn’t a zero-sum game? What if the problem of how to work with a big, growing pile of user ideas AND to engage with these passionate users could be solved with the help of Copilot?

Assuming LLMs are good at processing unstructured text, such as the ideas and comments submitted by users, maybe a Copilot could be trained to assist the PMs working at Microsoft. Allowing people to summarize and analyze the data, suggesting actions, automating routine tasks in administration, facilitating communication - it sounds like a pretty good fit for AI, right?

Now, what Microsoft should NOT try to do is replace the idea forums with a chatbot that talks 1:1 with the user. Increasingly, people are realizing that trying to remove human interaction from service processes entirely can end up also eliminating the customer relationships. As reflected by this turnaround of a famous AI advocate, Klarna:

I believe the right approach today is humans using AI tools, rather than autonomous AI agents trying to mimic humans. What this would mean in the context of product ideas and product development could be:

• Helping users better discover ideas that are similar to what they had in mind, to consolidate votes and comments.

• Providing MS employees with personalized summaries of recent new and popular ideas that are related to their area of the product.

• Offering them a quick way to acknowledge feedback from users and connect the ideas to ongoing release planning content.

• Automate status updates to users on their posted/voted ideas along the lifecycle of features development and delivery.

• Plan and design the public communication around the features as they are launched, to illustrate how signals from customers are leveraged by Microsoft in ensuring their investments align with the needs expressed by users.

The reason I feel comfortable advocating the use of AI here is that this process is not about exact numbers but messy, unstructured text. As a community member sitting outside MS, I don’t see much of a detailed, repeatable communication process in place today, so there’s not a big risk of things going wrong when facilitated by nondeterministic GenAI tools.

So, that’s my “idea for managing ideas”. I’d gladly post it on an ideas forum but… well, you get the idea.