So, as the Generative Pages marketing push and dollars spill out into the hot little hands of the various Microsoft Consultants out there who don't work for Microsoft but become their mouthpieces. This is all so very tiring.
A coworker sent me over this today and...yeah...this is pure Microsoft Fan Fiction.
What The Consultants are Saying
So this is the diagram this guy shared on LinkedIn and then spread all over the place.
This is perhaps one of the stupidest takes on future-state for the Power Platform I've seen. While it is definitely something that the Microsoft Copilot and Power Platform team would promote, there is no way this represents anywhere close to an accurate future-state.
Where to start...
C# Plugins
I have no words.
I'm not even going to spend any further time on this.
Power Pages
Everything about the existing Power Pages bit is terrible. Of course they'll keep the terrible Dataverse back-end and will then pivot to something almost identical to Generative Pages.
In truth, it seems likelier that Power Pages (as currently defined) essentially stops existing BUT it is replaced by Power Apps w/ Generative Pages. Under the hood, Power Pages wasn't very baked and was always wonky. So streamlining the turning Generative Pages into a Power Pages deployment seems far more likely.
There is already a pathway for sharing Power Apps externally, but it causes a load on Microsoft's platform as it currently exists. Generative Pages moves this load in some way off of the existing Power Apps model and more toward the Power Pages model anyway. Therefore, it seems highly likely that Power Pages goes away and Generative Pages w/ external sharing and altered licensing becomes the replacement.
Custom Pages
I have no doubt that the Microsoft Team would like to kill off Canvas Apps and Custom Pages. So this one doesn't seem like it is impossible. However, it will take a serious pivot for Microsoft to remove this capability and I'd expect they'd lose at least 30% of their customer base barring them flipping this to some kind of Generative Page Canvas App.
The point here is all around Model Apps being based on highly structured data and for the most part Dataverse. This forces the entire design of how you build/scope your applications back toward traditional scoping and project management. It either slows down initial delivery or it heavily restricts iterative changes. Potentially both, but definitely not neither.
Canvas apps (or Custom Pages) builds data structures on the fly and are more flexible during design and post-deployment.
I honestly don't know if Microsoft (or their AI) has it in them to truly flip the script here. AI does better under strict parameters and data structures. If things are too loose...hallucinations. So I think that's almost a guarantee that AI has to be much better before this could happen. Plus, Microsoft wants us to structure our data so that queries are cheaper (for them). Unstructured or semi-structured data flies in the face of this and also sets (current) AI up to fail.
Therefore, if Canvas Apps/Custom Pages is to die, then it is unlikely they will provide a replacement. Instead, they will attempt to price it out via licensing. Leave it in place, stagnate the dev and features (as they already are), and jack the cost.
Just so you know, I think my org will change platforms when that happens. If anyone wants me to help spec out what I think a replacement would look like, then let's chat.
AI Builder
Sure. This isn't going away unless the AI bubble hits us before 2030 (pretty please). But they'll probably rename it AI SUPER-Builder (or something equally stupid).
The feature prediction here is meaningless. However, it is certain that they'll tier this w/ altered pricing eventually to try and cover the spread on their ongoing massive losses on all things AI.
Power BI
BI will probably lean in heavier to free-text and AI generation for items, but that's already there. So I'd guess they'll do a bit more customization around this to grease executives to make them feel...special. If they don't offer up some kind of AI assistant offering "Business Insights" specifically targeting MBAs, then I'd be shocked.
Definitely additional Premium licenses for this of course.
Dataverse
His prediction on Dataverse flexing more toward Cosmos would make me happier, but it is highly unlikely that becomes a thing. It would fix some of my concerns around the flexibility of the data/app for Model Apps, but a precursor to this is an AI that does not hallucinate. This is a barrier that nobody has crossed yet and remains a huge TBD for the entire industry.
As well, Microsoft is not financially incentivized to do this. Period. They want efficient and low-cost queries. They like Dataverse over SharePoint because: $. However, I also know that simple single-table structures in SharePoint are less costly (in resources and of course actual $'s) than Dataverse queries on their standard inter-table model. So changing to a semi-structured/unstructured model for the core database would jack the pricing for Microsoft up and therefore pricing higher for us as well.
I don't doubt that they could include this as a premium tier to pull in Cosmos, but if anyone's looked at pricing for Cosmos, I'd say that's a non-starter. Cosmos pricing and features are targeting global companies replicating semi/un-structured data around the world. This would be targeting global enterprises for "follow the sun" data replication. There are some companies that fit this model, but lots of competitors in that space w/ better pricing. So I'd say...no.
Therefore, Dataverse continues as-is for the most part. Some more automation on the back-end to lower the needs for specialists on the platform simply to manage it is probably a best case there.
SharePoint
Perhaps the only 100% correct item he lists. SharePoint is too embedded much like Excel in existing customers. I'd guess they will disable more back-end features while adding fluffy things for UX. I do fear for the SharePoint APIs and how long we'll have cost-free access to those.
Fabric
Fabric is too new to know WTF they'll do w/ it. Certainly it has an appeal and I'd like to see some tighter integration as we go forward, but the future state for this is a complete unknown. I would be more likely to bet that Microsoft Access will still exist after Fabric is long gone.
Power Automate
Anything that kills Power Automate will not kill it enough for me. However, it is obvious that Microsoft has abandoned trying to "fix" the UX for it and will push this all into AI generated flows. I think they'll have some level of success here, but I will note that it will still be "Power Automate", it will just be that more people will use AI to build them.
This platform is terrible, but is essential. So if this changes, I'll high-five whoever makes it better. However, if it forces me to use AI to build flows and then I cannot edit them...
Power FX
I think it is a plausible guess to say that Microsoft will (try to) kill Power FX. However, I also think this is entirely dependent on AI's capability improving dramatically and us abandoning Canvas/Custom Apps.
So long as Canvas Apps exist, Power FX will exist.
Final Thoughts
I responded to this...<ahem>...idiot...on his LinkedIn post after I saw that he's got some speed-run course for turning you into a Solutions Architect for the Power Platform. Let me just say that if you've taken this course and you think you're going to interview w/ me, then you should just close your laptop.
 |
| If you know...you know. |
His course is the standard Microsoft's "best practice" design for "how to take the most money from your customers". It is entirely a copy of Microsoft's propaganda on Dataverse, Model Apps, and AI. Bring in any Microsoft consultant or MCP and they'll spew the same stuff. Every consultant we interview for any position around this platform has to be told during the interview that there is no Dataverse. There is no AI. If you mention it, we might fire you.
NOTE: If you're one of those consultants currently working w/ us, just know that I've already recommended you be released from your contract.
In all seriousness though, trying to predict any future-state right now that doesn't include "the AI bubble crashes and most if not all companies shut down their AI systems as never-profitable-boondoggles" as a higher probability than anything else is ridiculous. This is just a mess of a market where investments have gone into pure speculation to leverage access to property/utilities with localities that otherwise would say no. I think it's more likely that the Power Platform as a whole falls back to 2022 levels of "AI" and they jack our licenses for the same old feature sets than we get a miracle AI that builds apps for us from a loose prompt from some non-technical user.
Microsoft "MVP's" are lying to themselves and/or to you by pitching this future state. It is of course the standard "this is the future of X" kind of crap that we've all been guilty of pushing at one time or another during our tech careers. However, this time the bet is massive and the odds quite low. In some regards, I think they must know it is all a lie. If this all works, then they'll be out of a job. But if it DOESN'T WORK, then they'll come back to bill you to undo whatever they sold you last year.
The Power Platform is hanging on by a thread IMHO. It is still viable so long as Microsoft licensing remains similar to today, but there will come a day in the next 3 years (before 2030 that is) where Microsoft will adjust pricing to cover their losses and people will start to abandon it.
I have already put a project on my calendar to review alternatives for our enterprise of 40,000 employees. I don't know if there's a good replacement, but I'm already looking.
No comments:
Post a Comment
Because some d-bag is throwing 'bot posts at my blog I've turned on full Moderation. If your comment doesn't show up immediately then that's why.
DIAF Visualpath team