Currently Archicad AI is just a useless gimmick

[Wlad222]

The current AI feature for selecting fills in Archicad is not practical because it requires an overly precise description every time, including specifying things like foreground color pen and giving exact pen numbers. The main advantage of AI should be its flexibility and ability to understand natural speech, but right now, if one types "select red fill," the AI fails unless the full criteria are spelled out in technical detail. The regular Find command already does this well enough without AI—there’s no point in using AI if it does not make the workflow faster, easier, or more efficient.

True AI should understand and convert everyday language into the right commands automatically. Users shouldn’t have to memorize and specify all parameter names or numeric values. This rigid workflow defeats the purpose and convenience that AI is supposed to bring, and currently, it only complicates the process instead of simplifying it.

3/10 disappointed.

[Wlad222]

I great feature would be to select all text elements below or above fills... Not capable of doing that yet...

[Wlad222]

Another example where I have to specify EXACTLY what to select... AI is supposed to be smart and help me get the right layer. Even if I said only layer general, it should figure it out...

[Barry Kelly]

---

@Wlad222 wrote:

Even if I said only layer general, it should figure it out...

---

You need to say 'select text in layer .... ', not just 'select text in annotation general'.

Otherwise it could be a storey name or an elevation name, etc.

I agree though it should not be so picky with capitalisation.

Barry.

[vistasp]

I think it was the lack of a hyphen that stumped the "AI" at first. 😂

This is early days... it should improve.

[Bricklyne Clarence]

Oh my Gosh!!!!!

Are you seriously telling me that GS released yet another new version with a half-baked and 'not-quite-ready-for-actual-use' new feature tool, ...which they just happen to be heavily using to promote this new version?

I don't believe it!!

Not for one second.

That would be totally TOTALLY unlike them to do such a thing. Or to waste....er...... I mean....to spend invaluable (...and we're told,.....'limited'....) development resources to develop a tool and feature that wasn't particularly requested that highly in the wishlist forums .....or elsewhere - and only because they wanted to get in on the whole "AI" fad and bandwagon, that everyone currently is on - all while ignoring actual highly requested and popularly voted for wishes and tool improvements in the wishlist section.

And by "ignore", ...I  mean, actively neglecting by placing the majority of them "ON HOLD"*.

Don't believe me? Pay a visit to the ("New and Improved!) Wishlist Section, and take a look for yourselves. They were never serious about that whole Wishlist spring-cleaning initiative anymore than they were about the Roadmap that now just looks like an internal company memo that has nothing to do with what users actually want improved.

(* "ON HOLD"  disturbingly feels a whole lot like the wishlist section digital equivalent of, "Thank you for your patience. Currently all our representatives are unavailable but please stay on the line (and listen to this mind-numbing earworm elevator music) and one of our human technicians will be with you shortly to assist you. (...if you have several hours to waste with your phone at your ear)")

But let's give those users another shiny sparkly jangly jingling toy (AI Assistant, Vision Pro, ) to play with and distract them from the fact that we're not actually addressing what they need addressed in the program, nor do we plan to any time soon. As long as you a categorize it as a "Beta" tool or "Experimental feature" it's allowed to be as half-baked as it needs be and unusable as all hell (but they still won't hesitate to use them to promote the new versions in their marketing, ....for lack of other actual tool improvements to use to promote it).

And if we're really really lucky, after a few versions, they'll just abandon improving those "experimental" tools altogether and move on to their next shiny jangly thing - leaving those old toys in the vast graveyard of half-baked Archicad "new tools" that were introduced with a plethora of issues and problems but never revistied to be improved or in some cases just re-witten altogether for lack of functionality.

But hey, at least we can now customize the footings and handles on our cabinets in 10 millions ways, while also being able to rotate them perpendicularly 90 degrees with a keystroke command (something that their competitor rival software have been able to do for the better part of the last 15 years or more).

Oh, and Dark mode. YAAAAAY!!!!...

YES!!!

Finally!!!! (*Fist pump*)

Oh wait.....

I have to be on a Mac for that.

Never mind.

[laembert]

No way!

[Kamil Dylo]

Besides the fact that no one asked for this functionality, and that it seems like it will be delivered in a half-baked way, it makes me wonder: is this really how we’ll be using software in the future?

I keep hearing that future software won’t require an interface anymore, and that we’ll all be operating it through AI text prompts. What do you guys think about that?
Personally, I can’t imagine this approach being more efficient for an architect’s daily work than using mouse clicks and keyboard shortcuts, especially when we consider software with a well-designed, customizable interface (like Archicad). I’m curious what made Graphisoft decide to go in this direction.

[jl_lt]

others are doing it, so they said "why not?"

currently the tool is still on the sphincter retentiveness stage, that is why it needs you to specify everything you need it to specify. Give it time    (about a decade should be enough).

[Lee Hankins]

I personally have found it quite useful during testing. Yes, it’s basic at the moment and you have to learn how to specifically ask it to do things such as selecting elements but I am sure it will develop and be enhanced very quickly.

[joaomourao]

Been using it for a couple hours and already finding it useful... find and select is what is does mostly, for now, but writing things down can be much easier and straight forward than f&s. Yes it does not unselect yet but this will surely evolve quickly, thus my question... By using it more and more, results will be better but what can WE do besides using it and confirming that WE like or dislike when using a prompt? I am already fixing some shortcuts for naming but there could be more documentation about this Ai assistant... Anyone knows about it besides this link?

https://community.graphisoft.com/t5/AI-Solutions/gh-p/aisolutions

Cheers!