Possibly Coming Soon: Creating an App Directly from an Image

[Joseph_Ryan]

Hey everyone!

We’ve been brainstorming about ways to help new users get into Appsheet, and we had an idea that we’re pretty excited about. What if you could create a new app (or add a new table to an existing app) simply by specifying an image of a form? The idea is that you’d upload an image and we’d use machine learning to understand what the fields are - and then use that to build an app!

We picture it working something like this. First, when creating a new app (or perhaps a new table), you’d now click the “Create an App From an Image of a Form”:

Next, select an image from your computer. I’m using this simple receipt form I found on staples.com, but this could be any kind of form - including something you quickly slapped together yourself in a Google doc or slides or something like that.

Click submit, and voila:

A new table (or tables) and a new app with some default views, including a pre-trained OCR model for quickly adding new records.

Right now, all this is just speculative - we think we can make a feature like this, but wanted to gather some feedback before we commit. So, what do y’all think? Is it something worth pursuing?

[Bahbus]

I’ll say that its definitely not worth not pursuing. But ultimately what I am hoping for is that you use this feature to build upon the idea of more customizable views. Where one day we could draw a picture of how we would like app to look like, and it will automatically pick and customize a view (or many views if given many pictures) to match.

[Jonathan_S]

I don’t Agree, this will be a long hard journey, full of bugs and user requests for more functionality. This will cost a ton of money and will cost more than it will bring in. I don’t see this as a good idea no matter how amazing it sounds.

I don’t believe this would be worth pursuing,

I say focus on all the other requests instead of these EXTRAS

[Bellave_Jayaram]

I agree with this point of view. Not going to be any more helpful than Spec. In fact I would say Spec could be enhanced to include this feature as well as the UI mockup feature suggested in this thread.

[hcquadros]

Jonathan_S:

s for more functionality. This will cost a ton of money and will cost more than it will bring in. I don’t see this as a good idea no matter how amazing it sounds.

I don’t believe this would be worth pursuing,

I say focus on all the other requests instead of these EXTRAS

Agree

[Remco_Edelenbos]

I agree as well, i think there are a lot more other request from appsheets (already) costumers.

• Control users for instance is really really not optimal

• Get requests

• Weird overlapping text in Charts

ect ect ect

[retailpartnerco]

Sorry to say, but I don’t really see a need for this. It seems like the same amount of work to either build a spreadsheet form, or build a google slide and submit the picture. I think there are other features that would proceed this one. The only use I can see is if people use paper forms, and want to take a picture of their form to make an app, but most apps probably do not start this way, so it wouldn’t be worth making a huge feature for a small percentage of the “AppSheet Population”.

More Prominent Requests:
38 Votes

Action: UPDATE EXISTING ROW(s) using values from this row Feature Requests

This feature request is for a new Action-type that would replace the need for several actions to accomplish the same task: Update a set of rows using values from this row Currently we can: Add a NEW row to another table using values from this row, and Execute an action on a set of rows But to pass values from the current row to a set of existing rows requires a bit more setup. This could easily be defined in a single action and could use the same UI as the existing actions: [Scree…

33 Votes

Add filter capability on top of the columns Feature Requests

I’m still looking for this capability. Is there anyone to have an idea?

28 Votes

Custom Y/N Text Questions

Just added my Vote. That gets you a double.

26 Votes

Freeze 1st row of name Feature Requests

Anything on this feature yet? seems trivial. But would make a large impact on efficiency

[Bahbus]

retailpartnercom:

The only use I can see is if people use paper forms

Granted, I’ll admit it is possible that there is a decent number of potential new customers that would like this feature.

[Austin]

If all I have to do is take a picture of my current paper form and appsheet will create the sheet and even a basic view or form for it then I would be very thankful for that. Helps bridge that very small gap. It’s also a good marketing feature to have. Creating your own app is as simple as taking a picture.

[Austin]

I think so long as this is combined with the current OCR feature as far as development goes then it sounds amazing. I don’t think it will be too terribly different from the OCR feature so I don’t see it being a big deviation from current developments. I think it depends whole heatedly on what resources you guys leverage for this. We currently use Amazon’s Textract process which in a sorta similar way gives us this information. I know that finding a value that is likely a field in pictures is readily available with current tech. If appsheet is gonna develop this from scratch then I agree with Jonathan_S but if this going to use current services then I say its good.

[Ron_Mansolo]

Could be very useful in the healthcare industry where many businesses still use paper forms.

[Grant_Stead]

Aight, I’ll drop my two cents in…
It would be cool for inexperienced app creators that want to digitize a form. Beyond that marketing funnel, I don’t see much value.

[WillowMobileSys]

First, I agree with the sentiment that additional focus is needed on existing features and recent Feature Requests. Please, we need these features to be as spotless as possible.

---

I think most are responding from the perspective of creating new apps from scratch.

However, there is a huge market for conversions. If you think that many businesses are NOT using paper anymore then you MUST think they are using software (but see last comment). A lot of those businesses having aging software looking to migrate to something fresh and new. And most of those, I’ll wager, use both software AND paper.

I tend to think bigger picture. What if the feature described here could be extended to use screenshots from older computer systems in addition to paper forms?

Then the process could take care of a major portion of a migration with just a few keystrokes. Will it produce an end result…no. Will there be customization and movement of columns to fit a new structure…of course. But in large systems, it can save hours upon hours of normal migration work.

---

If you take all of the above away, I see this functionality being just as useful as creating a sheet with my tables and columns and then creating the app from that. I’m sure we all use that feature? This is one step further, you simply show an image and the sheet is created for you, the app is created for you, OCR rules are established. I think it is an invaluable step towards an even easier way to create apps for many.

By the way, I have done some research for marketing purposes on businesses still using paper as the primary documentation system. It will probably blow your mind how many still do!! Look it up!

[Caleb]

I think this would be very valuable. There are many companies still using paper forms for a lot of their business functions.

There are also a lot of companies where:

1. They don’t really have someone technical who can build apps in Appsheet despite how easy it is.
2. More likely, they think they do not have the time to invest in figuring it out.

This feature would bridge that gap even in just getting people started in a powerful and very quick way. Once they had this starting point, they would be convinced to spend the time to learn Appsheet or spend money on help. I would see some only using this feature and it still being of great value to them.

Another interesting thought is this could eventually lead to people being able to create an app from their phone without needing to go onto a computer, maybe even with pen and paper. Think of all the countries where a majority of people are starting to have smartphones but not computers. They are not an ideal market as of now, but in three to five years they might be.

[Dan_Bencsik]

I would he extremely excited about a feature like this. We have legacy forms we still use on paper because we haven’t found a good way to bring them into the web or mobile space.

[Jon_S]

Considering the amount of ‘thought’ that needs to go into creating a proper relational database in order for Appsheet to run like clockwork… this kind of liberal approach to creating ‘apps’ via images makes me feel a bit uneasy

But sure why not…

[Grant_Stead]

I think that’s what they’re trying to learn more deeply from… Hey a user uploaded this image. We made this app, and these are the changes they made to it to get to a final app. With this Appsheet can use AI/ML to learn how to better construct apps on the long run…

[Jon_S]

+1 Nice way to look at it! …

I guess someone has to think outside the box for slightly off-beat stuff like this

P.S I wish they used some of that secret AI/ML sauce to read through some comments here on the community posted by developers to improve UI for existing customers…and make their (developers) lives easier. Apart from some minor stuff, the last 5 months have been a bit dry in terms of useful updates (UX/UI-wise)… ¯_(ツ)_/¯

[Grant_Stead]

Yep, it seems like they’re doing a lot of foundational changes right now. I expect that within 3 months they’ll be through the woods on some of that, and some of these beta/alpha features they’ve been teasing will be more production stable.

[Fouad_H_Saad1]

I think It’s very helpful to focus on UI mockup. It’s very important issue, where It discussed in many previous occasions.

[Instituto_San_P]

Excelente!!!
Un camino para transitar de los negocios con documentos físicos al mundo digital.
La mejor manera de migrar y sin demoras
Adelante!!!

[aucforum]

I think you or we need to think more about what the marketing plan is for AppSheet. How to make it more well known among spreadsheet users the world over. I am assuming this is part of a strategy of seeking new users?

[Rousseau_Etienn]

This is a great topic . I guess hat a lot of apps in fact at the beginning answer to this king of need . Digitalise frequently used docs in the group , crop the info and transform it to other recap docs , or invoice , or any .
This is very exciting and probably changing the creation mode . Please continue
Best regards
Etienne

[1minManager]

I think would be a waste of time and like someone above said, no more than a marketing gimick. Even if you can get it to work… what then? You don’t really want a userbase who thinks an App can somehow build itself! Your going to get inundated with people who’ve created an App from an image, only to get stuck on changing a display name.

I think Appsheet has already set the bar low enough by showing someone they can create a speadsheet table with some sample data, then simply point Appsheet at that.

Like people above I feel that Appsheet should continue to expand on the functionality it already has. I dont think we’d had any useful updates in a while…

[Jonathon]

Agree with the overall sentiment; however I feel like there have been some pretty good updates over the last year.

[Koichi_Tsuji]

There was “Feature request App” to collect the feature request wide and broadly from Appsheet users. Now its gone, as far as I understand.

This community has a category of “feature request” and there is “vote” system to see what kinds of features would collect the support broadly from the app user and creator, presumably Appsheet team decide or see where the priroity resides.

After the Google acquisition, my personal feeling is the release and introduction of new features are pace is surely getting slower. I have not seen any meaningful “Feature Friday” almost nothing for this year of 2020.

I m not sure why and how the Appsheet team is collecting the opinions and requests for feature request through this community place these days? it is quite some time since I saw the voice in this community for new feature is actually taken by Appsheet and the feature is introduced in swift manner. It was and i saw such opporujnities before, but actually it was before Google era. To be perfectly honest, I dont see any gain and positive impact by Google aquisition for the time being. People may say we need to be patient for the time being it is around the corner, but I m not on that sort of feeling.

If this continues, I will give up to through the opinion as well as feature request to this community place, as it probably end up in vain, unfortunately.

[Caleb]

I think we need to be patient.

1. We received a lot of awesome improvements from the fall of 2018 to the end of last year. I had worked with other platforms before Appsheet and I had never experienced as many meaningful improvements in a year time span. Not even close.
2. I think Appsheet is busy integrating with Google Cloud and possibly GSuite. I can not even imagine the undertaking that is. I think that is going to bring us many new capabilities. There were probably a lot of updates that have been tabled because of the deep Google integration. Once the integration is complete, I think we will start seeing updates again.
3. Unprecedented times for the whole world.

[TDhers]

Thank you all for the feedback on this thread. Sorry Koichi to disappoint you with our pace of innovation.
As Caleb mention, we are working hard on upgrading the code to the level if expectations that customer have for any Google product, when it comes to Security, Compliance, Privacy.
We are also indeed moving our infrastructure to GCP which will give us an infinite capacity to scale the service without exploding our cost.
We are also looking at deeper integration with Gsuite and other Google service.
We are onboarding a significant amount of additional Google engineering resources in the Appsheet project which means we are spending a sizeable amount of time teaching and onboarding folks internally.
All this does indeed explain the slow down of visible innovation being released on a regular basis for the time being.
However this is a temporary slowdown before we can accelerate again with a stronger, larger team than we had before the acquisition.
We are still watching the community discussion daily and helping as much as we can there.
And Yes the Feature request section of the community IS the place to go to submit ideas and vote.
I shut down the app because multiplying various channels for customer input only end up diffusing the feedback and spreading the demand patterns… I’d rather have as single customer request forum that gives us a more accurate picture as to what is trending in the community in terms of feature requests and needs. I know we are not providing a lot of insights in terms of what we are planning on tackling as part of this feedback. @JCadence is working on this. But rest assured that we are still there, we are actively watching and listening.
I am super pumped about the future and all the cool things we are working on beyond Azure to GCP transition and GSuite integration. Be patient with us. The acquisition was only a few months ago and we are still learning to operate efficiently in this new world we are in, but rest assured our commitment to you our loyal customers and to innovation is not diminished one bit…

Thierry

[Koichi_Tsuji]

hi @TDhers
I m sorry for being a bit noisy boy here, but I will stay patient. I m putting the background music in my office, from oldies, Gns Roses, “Patience”
I started dealing with G Suite users whoe are asking integration with G Suite service, such as grouping and intergration with G Mail to let the appsheet collabrately work with bunch of G Suite tools.

I tend to be “demanding” for new features, but most of them, I m speaking out “on behalf of app users” who daily use Appsheet app.

Again, thank you for your words @TDhers, and looking foward to seeing upcoming series of new “Announcements” here.

Keep rock’n roll.

[11143]

I don’t agree.
it’s not usefull.
Invest your efforts in other URGENT future’s requested that is so much usefull as mentioned here.

As example,
Change column value ACTION with user input (popup) without using slices and link to row.

The email workflow is also complicated, why there is not ACTION send email.

These problems we deal them everyday.

Make table from image… it’s only a gimmick beside all the other important future requests.

[TDhers]

Guns N’ Roses. Nice!
I’ll share my response in a separate announcement for more visibility. Jennifer and I were talking about focusing one of the upcoming Office hours to have a candid discussion with you all about what’s been happening in the last 6 months and what is happening the next 6 months…

Thierry