We are about to launch a private beta of our platform called Bitskout. It allows to add AI capabilities without code to monday.com (more services to come). Using AI allows you to make decision based on context which in turn makes teams more productive and avoid waste of time.
You can create models via UI for images, videos, text and build your workflows based on selected parameters (e.g. text sentiment, key phrases, topics, image labels, image contents). There is a small extra - for each task which is approved by a workflow, the user gets a token allowing managers create reward schemes, convert them into existing point systems, run campaigns and etc.
Models then can be used a part of a workflow that is assigned per each task. Some use cases:
check the text sentiment to move draft article/PR to different groups for review
check image vs previous image to document equipment state
check automatically assess home damages to route to appropriate board for analysis
check that design templates have all the necessary buttons and text
check “before/after” picture to validates someones physical work (including location check).
If you are interested, please DM me and I will add you to private beta list. Otherwise, send your comments, we would really appreciate it.
Here is an example where my co-founder added a workflow to check sign-in page design templates to have necessary items like email, password, forgot password fields to avoid manual double checking. It is quite simple case for us but think about it in a span of larger team or organization. This tiny workflow saved us hours of unnecessary checks and sync ups + I got my token reward - in our team 1 token = 1 Fazer candy !
Honestly, I’m a bit stunned at what you were able to come up with using our SDK and framework! I think it’s an amazing app and I appreciate you sharing this - I love the candy idea to make the team more engaged and also to remove unnecessary manual steps from the workflow.
And another feature released today - document comparison. In the example we compare product labels to see the differences between the template and the produced label. Same thing with the packing list for shipments - using Bitskout you can automate comparison and avoid costly mistakes.
The result of the comparison is a joint image with highlighted differences.
And one more example of document comparison, this time supporting PDFs. Of course, you can compare easily two PDFs in Adobe Acrobat. The problem is that there is a lot of friction to achieve that - download the files, open Acrobat, select one file there, select another, then see the differences, and etc.
The difference here that comparison with Bitskout is a part of your project workflows - it happens where you work and it can be a trigger for the next step without you doing manual hassle.
You can use forms to upload documents and trigger sending results via email.
We’ve just released a new functionality called data extraction allowing you to get data from any document, screenshot or form and map it to Monday.com fields. In the video you will see how we get data from various sources like an invoice, a W-4 form, architectural drawing and a screenshot. Take a look:
We use a template matching functionality in our platform allowing you to setup a certain regions on the template and extract data from all similar documents.
Another use case that we see getting traction among our clients is related to capturing results of real-world work.
It works for clients from professional services, manufacturing, different vendors and also in property management. Any work that is done in real-world is being captured by a photo or video proof. And A.I. is used to analyse that the photo is legit.
This way companies can get several serious benefits:
Log a history of work
Capture expertise that are in their experts heads
collect proprietary data for a new wave of automations
We’ve just released our new feature that allows mapping the results of A.I. model execution to labels in your boards. This will facilitate more automations because now you can trigger automations based on that mapping.
We’ve recorded an example where we use that technique to detect document types and select appropriate labels:
Another example of advanced workflow with help of Integromat, monday.com and Bitskout that we’ve created for one of the clients. Using that workflow allowed the client to decouple work completion and reporting - once the teammate complete the task, the document was uploaded to his folder and Bitskout and Integromat took care of everything else. This allowed to remove a lot of unnecessary processes and simplified the work for the team.
The process is as follows: once the file is uploaded to a folder on Dropbox, we use Integromat to detect the new file. Then once we detect the file, we create a new task in monday.com with a file URL. The new task triggers Bitskout no-code A.I. workflow that checks what is the document and assigns the label. Once we know the document type via label we trigger a monday.com automation to move the file to a new board where, again using Bitskout no-code A.I., we extract information from the document.
The example is using supply chain documentation (bill of lading) detection but obviously it can be used with different documents.
Another use case that shows the power of monday.com combined with no-code A.I. - automating creation of Knowledge Base with help of monday. The key item here is the creation of a board and a form where all tasks are coming in. Bitskout A.I. is used to automatically generate keywords, analyze images and create machine-based taxonomy for your request that would allow you to classify your content.
Another demo and this time about sub-items. In this video we show how you extract a table from a PDF document and convert its rows into sub-items of a Monday task.
This functionality is really useful when you have a table that is populated based on user action (some kind of an order of something) or selection (a set of options). And also it allows you to parse tables coming from a document delivered by a partner - for instance, invoices or receipts where you want to have itemised list be available as sub-item of a main item.
Table extraction functionality is available right now.
Cheers,
Ilia
I’ve created a bit more enhanced example where I used JotForm configurable list field to make a dynamic choice for the user to add items to the table and then saved the submission to Dropbox. This way I’ve managed to convert JotForm submission into monday.com task with sub-items and extract information about the order to the same task:
We’ve created a small guide which describes how we’ve created the supply chain document detector in monday. We’ve used the same trick with another client detect incoming email RFIs from different contractor and moving them to respective boards and then extract information to monday fields.
You can use the same trick to track any documents you’d like.
Here are some simple examples how we helped several clients last weeks to improve productivity. Everything was achieved via Integromat, Monday.com automations and Bitskout.
The key message was to make monday.com task as a standard item of information flow in the client’s processes.
We are so excited to release new set of feature that are mind-blowing. Here is the video where I simply add a set of questions, then run Bitskout A.I. model, and receive correct answers from Bitskout for the text. This can be used in the variety of use case to automate workflows where human reading is required.
Check the review (extract product names, is customer happy or not?), etc.
check meeting transcripts (was something approved, any actions points, who was in the meeting)
press release processing (what company release is this, what the released, what was the key message?)
etc.
Here is another example of how automated summaries can help users with their work. In this video we demonstrate how you can organise the meeting tracker and be aware on a high level about all meetings that happen in your unit or organisation.
Having this information in text form would provide a new layer of data for reporting and allow you to get instant real-time slice of what’s going on into your inbox almost every day like a report.
In this case we’ve created a meeting tracker use case - each Zoom meeting that we have is recorded to Zoom cloud. Then Integromat scenario pulls out the recording and saves it to Dropbox automatically creating the task in monday.com with the meeting name and link to a video. Then we trigger the transcript which is uploaded to the same folder, and we detect that via Integromat and upload the transcript to the same task, but a different column. And finally we trigger the summary extraction using Keypoints from Bitskout.
