Integrating Smart Contracts With Intelligent Document Processing - DAML Webinar

Captions as YouTube publishes them, usually machine-generated, so names such as Daml and Canton can be misheard.

1. 0:00

   welcome everyone thank you very much for joining and hope all of you are keeping safe today we've got another exciting session a very interesting session that we present to you it's about integrating robotic process automation plus intelligent document processing or ocr as you might know it more commonly along with smart contracts um and using this combination to accelerate your digital transformation so we hope you like the webinar uh i'm you know and um as usual we like to keep this

2. 0:34

   insights oriented and actionable so we'll start with a quick background of what is rpa how does it help what is ocr intelligent document processing you know how does it go well with rpa what are some recent advantages uh or advances that have been made in this field including artificial intelligence and machine learning and you know processing structured unstructured information

3. 0:57

   uh you know some of those things we'll give you a quick background on what smart contracts are and why we think smart contracts could be a nice complement to rpa and ocr and intelligent document processing so we lay out a quick simple use case uh and we'll you know talk from that perspective uh and then of course you know as we always do we like to keep it practical so we'll end with a live demo of how the technologies can work with each other so we haven't implemented an actual rpa

4. 1:26

   bot but you'll see how capture fast ocr technology is integrated with demo how it sends data into the ledger and how the entire workflow can be driven from there thus making the implementation of the bots also much more contextual and uh you know much more relevant to the point of arrival you know if you're into technology or you know you're looking at obviously digital transformation uh rpa is not the only tool you know so how does how do how do smart contracts work with that so with that i'd like to uh quickly

5. 1:59

   introduce my guest today uh gurkhan karagoz he's the founder and ceo of capture fast gurkhan you want to introduce yourself briefly yeah sure uh i am the founder and the ceo of the captcha fest capture fest is a data uh capture technology company uh we are developing uh smart and uh capture text data extraction technologies helping for uh big enterprise and mid-sized companies our technology you will see in the demonstration

6. 2:31

   the technology is really adding a new uh breathing the development environment working with rpa tools uh so that's it for now uh we will be talking in detail in the demonstration session yeah and uh capture fast obviously works with a lot of other you know popular rpa firms out there like automation anywhere work fusion etc uipath so obviously you know so it's a very complimentary technology and uh i am

7. 2:55

   manish grover as almost you know many of you know i look after channels and solutions at digital asset i'm on the partnerships team um so you know so um i'm going to shut off the video and gurkan do you want to mind shutting off the video because it sort of pops up this annoying thing on the right hand side and for the recording you know so i'm going to stop my video and i'm going to um

8. 3:18

   and pro i promise you'll see more of us towards the end but for now let's shut off the video okay so you know the background for all of this is that you know there are islands that exist within and of course between two enterprises as well so you know there's uh if you take any business process it's comprised of different applications that support the business process you know and

9. 3:42

   later in the uh in the webinar we'll see a simple business process where you know you um a credit card company is issuing a card so there's a new account system there's an e-statementing system there's a fraud detection system there's a collection system you know um there's a credit authorization system you know so many you know customer service you know applications that need to access databases there's a preferences system

10. 4:05

    so there are multiple systems that make up any reasonable reasonably complex business process and because of that what happens is you know generally we use different technologies to build up these applications and we connect them through various technologies like api and middleware technologies and all that and we often run into these challenges as business evolves as the application portfolio evolves you run into these issues of you know

11. 4:30

    there are multiple same data in different places or different data in different places often needs to be aggregated workflows are not shared uh you're you're putting you know patchwork on top of stuff to integrate to applications you know different applications don't see the same metadata in the same way and stuff like that and because of that you know there are operational inefficiencies that come up you know process breaks come up

12. 4:54

    and you know so all of those common problems and what happens because of that is and rpa robotic process automation came up as a way to solve some of those problems for us um and it's uh you know this is a very uh you know cursory glance at what rpa does but in in brief as some of you may be aware you know if you have repetitive manual tasks that are happening um generally they are costly um second

13. 5:21

    they are error prone and they are unproductive because you know generally you do swivel chair operations so what swivel chair means is you're taking input you're putting it into one system then you're updating the status into another system then you're checking the status of something else in another system based on that you're doing something else you know so there are repetitive manual steps that are happening um and uh you know the where rpa fits in

14. 5:46

    perfectly is when some of these things are difficult to remediate from a technology perspective you know so many times you know there is a manual process and it's going to take it's costly to automate or it's going to be pretty complicated complex maybe it's not the point of arrival architecture so you want to move away from that application and you're moving towards the new architecture so what do you do in the interim you know how do you solve and how do you get those operational efficiencies out and then you know your application

15. 6:17

    portfolio might be evolving so rapidly that you don't want to you know put an expensive technology based solution behind the problem so rpa in that case comes to the rescue um it's non-intrusive um you know it works with ocr it takes all of your documents you know ai based understanding of the document it simulates the user behavior so all of the swivel chair operations you know accessing one system doing

16. 6:42

    something else in another system looking at another system then coming back to another system you know so step by step things that uh human does it improves the efficiency a lot and effectiveness a lot because those people can now focus on more value-added things like you know actually include improving the business process or maybe serving the customer much faster or serving more customer requests um so that's where rpa you know the bots

17. 7:06

    are bots are implemented and obviously you know the benefit is since you're doing it automatically it reduces the errors but you know many people also like to call it short term because of the things in the in the second box here you know so your application may not be a point of arrival so very soon you may get into the point of arrival architecture in which case you don't need the bot so you need to update the bots again um or your application portfolio may be

18. 7:30

    evolving very rapidly in which case you need to keep um you know updating the bots that you put in place uh and you need to train the bots again uh so it's evolving and the shelf life of your bots will obviously depend on these two things so some sometimes shelf life of what says several years but sometimes it could be as low as a few months or maybe up to a year so with that um let's now discuss how

19. 7:54

    ocr comes in gurkhand do you want to talk about this yes sure thank you manish uh austr and arcade technologies are tied together uh i know that and we all know that rpa is looking new in the market but it doesn't the technology is coming from those mainframe years and i know that there are some rpa tools but they were not named as rpa before coming from 30 years 40 years they were

20. 8:25

    as mentions mentioned those tools are handling the repetitive tasks mostly and the ocr is extremely important for them because our paid tools uh just change the position of the data from point a to point b so to do so uh the rpa tool needs to have the data first so ocr technologies are helping them as you can see in the first items while rpa tools uh solving the problems of human

21. 8:58

    processing those are two solving the problems of human data entry without any austral technology our plrp tools will be have so the benefits are uh clear the the reduction of costs are extremely important and manual data entry uh eliminating the manual data into errors are extremely important because the data accuracy if the data is not well extracted from the document the operation uh loses money and time so that's why uh even if the ocr uh accuracy rates

22. 9:40

    even with the human uh typing increase rates uh the data quality the data speed is extremely important for rpa tools yeah and uh just so some simple examples could be that you know your medical assessment questionnaire that you're filling up when you go to a doctor for example or when you have a contract or a purchase order from in procurement or manufacturing processes all of these are paper-based documents

23. 10:04

    so something something needs to happen to get them into the system and that's where ocr document processing comes in yes definitely on the other hand not not the paper the paper in the world the usage of paper is looking decreasing uh but the digital pdfs are also documents and they also even if the data is digitally printed on uh pdf files uh the systems the rpa tools needs to

24. 10:30

    understand the layout of the document so that's why the term also is uh coming from the paper scanning but now uh we are it replaced uh it it's it means uh not only taking the data from the paper but extracting the text layer from the digital pdfs so the invoices purchase orders claim documents uh legal documents now everything is based on pdf files so the extracting the data with an ocr technology is very important to operate for the companies for the

25. 11:02

    organizations got it okay okay for example for the insurance companies i can say that this is extremely important uh the competition is so tough and uh when you are waiting to insure something you are sending your request by a portal and that portal is distributing the request into multiple insurance companies and we all know that who is accepting first we start working with that company so the time is more valuable than the data accuracy sometimes for

26. 11:38

    uh issues companies got it yeah uh so uh what are those challenges first of all uh we have to be emphasized that the ocr techniques are not the same uh it is not that simple it's not as simple because we are working on analog data and converting an analog data to the into digital data there are different approaches so uh all all settings are not the same you can see here the weak or inconsistent data is the main problem on ocr technologies because you have to

27. 12:11

    accept some error rates converting the paper data into digital data so the power of ocr tool is extremely important rpa tools are optimized on 100 percent digital uh flaws workflows uh because rpa tools are uh are developed to replace humans so how can you add the human in the loop this is ironic i know but if you are dealing with an analog data so that could be some help of the human because uh trying to add a technology for small touches could be

28. 12:48

    extremely expensive and not feasible replacing the the the whole human interaction so the second item again one of the important points of variable layouts uh the oscillos are most of them are uh location based just like the legacy forms process solutions but today our clients uh are working with hundreds of different vendors for example and invoice processing they are receiving one thousand two thousand different thousand vendors and each inversions are totally different layouted on the other end the insurance claim documents

29. 13:27

    legal documents uh thousands of hundreds of thousands of different layouts are coming with the same name as notice of hearing so the austere tool must understand the layout finding the exact word anywhere in the document just like the human brain as an human you just receive a document you look at classify it in your memory in your brain because you trained it before and after

30. 13:52

    that you extract the data for the document type our systems are working like that we trained the entire system with the type of thousands of different type of documents those system automatically classifies and then exactly recur information from the document so uh do you um so obviously there's data can be in structured form um and it can also be unstructured right yes definitely the analog data or the digital pdf could be structured unstructured non-structured semi structures for example invoices we are naming them

31. 14:26

    as semi-structured we are going to do demonstration as a medical historic document this one also is a semi structure what is the meaning of semi structure we know structure is is totally structured you know each data in what location we know on structure it's some kind of a letter we don't know anything about this layout but the semi is most of the documents in business world is semi structured documents such as the demos

32. 14:51

    we will show you the demonstration a semi-structure means you know you're expecting some certain fields but you don't know where that field exists or uh the location or the type such as the the the medical form you will see that you must accept to receive the name you must accept the receive insurance policy number a social security number but you don't know where because everybody can use different layout different

33. 15:20

    documents the the companies organizations your facilities are using different type of formats so that's why the ocr tool must smart enough to find the required information instead of analyzing finding the location okay got it and uh yeah this is a slide on capture fast competitive advantages yes definitely the information i gave till now were general uh problems general challenges general benefits of

34. 15:53

    an rpa tool or an ocr tool these are the the competitive advantages of capture fest uh in capture set here we developed and hybrid auto classification which is totally machine learning based but uh we used a totally different approach we are not using the technology the machine learning deep learning technology to classify the document but we are using the machine learning technologies to set the rules so this approach is really increasing the classification

35. 16:24

    uh accuracy why is that important because the machine learning technologies are mostly optimized classifying the objects in in cameras such as cats dogs apple cars pedestrians buildings birds so you can the cameras can capture 30 frames in a second so you can easily accept some error rates but in document processing you have only one document in other words you have only one single frame so you

36. 16:58

    cannot accept the error rate that's why the classical machine learning approach is not working in document processing that's why we added the hybrids we are not using the machine to classify the document but we are using the machine learning to set the rules this one is really helpful and increases the classification uh accuracy the second and third is our similar capture fest is has an expertise on

37. 17:24

    extracting the difficulties we always name that capture fest is solving the unsolved problems so if the document is easy you can simply use any uh rpl tools existing primitive capture technology but the documents are too complex capture fest can easily help you and the final bubble is also extremely important why rpa develops why rpa tools are created the answer of these final bubbles uh is sacred in this question why rpa

38. 17:58

    tools are developed because apis are not working because the it departments developers are not that fast i know that very simple tasks or requests are waiting for months in the queue of id departments in big organizations so rpl tools helping them because it is very easy you can easily simply use the graphical user interface and complete the text task so the the robots process does it so you don't have to ask any integration help with apis to the software developers uh

39. 18:36

    with the help of rpa one single uh analyst can easily connect the two systems without asking any help from those two ends something like that capture fast is also has a graphical user interface you don't have to develop one single line called you can easily connect the system with the help of our integration point you will see how easy to do during our presentation yeah and uh you know so that brings us to a

40. 19:09

    good point you know so um there was a study done by pega pegas obviously a pretty big you know this um it's a you know technology company that is also doing a lot of intelligent automation these days um they ran a survey of uh key you know stakeholders enterprise technology and business stakeholders and 85 percent of them said that rpa is an important component of the digital transformation and that's

41. 19:33

    exactly you know why you know you know that's uh based on what guru khan just said you know um things take time to change and in the meanwhile you still need to drive the operational efficiency you still need to you know look at how do you you know do stuff in the absence of a appropriate technology remediation and that's why rpa is really becoming important but at the same time

42. 19:58

    and that's the reason we are doing this webinar today is that rpa by itself doesn't solve all the problems you know as you know as you probably have guessed you know so rpa is is a surface level automation technology it com it bridges two different um data islands um you know because the bot can access both of them and you can train a bot to access two different applications but obviously

43. 20:22

    your underlying data islands and the underlying issues uh you know it you still need uh remediation to solve you know or you know help those data islands be harmonized second thing is you know your business processes as they are changing um rpa doesn't change it you know rpa uses the existing business process and it trains itself to execute it uh with you know little or no human effort but

44. 20:47

    there still needs to be in your digital transformation charter a way to harmonize your business process so that's the reason and the third one is of course you know improving your analytics and ai maturity which is important for every enterprise today not only for process improvement but also to drive savings uh also to drive customer experience you know more contextual experiences uh you know improving your uh you know

45. 21:12

    cross-sell and marketing efficiencies and stuff like that so uh you know um all of these things form part of the digital transformation and uh you know the the thing is like how do we handle those um and smart contracts then are coming up as a way to help with some of those things and that's why we decided to put this thing together um so what do smart contracts do you know so for anyone who's familiar with

46. 21:35

    uh daimler um you know or who's not familiar with tamil um essentially what smart contracts do is they provide and provide a way for all of your applications in your in your enterprise technology landscape to be able to access a single secure universal record of truth which means that all of the applications will see at any one point in time the same state of the business process contrast it with today where different applications have their own databases

47. 22:09

    and obviously you know they must keep in sync with each other in order for for them to be able to see the uh you know what that state of the business process is but there are obviously a lot of breaks that happen as you do that as you try to reconcile data from one system to the other as you try to come you know um send messages from one application to the other so smart contracts sort of provide uh provide a layer of layer between between the

48. 22:33

    applications so that you can they can all access a single golden source of truth so here's an example of what that looks like so you know this is a virtual pipeline right so any application um any business process in an enterprise is composed of say these uh you know multiple applications that are working together so here is a simple example of how a new credit card is approved it's a new account system comes in

49. 22:57

    underwriting approves it you know it goes into your estate printing systems it also goes into your credit authorization systems customer service can see the new card payments are being made on it those payments are being seen by your customer service team as well and your restatementing team as well and then obviously uh you know it might go to collections where a payment is not made so the collections team also needs a copy of the record that this customer you know

50. 23:22

    has been provided a new credit card all of these applications traditionally uh or many of these applications would typically report into different organizational hierarchies so thereby they would have their own databases own technology own system and they would try to interact with each other through either apis or file transfers or you know a middleware system that you know you send a message to the middleware and the middleware then sends a message to the other application

51. 23:47

    where their data is updated in all cases um you know you've got multiple uh versions of truth lying all over the enterprise and there's a significant effort which is spent in trying to align all of these data sources together whether it's to whether it's for customer service to see when a customer calls in whether it's for the customer service to pull data from all the different systems

52. 24:10

    or um you know if it is just to produce business intelligence reports or you know do analytics on you know what are some of the levers that we can pull to improve the business performance and uh you know so those are some common challenges and smart contracts provide a simple way for each of these applications to then be on the same page you know so it mutualizes the business process everybody sees the same data that's in the smart contract store so you avoid the issue of reconciliation

53. 24:36

    and it also maintains privacy so you might not want your you know your customer service for example um to for whatever reason you may not want your customer service to see that there has been a you know open fraud ticket against uh customers um you know against a against a customer's uh card for example or you know you may not want you know your e-statementing team to uh know that you know there is a there is a open you know collections uh process underway so

54. 25:06

    in order to avoid you know allow all that smart contracts provide an easy way they maintain privacy and they also provide a common smart contract store each of these applications can obviously have their logical views of the underlying data so this is why we think that rpa plus ocr plus smart contracts may be a maybe a good combination because they allow your bots now to access a common api instead of having

55. 25:31

    to access different apis because you know your bots are becoming more intelligent now instead of them accessing multiple different apis you have a common api that is exposed by a smart contracts layer and they don't have to do you know too many swivel chair operations as well so your your uh you're progressing on your digital transformation charter while still allowing your bots to you know um

56. 25:54

    exist where they should you know where where you have a need you still have the bots but then you're progressing along your digital transformation charter as well so you know solving your data island problem harmonize your business processes and obviously improving your ai maturity as well so this is what demo looks like you know just i won't spend too much time on this but essentially the smart contracts

57. 26:16

    language has the ability to run on multiple blockchain platforms as well as multiple sql databases which means that you can obviously this process that we saw this is within an enterprise but then tomorrow if this company wanted to interact with an external entity external enterprise and if they were deploying a blockchain you know between the two enterprises then obviously this

58. 26:38

    application that you develop using demo smart contracts would seamlessly and atomically interact with the um you know with the with the with the other network that you might be setting up so i think those are some great benefits and obviously damal is fit for purpose for multi-party systems which means that there's by default privacy between the actions that one party is taking versus the you know actions that the other party is taking so by default you have privacy and

59. 27:03

    obviously you can open up the privacy but you're able to control the privacy in a granular fashion so with that yeah and feel free to ask questions actually i forgot to mention that so feel free to ask questions if you have any questions or any clarifications that you need but i think now let's get into the the demo portion and for demo what we've done is we've taken a simple use case which is something that almost all of us would be familiar with is going to a medical provider um you know filling out the medical

60. 27:42

    questionnaire form uh and then you know when you go the next time looking at the document and updating the information which almost never happens you end up filling out the whole form again most of the times or somebody prints out a paper form for you to um for you to update and then that thing is fed back into the into the system so we thought let's take that simple process and model it out using smart contracts uh

61. 28:07

    and integrate it with this line that you see the document processing board this is capture fast capture fast is looking at these medical questionnaires um you know ingesting the data making sense of it and then sending it to the smart contract layer right now and you know thereby kicking off the business process and being an integral part of the business process um there's a blog uh that we wrote on

62. 28:31

    this uh last week which you can which you know which the link will be on the last slide and i'll also send it out in the follow-up email that we send send along with the recording um so gurkan uh do you want to take over from here and show the demo yes sure so i'm going to stop sharing yeah and probably you know yes i have the permission now i'm sharing my entire screen here is the capture fest application all right i'm gonna resize a little bit so

63. 29:03

    yeah capture fast is a web application uh you can see we have dashboard here uh with the help of forks and with lots of connectors uh you can easily define your document types and upload the documents and extract the data before directly uploading the document and looking at the results i just want to show you how easy to do to define a document type into the system you can see here the document types is administration only

64. 29:33

    tab only one administrator can define a new document types into the system you can see we have here a medical forms already defined by us and when i click on it uh you will see that we have four steps to complete uh the entire document extraction tool the first one is uh adding a sample document for defining them and the next step is looking at the adding the fields as you can see here everything is automatically captured and and and added into the system and each label

65. 30:11

    has different uh properties uh in the user interface so with the help of advanced options uh you can dive deep in the properties or some advanced options to extract the data perfectly so the final one not the final one the third one is the destination as i mentioned we have a user interface here uh we are not typing any code we are not asking any id help uh we are just typing the connecting endpoint of the uh destination in this case in this

66. 30:45

    example uh our endpoint is demo and you will see that the url the api url is here uh we are defining uh the fields the labels here on the left side you will see the fields the tags are coming from the web servers of the demo here and on the right side these fields are coming from the capture face so you can easily click and uh map one by one without typing any single line of code and after this when you complete

67. 31:22

    when you go to the next you can invite your uh users capture fast is a license per page so it is not asking any limit on the number of users this one is also another advantage uh you can even invite your end users so the system is not checking for the domain for the even for the free version you can invite millions of people as the verification or the capture users the capture users are mostly sending the documents into the system so who are these human verifiers people

68. 31:59

    who will see whether the data was extracted properly human verifiers the verification is very important this is the human in the loop option sometimes we especially in united states uh the culture is coming from the the decades of facts machines lots of companies are using multifunction devices you know those big xerox machines and they all have computers in it and you can use them as the fax machine email server uh document scanning printing for everything so

69. 32:38

    the companies are still using the settings coming from fax machines so the document resolutions are very low so sometimes there are some scratch on it some dirt on the the machine so the the analog data the scan papers uh can fail extracting the entire data with 100 percent accuracy at that time adding a human just to verify the data could be helpful you will see during the demonstration so these verifiers can be

70. 33:11

    in the company in the organization in the or anybody in the world you can just add people and invite into the system so let me upload the documents one document here i am going to use uh one of the sample documents i'm uploading the upload is completed we have lots of input channels that test at that point i just want this i simply used our user interface but we have mobile capture created by capture fest

71. 33:42

    or we have some apis any third party software can easily send the documents and most of not most of all the rpm tools has connectors into the system and rpa tools are sending the documents with the help of our api the system is processing the document you can see it is waiting in the verification uh the document is digitally created a pdf file normally we shouldn't verify the system extract the data and

72. 34:15

    automatically populates in seconds but in this uh demonstration i just want to show you how easy to verify a data with the help of capture frag that's why i enabled the verification i enforce the verification for the medical history forms here uh when i click the verification this the system pop-ups the uh the data and the image side by side so you can easily check out uh check the data and

73. 34:46

    while if the everything is perfect you just finish and the data is going to be transferred to them sorry just one quick question that has come up from the audience it's come from amit what he's asking is does it support handwritten text yes in enterprise level uh we have handwritten recognition engine uh but it is not scripted languages you know the language that the letters are

74. 35:14

    connected to each other this is the handwritten scripting letters our system is optimized on the the separated letters okay got it by the way the handwritten technologies are not uh developed anymore there are some companies created handwritten recognition such as google vision app so we can easily add into the system uh but these days everything goes digital uh so that's why we don't have enough

75. 35:44

    clients who are asking us handwritten recognition but it is extremely easy to add our technology is really open and we have user exist technology and these technology these topics is not today's uh topic list but we can easily arrange another private uh session so i can explain or i can show how i can handle the handwritten recognition okay perfect thank you for each field again without typing any single line called you can add lots of regulations

76. 36:16

    you can add regular expressions you can add some rules mandatory if you you can force the mandatory fields and when you type anything uh when you miss anything the system will not let you uh approve the document you can see the last name is mandatory the line is read when you start when you try to finish the the document the system will force you to type to extract the last name so that's why i'm gonna retype it and now the all fields are green i don't need to do

77. 36:48

    any more control so i can easily finish and check the documents uh in demos now the data is extracted and transferred from capture fest capture fest is done here let's go to demo and check the data extracted in the here this is uh yeah so for everybody for those of you who don't know project double project double is a cloud environment to host demo applications so applications that are written in demo can easily be hosted you know out of the box actually so just upload your demo model and

78. 37:29

    dabble allows automatic apis and you know authentic automatic authentication and all that stuff so it's a pretty a nice uh platform i'll include a link to that in the follow-up email so that's where this application was developed and uploaded yeah as as you remember uh during our document type destination point uh i mentioned that uh there is a url which is the api of the demo

79. 37:52

    and the system is going to populating this web service the sdk and now here is the projects and when i click on capture fest and the live data here this is yeah yes this is the data we just extracted you can see in json format on the right side remember the last name uh lake here yes in lake and the all data is taken from capture fest and sent to demo in json format so this is it from my site uh probably

80. 38:38

    uh before switching to questions and answers yeah there is one more slice probably yeah gurkha just one more questions come up from the audiences how can we add scanned reports to it scan document okay the system uh this system is not looking if the document is scanned or or digital the system when you upload any file even if that could be a digital or scan this system will automatically understand and take action per the type of document type so you don't worry about the

81. 39:10

    the document is scanned or pdf you just send the file but of course you have to take the file uh you have to scan the paper first and create a save into a file okay because any digital any digital input is acceptable yes anything is acceptable the system accepts pdf files png jpeg tiff words there is no any limit in the in the file format okay perfect and uh yeah so i just wanted to just while you are here gurkhan if you can click on

82. 39:47

    um if you can click on deployments yes and if you click on that captcha yeah over here click on main.demo yeah so this as you as everybody can see is the demo model so here is the data structure in the demo model that we expected from capture fast which has been declared here and then if you go down a little bit further below scroll down a little here you see the workflow so it's only i think 60 or 70

83. 40:17

    lines of code demo code that were written to onboard multiple providers and be able to accept this data from capture fast be able to approve the data be able to give the customers access to their own medical information which is kind of difficult today um but you know innovators already working on that so that's not the topic for us now but um you know it allows us to do a lot of things within 50 or 60 lines of code you know smart contracts

84. 40:42

    so that's how it that's how it sort of uh helps us and uh it obviously maintains privacy so if you see a consumer role um you know the is consumer and provider which just means that a customer cannot be created without the permission of both the provider and the customer on the system so automatic consent management permissions management uh also comes out of the box um so yeah

85. 41:06

    with that uh gurkhan i'm going to stop sharing and go to the screen now um yeah so this is the overall workflow that you just saw modeled in demo so this was the function api that capture fast was calling but of course all of that has already been built out you know so if there's a customer the doctor must approve the document you know before it gets uh before it gets populated in the system and then obviously here is where you'll

86. 41:39

    see the automatic benefits so there's provider two here's another medical provider a completely different network if they want to request this customer's record for example the customer went to provider two and they want to leave provider one you know and they want to take all their medical records with them it becomes uh it's a pretty complicated and costly process for uh for us to do that so in demo you can

87. 42:03

    just model provided two as a party you can create a workflow where uh you know you create a request the customer approves the request and then you know you'll uh request provider one to transfer the transfer the record and the record is automatically transferred to provider 2 and optionally if that's how you want the system to be provider 1 does not have access anymore to that information from the customer so you can obviously meet all of those privacy confidentiality requirements as well so all of this was modeled out in that simple demo

88. 42:34

    but you know that's how rpa ocr and smart contracts can work together and obviously now you can put a lot of you know you can still implement rpa technology but now you know backed by this nice you know business process that has been developed and that is sort of spanning all of these different systems that are that could be you know implemented at a at any medical provider um you know it makes your bots much more efficient and it also makes them easily auditable as well because whenever they take an action it can be

89. 43:05

    logged into the onto the ledger the smart contract ledger which means that now you have an immutable record of you know how the word functioned or you know what actions if they take obviously so it's good from a compliance perspective as well so um i think okay so i think that brings us to the end uh so as we promised uh oh one second there's another question uh let me uh gurkan can you

90. 43:34

    see the question that came up yes i see uh rohit's question can you explain a bit more project double conception is that one okay yeah good so uh yeah i couldn't see it because then i'm sharing for some reason zoom doesn't oh okay okay then let me complete the question then can you explain a bit more project double consumption and how smart contract can leverage it how users doctors patient provider can use this json provided by system just demo yeah exactly so

91. 44:07

    project dabble is is the cloud environment for um for demo so you saw the workflow this workflow that we developed in in demo um this workflow is then uploaded to project double um it automatically gives you uh you know the ability to expose it it automatically exposes all of the demo api and uh you can also host your front-end um user front-end on the on demo on on project devil so what it does is it gives you a very quick fast time to

92. 44:39

    market to take your demo model take your workflow take your business application and deploy it without having to worrying without having to worry about all of the other infrastructural complexities that you might come up you know how to host it performance security authentication all of that stuff is automatically taken care by project devil so that i think is what project double does and i'll include a link in the follow-up email that i send out the other question was you know how do

93. 45:03

    you go for adoption um and uh you know that's where i think some of these these properties of damn will help so you know your dml application um is deployed on project dabble right here um you can either choose to onboard everybody to the same ledger or if there are different providers who want to use the demo application but they want to deploy it on their own ledgers or their own networks but you still want to be able to

94. 45:33

    provide a connectivity or an interoperability between those two networks that's also possible so in the ideal case of course you know not every medical provider is going to you know incur the overhead of uh you know trying to set up their own network they will probably just outsource that infrastructure management to to someone and trust you know like amazon for example or microsoft so you could deploy this on amazon or microsoft cloud

95. 45:58

    um you know people are pretty happy with that and then you run your run your application on top of it so that's what project devil allows you to do it gives you a completely managed environment for for your demo applications uh managed we have another question uh conor says asking can we rule out medical legal cases uh i'm sorry can can we rule out medical legal cases i'm not sure about medical legal but of course we can rule

96. 46:30

    out anything if we define in the flow yeah can we rule out medical legal cases i'm not sure we understand the question kona so if you want to clarify it i think if you're looking at decisions based on the data that is being captured automatic decisions you can definitely you know write a smart contract that automatically executes when something is encountered and it can filter the filter the data or move the workflow

97. 47:06

    along in the right way so that's definitely possible i'm sure capture first can do certain certain rule-based action as well guru khan when you when you extract the data can you make some data you know can you make some decisions based on the data that you're capturing yes of course we can do we can add fields or we can add some ai to make decisions we have a simple workflow in capture invest and according to the

98. 47:36

    extracted data we can forward the distance we can change the destination point yes that's possible yeah and uh i think there's another question on what if a doctor does a wrong treatment so obviously you know this workflow is not intended to do that but obviously you can build out a complete workflow in smart contracts and you could integrate that with any existing hospital management system as well on what happens you know in that case when

99. 48:00

    when incorrect treatment was done you know what are the rules what is it what should the workflow look like so obviously you know it doesn't limit you the workflow doesn't limit you at all to implement whichever business use case you might want to okay perfect so i think that brings us uh to the end of the uh questions and uh we just wanted to show you this so if you want to get in touch with gurkhan obviously gurkhan at capturefast.com you know definitely read this blog it'll come out in the presentation i'll also

100. 48:33

     send you in the follow-up email how dammit can complement robotic process automation so that will explain the workflow that we just discussed in a little bit more detail uh you can watch a demo on youtube actual capture files demo and then of course you know if you have more questions about damil then definitely join the community uh on discuss daimler.discuss.com and um download the demo sdk also at docstoredemo.com

101. 48:57

     um so with that i think there are no more questions and we are at the end of our webinar thank you so much for uh for joining um oh there's just one more question that popped up um oh conor saying thank you thank you thank you everyone yeah so um thank you manish you arranged this webinar and that was wonderful for us and thank you everybody joining us today talk to you again yeah thank you very

102. 49:25

     much bye you