Hello and welcome everyone we're really excited to have everyone here today for our weekly oci deep dive I'm Maria Philip product marketing manager with the oci go to market team and today we're joined by presenter Harris kesi a integration specialist on the Oracle EMA Ai and a modernization team and he'llTake you through the live presentation and demo we also have Juan Carlos Gonzalez Carrero who is QA lead for this webinar he's also app integration specialist on the Oracle EMA Ai and app modernization team the topic of today's webinar is empowering insights automating sentiment analysis from customer feedback which is about theTransformative potential of oci AI language service you learn about the approach to harnessing the power of language processing focus on sentiment analysis derived from customer feedback data our presenter will also cover how to use Oracle integration Cloud for seamless automation facilitating the extraction of sentiment insight and now that's enough for me soHarris please go ahead and take it away thank you hi everyone thanks a lot Maria for the introduction uh my name is hares kesi and as Mario mentioned I'm joined by Juan Carlos who would be answering all your questions so please keep him busy um today's uh topic we will beCovering uh how you can get insights from sentiment analysis from customer feedback and we will be covering a live demo so let's start you know uh in the world that we are living today every business every action that we take we are always looking for feedback from making aPurchasing decision if I go to a shop I need to buy something I always looking for some kind of a feedback to watching a movie maybe I go to IMDb first to see what kind of a rating that uh movie has uh to going to a trying a new newRestaurant in our world today every decision that we are making we are always looking for our peers uh feedbacks and in that case the businesses are really really interested in what customers are saying about it now while uh I have been talking about this I'm sure you some of you would haveRead some of the feedbacks that I've shared over here about uh burger joint and for you and me it is very easy when we read these sentiments to understand what uh uh kind of a u sentiment this person has towards this business you know they can be uh negative uh or theyCan be really really positive but in some cases within the some of the feedback they have both the feelings you know they might be happy about something they might be uh not so happy about something else so we say you know that's kind of a neutral uh feedback now inToday's uh world uh we always notice that customers are taking race or sorry businesses are taking real action on these feedback if you go and look at a business reviews maybe on the Google Website you would see that businesses are responding to these feedback they want to keep theirCustomers uh happy and in this scenario where someone is providing a really really bad feedback maybe they would uh try to reach out to this customer offer them some kind of a free meal uh to make sure that their sentiment you know changes towards their business and theyDiscuss uh about their business in a positive term now you and me we can read this we can uh understand it but for businesses the the reality is that they are getting feedback from different sources and the amount of feedback that is coming is huge uh now imagine youNeed to do the same thing just multiply that those three feedbacks to thousand feedbacks how you can automate this whole process use AI to understand this sentiment and then uh take some action and today it's all about this before I run with uh demonstrating end to endLet's uh understand a few other things you know from Oracle perspective we are covering AI in many different ways we want to provide uh solutions to our Enterprises not in one manner uh all of us have been discussing wherever we go we read about AI what's oracles strategyOn that number one uh the world's most of the data is relying with Oracle right so it always starts from the data level U there are different aspects to it uh one of the aspect is uh you have data scientists they want to run their own machine learning uh algorithms they wantTo uh perform uh data science tasks of in in regards to that we are providing HPC we are providing data science Services we are Prov providing machine learning built into our database whether it's Oracle or MySQL WIA Heatwave if you want to run High uh and uh uh things weHave GPU infrastructure as well as part of oci so you can go and you can get the best uh GPU available to run your machine learning algorithms we all have been listening and hearing a lot about generative AI so uh in that respect Oracle has gone into an agreement whereOne of the leading generative AI companies called cair and they are focusing on um Enterprise use cases and Enterprise customers this is service is currently available as limited availability and you know in the cloud world you will find very soon it would be generally available so if you haveOricle Cloud infrastructure very soon you would be able to benefit with cair services or you can join in The Limited availability program and have these Services as part of your oci now besides providing all this we also are offering AIS services and applications uh whichIs on the right hand side now not every business have got the capability or resources to run to hire data scientist or to run these machine Lear Al algorithms so what we are doing for them we are providing AI services like digital assistant where you can have a chatbot running for your organizationsOricle AI speech which you can translate any kind of a conversation like this conversation we can record and it can transcribe that and can translate that into text vision service which can detect any kind of object so you you pass uh some uh images or you have a useCase where you want to let's say read the uh licensing plate from a car the vision service can detect that can read that and can uh provide you those details document understanding uh I have run a a similar session earlier you can go and look at the recording whatDocument understanding you can actually pass different Enterprise documents like invoices resets passports licenses it will read those documents it will extract the key values and will provide you those information and similarly anamal detection which is all about you have data coming from different sources and usually it it follows a certain pathBut then there are certain anomalies coming in certain scenario so you know you want to detect that today it's all about language so we will see how customers are utilizing oci AI language service to understand feedback sentiments and an analyze uh uh record so I will talk about that uhShortly apart from that we are also building AI features inside our applications you know whether it is Fusion application netsuite or cner all these applications very soon you would see uh a simple use case can be there is an HR person he wants to create a newDrop description he wants to uh has got certain parameters and it would provide uh those prompts those commands uh and get that job description generated for them right so from Oracle perspective we are providing AI we are building AI at the application Level we are providing different AI services for developers toBenefit we are partnering with goair to give you generative AI as well as we are giving you infrastructure when it comes to GPU and building AI within our database uh layer also now when it comes to oci language service there are different uh capabilities that this service does asThe name stands uh number one it will detect different languages so I think it can detect today around 70 different languages you can translate text from one language to another but at the same time it is more than that so if you pass some text it can uh recognize named andEntity so in this example if you see on the right there are certain things you are seeing color coded like Sky Train That's a product um uh Scotland that's a location so when you pass the text it is able to recognize those entities sentiment analysis looking at the textIt is able to analyze and say okay what's the sentiment of this sentence or what the person is experiencing is it a positive negative or neutral it is able to detect personal identification so I I come across a lot of customers with use cases where they say I have got thisData I want to run some analytics on that but I cannot do it because it has a lot of personal identification uh maybe credit card numbers maybe uh bank account numbers maybe date of births I want a service which can you know either remove all this mask this or uh replaceIt with something else so using the uh the pi identification capability you can pass that data and it will detect all the personal identification and as an output it can you know you can set the output it can remove that information it would mask it for you it detect keyPhrases as well and as well as text classification so when you pass a document to this it can classify and tell you what kind of a document it is you know a use case can be uh someone is trying to research uh writing a researchPaper and they are have got a lot of documents they can pass those documents to the service and it will classify that these documents are let's say about sports these documents are let's say about food and if their research is about say food so you know they have allThose documents in one bucket and you know they can move forward from there so within the service you have a lot of different uh capabilities from the sentiment analysis perspective which we will be focusing today um apart from the others uh what the sentiment analysis is it is doing itAspect based so if you look at a sentence if again on the right hand side if you look at the example and I just pick one sentence where it's saying the hotel is positioned for the sky train for exploring downtown with the usual Comforts now this is a sentence whereThe positive aspect and uh feeling is about the hotel so you can see what the service does it is able to identify that it picks up okay this sentence is talking about hotel in a positive manner so it is identifying that it is giving you a positive feedback to that with aCertain score so that is what aspect an sentiment analysis is all about as at the same level it is doing that at the sentence level as well so if you are passing the data you want uh the aspect analysis at the sentence level you can do that or at the entity level nowEnough talk about our services let's look at uh language AI service in real time I presume you can see my oci language console now so this is oricle cloud infrastructure console uh for going uh reaching this service all you need to do within our analytics and AI there areAll these AI services and number one is language right so you can go to language if you have oci uh oci tency uh you will have access to that first it is you know giving you a lot of information you know the things that I talked about languageDetection it talks about 75 languages text classification sentiment analysis key phrase named entity and Pi detection it gives you reference to different plogs uh live laes sdks apis now let's look at what this service does when I click on the text analytics here is some text let me put my own text hereSo you you I I want to pass this text and I want to this service to analyze you know and it is talking about uh uh some kind of a hotel and loyal customers number one thing you see from a results it detects okay this is written inEnglish since it is talking about hotels and food and you see it is mentioning about lunch menu many different times so the text classification it categorizes it as a food and uh restaurant now in the named entity you would see that it identifies different entities and theseAre the color coded just uh for our benefit so 10 years that's a duration lunch that's an event onion salad is a product restaurant is a location and facility so it is going in and it is identifying these uh entities for us it's able to extract these key uhPhrases for us as well so it's not only identifying it is extracting and I can utilize this for my next uh use case where I place these entities into a table I add the aspect to it and then I show the analysis now this particular uh feedback overall was negative so we seeThis feedback is you know it's gives me at different levels the document sentiment is negative aspect based if I read just one sentence we' have been going to this restaurant for over 10 years now and visit at least 2 3 months all I can say is that they do notAppreciate loyal customers now from a sentence like this it is able to decipher that it is talking about the restaurant and it is all negative so that's what the EXP based uh sentiment analysis is and then it is showing me at the sentence level as well right not all sentences were negativeNot all sentences were positive they are certain which were neutral so at the sentence by sentence it is showing me that then it is also giving me uh personal identi ification information so anywhere where it identifies something like date time something like uh although in this example I don't haveCredit card numbers credit card numbers or different things it will identify and will mask I have the option to configure mask so there are different options I can mask by saying okay make the last uh four things visible or last four thing isic replace it with something else orTotally remove it you know whatever option I select that would be the feedback over here so I said remove so that uh those information has been totally removed from the text right so this is how the interface looks like right but for us when we want toAutomate all this we are interested in calling this passing the data to this and getting the feedback so in that regard you can see on the right hand side there is the Json so I can see the Json and now I know when I call this APIHow I will be getting the data so if I look at the named entity I can see the key this is the offset where it is coming what is the type uh and what are the values right uh similarly here I'm seeing you know uh text is two threeTimes it's a quantity if I go down it text is per month it's quantity and so on right so what I will be using in My Demo I will be actually passing some customer feedback to this uh API of oci language service I will be reading thisInformation and then I will be taking some action right so when you see the sentiment again this is the text and the sentiment is negative now imagine you have feedbacks you are getting feedbacks you are getting feedbacks in real time or you are getting feedbacks all theFeedbacks have been stored in some kind of a file you can pass all those feedback to this uh API it will give you analyze the name entity it will give you the aspect positive negative and then you can store it somewhere and you can take action that's what we will see inThe next uh part the other um thing that it provides it provides you the translation service so I can have a text there are different languages that it supports and I can easily translate that and again there is an API available which I can utilize to translate any text from uh the languagesThat we support as source and to the languages that we support as Target and there is quite a b large set of languages including Arabic Czech Danish and so on okay now that was all about uh our uh core service when it comes to how we areGoing to read the uh sentiment from uh some kind of a text the thing is we want to automate this N2 and to to able to automate this we would be using Oracle uh integration cloud service now Oracle integration cloud service this is one of the leading iPassServices out there if you go and look at Gartner we are in the top right when in the leaders quadrant it provides multiple capabilities not only integrating but also process automation where you can build workflows visual Builder where you can have uh visual screens and mobile applications you canBuild using that B2B file transfer capabilities oci provides you a lot of services oci language is one service so we can easily integrate with Oracle as well as non-oracle services so it it it sits at the core of any of these use cases where we need to do theIntegration and what I would say is automation oricle integration Cloud provides a lot of adapters out of the box uh from Oracle applications to non-oracle and in this case I will be utilizing multiple adapters I'll be using the autonomous adapter to store the results I would be using the restAdapter to call my language service I would be using oci streaming adapter for realtime replication which is Oracles uh CFA based uh uh streaming uh Service as I said while when we as a person I can go and I can read uh the feedback and get the uh sentiment out ofIt from the language service I saw that I can go in and I can type the text and I can get that information but end of the day what businesses are trying to do they want to do automation end to end these sentiments if you look at theInest they are coming in different shapes and forms it might be coming as a feedback might be coming from an email a customer is sending an email because any business you can see there's a contact center and they would be providing they might be sending an email for uhComplaint or they might be happy uh you might have a lot of feedback in some application or in some kind of a file today if you look at social media whether it's uh WhatsApp Facebook Instagram uh Google reviews feedbacks are coming from all different shapes and angles what aricle integration Cloud canDo and simplify it can read it can connect to all these different Services it can read the data even if there are uh things are coming as a file attachment it can actually download those attachments and extract data from that then once the information is available to Oracle integration Cloud weCan call different AIS services in today's scenario we are talking about language for other scenario it could be something else we can combine different applic Services as well uh one use case can be I have got a lot of audio conversation about my uh call center IUse the oci speech service to convert that into text and then I pass that information to oci language service to get the um sentiment analysis so once we have the information we can call the language service and then we can put this information to our database and uhHave some analytics dashboards spiled uh on top of it so in today's demo what I will be showing you is a similar example I have a lot of customer feedbacks I have been provided in a in a file which I will be storing in our object storage service and I wantThis to be real time so within oci there is a service called events so as soon as a file is dropped in the object storage I can actually trigger an event which will send a message to our Kafka streaming service and Oracle integration cloud is waiting reading at that serviceAs soon as as a message comes it uh reads uh that message it downloads the file it sends it to oci language it gets all the entities it gets all the sentiments it down uh it stores all the feedbacks to our autonomous database where the analytics dashboards have beenCreated and you know you can as a uh customer success person can monitor those dashboards on a daily basis and say okay what is the sentiment if they need to take action they can take action so before uh further Ado let's look at this uh n2n demo okay now let's start with OracleIntegration Cloud uh within Oracle integration Cloud I have two Integrations built the first integration is actually as I said subscribing to the oci screen so here I have a trigger based integration which is if you look at the connection this is subscribing to one of the oci scam and what I'm doingIn this scenario is this is the oci stream that I have I am uh getting that information and uh I have a simple message uh there as well what message I'm ready to read from that stream once I configure that all I'm doing is as soon as I know that there's a messageWaiting for me I call my child integration which is going to call the language service so if I just show you the mapper here you see that the message that I receive on the left hand side I am receiving the request wrapper I'm receiving receiving the content and asIt is an event I'm receiving different data data telling me about the file that I have uploaded into which compartment into which bucket and so on and I'm passing that to my child integration as event ID and all that data information now I'm calling this child integration which is the integration numberTwo where all the business logic is and in this integration once I receive that file the first thing I do I am actually downloading that file from object storage so here you see I'm saying read that CSV file and I'm actually downloading that CSV file into our own inmemory which is the stageFile then I'm going ahead and I'm performing different tasks you see I'm doing it in a loop because within that file I don't know how many uh Records might be there so I'll run a Loop and first thing is I'm actually saving the records because I'm getting the customerFeedback inside one of the tables of my database so here uh is my invoke save reviews to the database and I can show you our database uh adapter very quickly what the database adapter uh supports it supports different operations in this case it is insert IGo in I select the table which is the name of the table is feedback and I basically going to store the results into this feedback table and again using our mapper the information that I'm getting from the file all the records one by one I am inserting it inside theDatabase table so that is simple mapping and you can see you know I'm performing uh one to one mapping there next step I've got two more tables one table I want to detect all the entities and then the other table I want to detect all the sentiments so in thisCase if I open this scope I am calling the language API so what I showed you from the front end what I can do I can perform all those tars as part of uh the API so here I'm calling the detect language entities I am going to pass the textThis is my sample because that detect API just requires the text okay give me the text that you want me to detect all the entities and in the response I will be receiving a response something like that this is the sample all the entities inAn array it tells me the offset so that tells me within that text where this information is what is the length of that text what is the type is it a person date time or something else right and the score so that's uh what I would do so again WithinOne sentence or within one text there can be multiple entities that's why you see I'm again going in a loop and I'm now storing the text that I have passed inside another table where I am storing the entities so that's one table and in a very similar manner what I'm doingHere is calling I want to detect the sentiments right so again calling the API this time I'm asking it to detect language sentiments in the request again it just requires me to pass the text so I pass that text and in the response this time I would be receiving a response which isThe aspect analysis so again it gives me the offset it gives me the text it gives me sentiment whether it's positive whether it's negative whether it's neutral and then the score of that positive negative as well as neutral this is the information which I would be then storing inside my databaseSo I can quickly show you from a mapping perspective that once I receive the sentiments I am actually storing it in one of my tables which is here right so I've got the feedback ID aspects sentiment and so on so this is what my integration is doing as part of Oracle integrationCloud so you know very easy very quickly I can put together this integration and here you are seeing I'm using some connections which are connections to my streaming API service connection to my oci language service the database and rest trigger now okay let's lookAt where we will start so now I have oci within oci we have object storage so if you remember in my Uh yeah in my architecture I've shown you the language side I've shown you uh the integration the Integrations are built now look at the uh pieces which put all this architecture together so we are going to object storage within object storage we have buckets where weCan store our files so I have got a bucket which is called HQ customer feedback and within our uh oci Services many different Services they can trigger an event so if you notice over here I have emit object events enabled right and what I have enabled over here is ifI go to the event service which is within our observability and management I can very easily create an event and I'll show you what this event is going to do before I you know download a file there is this so I have an event uh event and within the event the rules areEvent type is object create so anytime an object is created ated or uploaded into that bucket this event would be triggered it have attributes compartment name and bucket name are the attributes that it captured uh and this is what this event is doing from a triggerPerspective if I do an edit rule you can actually see for example the example over here and this is the Json object which I would be receiving and you can see over here I can read the file name I can read the compartment name and these are allThe information that I need in my integration to download the file and then there is an action there are different actions that I can take from this event once the event is triggered I am actually sending it to an oci streaming topic which is our uh oci streaming which is our cfus serviceTopic and orle integration Cloud as part of my first integration has subscribed to that topic so I drop the file it fires event it send that message to the oci streaming and OIC now has got that event captured and it takes it uh forward from there that's not the onlyThing that I can do if I wanted I could have sent a notification maybe I sent that as an email or I can call an oci function uh which is our serverless service to toose take some action also so I have a streaming service uh this isOur uh streaming service where I've got this topic created uh and this is the topic which is just stores HQ customer uh feedback stream and you know the message comes over here which article integration has subscribed to now let's trigger this uh whole integration so I'm going back to theBucket so I have this customer feedback bucket and I want to upload a file so I'm just selecting one of my files which have got a lot of customer feedback and that's the hotel reviews. CSV file and I will upload this right so it's a small file whether around 15 20Records uh but it can be more but for the demo purpose I have got you know 15 20 records there so that's been uploaded now as soon as this file has been uploaded I expect that the event has fired and there is a streaming uh it reached to streaming and within OC OICNow I'm coming back to my Oracle integration Cloud I go to observe which is our monitoring capability and H here what you see it has already received a message on HQ so you can see it's just now so on the HQ subscribe screen customer feedback I've received a message rightSo I just uploaded a file and because it filed an event it send a message to oci streaming and I can see within the monitoring that my first integration is already triggered and I can see it is all successful and it is called the child integration in The Wire message in our activityScreen I can see this is the hotel reviews. CSV file that I had uploaded and all the other details are available there as well right so so far so good as soon as I upload a file I get uh this has been triggered and let's look at what's happening with the otherIntegration so it triggered the other integration and that has also processed successfully so if I I go to the other integration I can see the resource name Hotel reviews. CSV came and this integration as I was uh talking that you know the record has around that the file had around 14 records itDownloaded the file it then one by one inserted that information into three tables one was uh entities the other is sentiment and then the third was uh feedback right so I can see all this has been uh completed successfully now this is my Oracle uh autonomous database and here here areThe tables that I have right so when I look for records inside the feedback table I think my session might have F but let's see yeah so these are all the records which came from that file right so I have got these 14 records I've got different uh Fields ID feedback dateFeedback and location similarly I've got another table sentiment and I've captured the sentiment and I've uh stored that in the sentiment table so what happens in the sentiment for the named entities it will give me the score uh whether the sentiment is positive whether the sentiment is negative andThis is all coming from uh the API that I showed you in my integration so here you can see the aspect about view is positive for this feedback ID server positive and so on if there is any negative feedback which I think is here for crib there was a negative feedbackRight so we have captured the aspect we have captured the entities we have captured the feedback now on top of this we have built uh an analytics dashboard now this uh dashboard is actually capturing the data from uh the database tables that I showed you so it's just asking me toRefresh and this is all now is real time right I did not have any records in my database table and you know know I ran this and I can uh have uh my dashboard sentiments uh available also let it just load okay and this is what you know theIt's a very simple dashboard that I have created I am uh counting all the aspects and sentiments you can see by the size of the text that you know for employees for hotel there's a lot of positive feedback as well as there is negative feedback so I can uh compare uhThem so here you can see I can compare you know there are uh around 40 uh negative uh aspect Fe feedbacks uh about hotel while as there are 120 uh for positive uh similarly I can see about room there are certain negative feedback I can read the actual text related toThat uh particular feedback over here okay these were the two uh requests or or two feedbacks which were uh had room aspect has got a little bit negative I am counting the different entities here as well so this is just a very you know uh basic dashboard that I have but youGet the context right so with the feedbacks that I'm receiving I am not only able to uh understand the aspect I can actually automate the whole end to end process I can have a dashboard for my customer success people to monitor this and I can actually take someActions as well with the help of process there I might uh trigger some events which uh I want to reach out to those customers and send them an automated email or maybe send them some um as I said some coupons and things like that so all these things can be done in realTime so from the demo perspective this is uh the overall what I wanted to cover uh on uh if there are any questions or there is any specific thing about this demo that you would like me to show you please uh let me know okay great thank you so much uh HarrisAnd thank you so much everyone um for joining today we greatly appreciate appreciate you taking the time to join us um and thanks to Harris and thanks to Juan caros for presentation demo and Q&A so we have received some questions um so far thank you for those and uh pleaseRaise any remaining questions in the Q&A box and while you're doing that I'm going going to pop up a conclusion poll here so we have three questions in this poll so make sure to scroll down to see them all and I'm going to uh let thisPoll run in the background uh for a little while uh so that Harris and Juan Carlos can go over uh any the Q&A questions we have received so far so Harris and Juan Carlos you can decide um how you want to cover them but you canCover from from the list of those have been answered so far and if you could repeat the question as well along with the answer um so I like the uh thank you for responding to the poll and I'm going to let it uh run for a little bit longer than conclud it soUm yeah I can see there's a question from Mr Hen uh can the same applied using audio recording files yes absolutely so as I mentioned we have different AI Services uh the oci AI uh Speech service it can actually take any kind of an audio and it can uh convertThat into text so as part of the automation process we will actually first convert the audio into text and then we will pass it to the language because language uh uh input only is text and using those two Services we can uh perform similar task uh to audio recording files as wellUh there's another question is this feedback getting processed using Oracle AI uh definitely and I expect uh all green and positive okay let me look at some of the questions ju if there's any question you would like to pick please let me know I can um I can repeat it repeatSome of them um so one one of those is Cent language service read from voice videos and images so basically um I mean the O language service operates over text so what you can do is for um voice and videos you can use o oci um speech HTo extract text from audio or video files and then sending that to oia language to do text analysis whatever you want sentiment whatever and uh for images you can use oci Vision to extract metadata from those images and then passing that to O language to do whatever you want like name entityRecognition or something like that um um someone was asking for reference link to an existing um example practical example to to like an step on step step by step um I posted some u a link of an step by step to do real time an event driven Twitter sentiment analysis usingOci AA language and Oracle integration haris I don't know if you have another link that we can share as well yeah so there are if you go to Oracle live Labs there are individual live labs for oci language Serv there a lot of live labs for Oracle integration Cloud the linkThat Hanan has shared with you uh that basically combines both of them uh but do check in live Labs we are always adding more and more live Labs uh I'm sure one of these uh demos that we are showing would appear in those live Labs sometime soon um there's another one reallyInteresting um can we convert these sentiments in terms of a star ratings as well like one star to five stars so basically um one one thing that you can do is you to leverage another capab ility that we have within or integration Cloud that is called dmn rules so youCan have like a decision table where you can have like um based on the score that this oci language sentiment analysis provides you that is from zero to um uh I mean from zero to one sorry um what you can do is basically like um depending on the ranges of that thatScore H you can convert that from you know to an A Range let's say from one to five and the cool thing here is that when you define these dmn rules within Oracle integration Cloud process automation H those rules can be exposed or let's say when you publish orActivate them they are exposed as rest services so then you can use uh or Oracle integration Cloud to call this service after you apply sentiment analysis and as haris show we can do bat sentiment analysis because the API can receive multiple texts let's say so umThis dmn rule or this decision rule can also receive an array of text or scor sorry in this case and then you can have like an array a result array with the let's say um a score from one to five so that will be let's say the approach thatUh you could um follow to uh um implement or um support and cover these um these kind of uh transformation or star ratings right yeah one we have just received a new question so I'll just quickly answer that so can this be integrated with third party survey toolsUh yes uh absolutely there are different ways from a use case perspective you can integrate it with third party tools first thing from an oracle integration Cloud it can connect with Oracle non-oracle applications any apis so from a third party tool if you have receivedUh a lot of responses and you have that information available either in text or there is an API where where you can download all the uh feedback you can then pass that to the language service and you can easily integrate that matter uh I understand these days uh when weAre doing uh third party service while there would be certain questions which are you know multiple choice or single choice but we also provide a lot of text for people to provide their feedback and I'm I'm assuming this question is about okay if there is a text in those surveysHow can we get that text so from those third party tools like monkey survey and others they provide apis you can call those apis receive that uh feedback and then the rest is very very similar okay and see Mark Mar has just provided a link to one of the live Labs uh soPlease uh do uh look at that okay so if there is no more question you can continue asking I'll just quickly do a wrap up so from our perspective today what we covered we in the demo we showed two Integrations one integration was reading the data from uh oci streamingService uh where uh and the other was actually downloading the file from object storage get passing all the text into uh to our oci language servers getting the sentiments getting the entities storing it inside the database for this demo purpose I used a file which had some uh feedback but that'sJust for this demonstration uh as we discussed uh we have other use cases Juan Carlos has already shared we can actually get data from a Twitter live feed we can get data from uh Google reviews for example we can get data from WhatsApp and so on or the third partyServs as the question came so once we have the text in this demo we were actually uploading the file to a uh bucket which is our object storage service we were triggering an event because we wanted to do it real time that event was passing the informationTo our oci uh streaming service and once that service was there OIC reads it and finally it was storing all this feedback entities and sentiment inside our autonomous database service and we had built a analytics dashboard which was showing uh all the information about the aspect the uh the positive negativeNeutral uh feedback that you know people are providing about uh customer feedback uh I hope you enjoyed uh the session uh if you have any question please ask now later on you can reach out to us as well you can easily find myself Juan caros onLinkedIn uh and you know if you need you can you know get our email addresses and details too thank you thank you so much everyone for joining uh so feel free to uh drop off uh from the webinar as you like thank you everyone have a good Day