hello everyone welcome welcome to day one of reinvent thank you for being here we are super excited to be here with you today we're going to be talking to you about customer technology Journeys and different personas and how we work with them here at AWS so first I want to introduce us my name is Teresa Meyer I lead a team of solution Architects that are focused on customers new to AWS and helping them get started and really build their foundations with AWS I'm joined today by Eric espe who is another leader here focused on similar things Eric himself was a customer and so he's got a unique perspective that he can bring in working with our customers as they migrate to AWS uh next is sorab annotti and sorab is uh a leader here at AWS with over 18 years experience working in migrations and helping customers with their journey to aw us and last but not least is Jay Vale Jay is uh one of our experts in data and analytics and has similar experience so we're super excited to be here today I'm going to talk through uh the flow that we're going to go through in the session we're going to introduce you to our core personas and share some of their pain points next we'll take take you through how we uh work with our customers to address those pain points um uh in their journey we'll analyze some best practices for those pain points share some of our real customer use cases related to those pain points and then finally talk through some architectural patterns and guidances that we uh work with our customers on so first let's introduce you to our core personas we've got four core personas that we're going to talk about today the first is who we affectionately like to call Data Center Dave uh so as his name applies data center Dave is an expert in all things data centers he has uh established his career on building and maintaining data centers but Dave is uh customer similar to Dave has realized that the backup and Recovery requirements and uptime requirements that his company is asking him to maintain he needs a little help and so how we usually begin working with customers of this Persona is in that backup and Recovery space so customers similar to Dave are really looking for how can they ensure greater uptime and um look for that data Disaster Recovery strategy in 20125 next we have energetic Elsa Elsa and her team have some experience with AWS uh they have mostly engaged with AWS in that data uh Center disaster recovery backup recovery scenario but Elsa understands that there's more that AWS can offer her and her team and her company and so they're really looking to take the next steps in migrating some of their workloads to AWS really focusing on cost optimization as well as those those requirements around data governance and um regulatory requirements that her company is asking her to face and so we'll go through some of the strategies that we work with are are Elsa energetic Elsa personas on how we guide them into those first steps into migrating to AWS next we've got strategic Ian so Ian is focused on growth so he him and his company have already migrated a significant number of workloads to AWS and so they're very familiar with the AWS platform and the migration uh path that we have set out for them mostly what they've done though has been in a lift and shift capacity and so customers like Ian we really engage with them to help them focus on how do they leverage those Cloud native applications to really begin to take advantage of cost optimization and the power that AWS has to offer their company and so we we typically work with uh customers in the Strategic Ian Persona um around that space of migrating to Cloud native Solutions but then we also work with their lines of business to really really help them understand what are some new opportunities that their company can begin to explore while leveraging AWS Technologies and then last but not least we've got AWS expert Emma so as her name uh implies Emma's company and Emma herself is very familiar with AWS and in in some cases these companies are even born in the cloud like we used to say and so typically the way that we engage with customers that are in the expert Emma c u Persona is around how do we help them with their data strategy and their gen Ai and genbi strategies really focused on how can we help them develop new technologies that give them a competitive advantage over their their customer or their com competitors and so we'll definitely be walking through today what we have to offer in the geni and genba Spa bi space for our customers uh helping them to to get that competitive Advantage so before I hand it over to Eric uh to talk through um how we actually begin to work with customers in these phases I'd like to take a poll it really helps us here on stage know who's in the audience and and helps us tailor our conversation for you so just by show of hands how many of you in the audience can really relate to the data center Dave Persona and you know his cowboy hat here excellent thank you what about Elsa energetic Elsa anybody already worked with AWS a little bit and is really more focused on migrations we've got a couple perfect strategic Ian you've already migrated a bunch of things to AWS you feel like you're really getting your feet under you and now it's time to move to more Cloud native excellent and then any experts in here that would classify themselves as AWS expert Emma's less so but that's that's the whole point there's still quite a few so thank you uh really appreciate that like I said I'm going to turn it over to Eric now to walk through how we actually begin working with these personas all right thank you Teresa so um we're going to start with uh data center Dave and uh as you may recall Dave is relatively new uh to the cloud so he's going to have a lot of questions and the AWS account team is going to be able to help Dave understand what the cloud can do for him start understanding how to uh anticip P cost in the cloud and um you know also start to figure out how to uh help Dave and his team become more familiar with the cloud the first thing that the AWS uh would help Dave with is to get a understanding of what that Eco system on premise would look like so that we can help uh start developing a disaster recovery uh scenario uh that would help Dave meet his slas and slos and uh make sure that there is a high confidence that AWS could solve that problem for for Dave they would then go ahead and and execute that solution um for Elsa Elsa's already had a successful Dr Solution on AWS um and she's much more interested in making sure that she is setting up her AWS environment according to Industry best practices as well as um making sure that uh they're they're meeting her compliance and governance objectives um security is going to be really important for Elsa as well at this stage um and uh she would work with the account team to start understanding what does her current environment look like and start designing what her new environment would look like so uh part of that is going to be education and this is where the AWS account team could help Elsa very uh significantly through the cloud Foundation accelerator which is a series of modules that talk about the different aspects that go into the technical components that put a good foundation in place to grow on AWS we would also help Elsa understand how to set up a multiac account strategy uh anchored on control tower and Landing zones and create a plan for her to go ahead and do that uh to uh put those pieces in place and then uh we' make sure we would work with Elsa to uh implement the control tower and Landing Zone Ian is a little further along he's already got confidence in his AWS environment he's got governance he's got compliance he has high degree of confidence and his security he's much more interested in figuring out how to uh leverage strategic workloads in the cloud how to modernize those applications so it's really important that uh we work with Ian to help him understand what are the applications that the line of business is using and where are the pain points that those line of business business owners are experiencing so that we can help uh figure out how to solve those problems um a big part of this would be executing uh what we refer to as an OA or an optimization and Licensing uh assessment this takes a look at the ecosystem that's running understands what the utilization of those assets are as well as the licenses associated with that so that we can help in make uh informed decisions around modernizing his uh applications we would then take the output of that Ola and present a strategy and a plan uh on how to implement uh a modernization strategy within AWS so um and we would prioritize that for the most impact moving on to Emma so Emma is done all of the the heavy lifting to get AWS up and running she has control tower in place she has um uh governance compliance security has her strategic applications running in AWS but Emma is really interested in innovating and this is something that AWS can really help Emma with because we have uh experience dealing with customers from a variety of different uh Industries so we can help Emma understand who are her customers what are the things that they're important to them what are their pain points and we can then work to help em uh identify Innovation opportunities uh deep diving into those customers uh what are the current market trends or what are other Industries doing that might be applicable we would help her by working on a digital Innovation session uh a working backwards session uh we could uh perform an executive briefing where uh we would bring in AWS experts to help understand what's going on in the industry we could do a culture of innovation uh session and coming out of that we would work with Emma to identify three workloads that would be good candidates and opportunities for Innovation and uh then we would uh work with Emma to build a business case to present to the seite and start innovating so with that let's dive into the architectural patterns and guidance associated with the different personas and for that I will turn it over to sorab thank you thank you er um all right so based on the uh uh personas and the customer Journeys that Eric and Thea just talked about we have categorized the workload into four different categories the first category is what we are calling as starter workloads and this resonates with the Persona of data center Dave and here we will talk about backup and restore and Disaster Recovery the second category is technical foundations and this resonates with energetic Elsa and we are going to talk about multi- account strategy identity control tower and networking in this the third category is strategic workloads and here we will talk about lift and optimize shift the surus workloads and contact center Solution on AWS and the and the fourth category resonates with expert Emma who is looking to do the Innovation for the company and this uh uh and in this category we will dive into the workload such as the data analytics generative Ai and bi so starting off with the first category which is the starter workload which resonates with the Persona of data center Dave so Dave wants to implement a disaster recovery strategy for his organization so from the approaches standpoint there are four approaches that uh we have for the Dr backup and restore pilot light warm standby and active active all these strategies are based on their RPO and RTO objectives RPO stands for Recovery Point objective which means that how much of the data can you afford to lose and RTO stands for recovery time objective which means that how much of a downtime can you afford So based on that the backup and restore is the most common one in which you backup your existing applications and servers to a vault and and you can recover whenever uh you need to the second strategy which is the pilot light in this strategy you do a continuous replication of your data to AWS Cloud but the application servers are not running yet only the continuous replication of the data is happening which is why it has the RP of minutes and RT of hours for the warm standby you do the continuous application of the data as well as the application servers are running however they are not running in a fully scaled up mode uh again which is why it has RPO and RT of minutes now with the fourth strategy which is active active it has an RP of seconds and RT of near real time and this is when you're serving your workloads from multi- of AWS the traffic is being actively served from multiple regions so looking at this Disaster Recovery Spectrum when it comes to the backup and restore from on premises to AWS so towards the left hand uh of the slide you see the on premises and towards the right we have a cloud so data center Dave wants to back up the applications and servers from on on premis to Edis cloud and the question that Dave asked is do I have an existing backup software that has the AWS Cloud native support and if the answer is yes to that question then Dave can use that existing backup software to backup the data on cloud but if the answer is no or for some reason they decide not to CH uh go with the backup existing backup software then Dave can utilize AWS storage gateway to backup the data on AWS Cloud a storage Gateway provides you that hybrid cloud data storage solution it also provides the local C for low latency access it is compatible with many types such as file volume and tape and it also integrates with a lot of storage services such as Amazon S3 Amazon S3 Glacier and EBS so now Dave has completed the backup and restore but Dave is even more excited and he is looking for more and more he uh so now Dave is so now Dave wants the RPO of active active RTO of warm standby but at the cost of pilot light and this is where AWS elastic Disaster Recovery or we call it uh a DRS service comes to risk here so so AWS DRS service will monitor your primary environment for any failures and it will provide you an automated Dr option and in the next slide we're going to talk about how the high LEL architecture diagram or how does the DRS works so in this high level architecture diagram when you start implementing DRS the first thing that you do is you install the AWS replication agent on the source servers installation of these replication agent does not require any reboot of the servers or or it does not impact the performance and once these replication agents are running on the source servers they start continuously replicating the Block Level data into AWS Cloud the data that they replicate goes to the staging area subnet which is automatically spun up by the DRS for you and once uh yeah and the staging area subnet is designed in such a way that it provides an affordable storage as well as compute solution it minimizes the compute to just maintain that continuous Block Level application but at the same time uh it it compresses the storage to provide you an affordable solution and with a click on the DRS console you can now recover your instances into the recovery subnet uh and and this process typically takes just few minutes so um out of many customers that use storage kateway and elastic Disaster Recovery these are uh the two references that we have Kelloggs which is an AWS customer they were able to reduce their backup cost by 90% And they eliminated the traditional backup solution using the a storage Gateway on the other hand uh Tyler Technologies uh used AWS elastic disaster recovery and they were able to recover their mission critical work workloads 12 times faster compared to what they used to have so now we have helped Dave Implement a backup and restore and a Dr strategy and before Dave can do more uh energetic Elsa comes in and she wants to ensure that before they scale out the cloud adoption we are having a solid foundation from the technical standpoint so before we dive into the technical foundations let's understand why M account strategy is important when you're new to AWS uh so imagine that you have a house with many rooms and all the stuff that you have you just plan to put in one house in the beginning it might seem convenient but over a period of time as you start as you start buying more stuff accumulating more stuff it will be hard for you to organize that and keep it secure so similarly when you have the single AWS account it results in the gray boundaries complicated per Mission sets and there is often a resource Collision which is why a multiac account approach is is beneficial it's more effective it is more effective to manage your AWS infrastructure you can group The workloads based on their business purpose and ownership so much like that big house a bedroom is for sleeping a kitchen is for the food and so on right so you can group your accounts based on your on the business purpose and the ownership a multi account structure also promotes The Innovation and Agility uh because it allows your teams to independently work uh and increase their speed from the cost management standpoint as well a multi- account structure is very effective because now you can measure the cost of your workloads and your environments you can segregate them out so knowing that Elsa wants to implement the multi- account strategy the first step that she does is AWS control tower AWS control tower can help you spin up a well architected multiac account environment in less than 30 minutes it brings it has 400 plus controls out of the box that you can apply to improve your security and compliance posture in order to spin up the control tower we we recommend to start with a management account so you start with a management account and you set up the control tower over there with just four simple steps and as you are in the process of setting up the control tower under the hood it uses the AWS organization and and there is a foundational organization unit that is created in this scenario we are calling it as security OU but within that security OU there are two AWS accounts it spin up one is the log archive account and and the second one is the audit account the whole purpose of spinning up these two accounts is that we are going to centralize the logs from all the different child accounts into one place and give access to the security and the audit so that they can do their analysis on these accounts now once this Baseline control tower architecture is complete now Elsa can bring in that existing disaster recovery account into the control to governance and uh and in order to do that Elsa uses an account Factory feature of control tower and so using account Factory you can spin up new accounts or you can also enroll an existing account and bring it under the same uh governance model of the control tower and finally uh we need the access for the users and AWS IM identity Center it used to be called as SSO earlier provides that single sign on access to all the child accounts you don't have to really provision the IM users in every single child account you can just Pro uh so you can just provision the users in just the management account and from there uh you can grant access to all the child accounts AWS IM identity Center also integrates with the external identity providers such as OCTA Azure and so on so now Elsa has set up the multi- account structure and now she wants to ensure that the networking architecture is Baseline so on one side we have the corporate data center on the other side we have the existing Dr account what we generally recommend is to to set up a dedicated network account for the organization ation so this dedicated Network account will act as a centralized Hub and then it will have you and then it will have further spokes right so you can deploy a Transit Gateway on this dedicated network account and connect that to your corpor data center uh using a VPN or AWS Direct Connect and then connect the spoke back to the VPC of the of the disaster recovery account and like that you can keep on adding your rest of the accounts appd accounts the data itics accounts and so on uh and finally you can also have an inspection VPC and a firewall running over there which can inspect the inbound as well as outbound traffic so up until now we talked about uh Dave we addressed Dave's need we addressed Elsa's need um as well and now we are going to talk about the workload that strategicon cares about so so one of the objective that Eric talked about for Ian is to minimize the technical debt and and in order to do that uh there is a lift and optimized shift with AWS that we recommend right uh So based on the customer uh like Ian we have seen some challenges across uh often customer who are doing the lift and shift to AWS of their data center they run into challenges of schedule overruns the the project can run upwards of of more than 18 months the customers are looking to find the best guidance and find the best experts uh in the field and sometimes they need a place a centralized place to collaborate with their internal teams as well as external partners and that's where AWS migration Hub really helps AWS migration Hub helps you plan track and assess the migrations to AWS it really acts as your comprehensive assistant to uh in your migration Journey so let's understand how does AWS migration Hub works the first service that we have is AWS application Discovery service so um so consider that now you're building up on the same house example uh now we are just moving the house and you have stuff scattered all over the place into the multiple storage units into your friend's place into other house and so on right so AWS application Discovery service is really like the I mean if you want to hire a a bunch of explorers who are going to go out there look at your assets and inventory and make a list of them uh it is essentially what AWS application Discovery service do when you uh use the AWS application Discovery service it starts maintaining the assets and the inventory of those assets for you on your behalf and once you have those assets inventory then AWS migration Hub applications will give you a centralized dashboard where you can view the the applications and its dependencies as well and once you have these two then comes the a migration Hub recommendation so AWS migration Hub recommendation is really like that professionals uh mover and Packer who are going to pack the stuff and move the things for you right and they will provide you the best recommendation on how to do it so using AWS migration Hub recommendation you will get the recommendations on what applications are a good use case for repl platform what applications are good use case of rehost and uh and so on and finally moving a house is a very complex and cumbersome process and you need a detailed plan and a checklist and in order to do that we have AWS migration Hub Journeys which will give you a detailed plan and a checklist and help you give that seamless migration so this slide really puts it all together on this slide as we uh as we classify the migration Journey it is four phases assess mobilize migrate and modernize and on this slide what you see is the AWS provided migration tools migration Hub tools and then also the partner provided tools uh and all you have to do is pick up a template use one of these tools in each of the phases and then uh it will help you out with your migration Journey so foran we have minimized the technical that we have helped Aon to move the things uh and and the data center and the servers to the cloud and now aan is asking for even more right so the first workload that we going to talk about of this category we will talk about is the seress workloads on AWS but before we dive deeper into into serverless your workloads we will talk about why serverless prior to joining AWS I I used to have a development background and and if if I remember those times uh if we have to build the application we used to rack and stack the servers on data center we used to uh do the security patching or Baseline the server and then on top of that we used to install the relevant software then came the cluster management headache then came the automatic uh scaling headache right with serverless all that pain kind of goes away so you don't have to provision any infrastructure there is no automatic scaling required for you to build and develop because it automatically scales and up and down you pay for the value and it is highly available as well as secure so if we look at this table right here uh if you use well as services to build your workloads then a lot of responsibility is is taken care by AWS and it reduces the operational overhead for the customer primarily customers have to focus on their business logic as well as just the apis so if I have to classify the serverless application essentially you can classify them into two broad categories the first one is the API driven use cases and the second one is the event driven architecture for the API driven use cases Ian is looking to build a mobile application and when Ian wants to build a mobile application on the uh the client that you see is the front end of the mobile application and the way client communicates to AWS cloud is via the apis and we have Amazon API Gateway which is a fully managed API Gateway service by by AWS it acts as the front door to those front end uh clients so when the request let's say that the the the client wants to raise a request to the back end that get me the data from the contact table so when client sends a DAT uh request HTTP request to Amazon API Gateway that request is then translated into the event payload and then passed over to Amazon Lambda uh function so AWS Lambda is another serverless service where uh you essentially write your business logic and code and when ad Lambda receives that request it sends the request to the Amazon Dynamo DV table from where the data is read and then the response is processed further by Lambda and then further passed on back to client so this is one architecture of a microservice and like that you can scale this architecture out and have multiple Lambda functions to create a a distributed microservices architecture the second one is the event driven architecture so Ian has a use case in which Ian is looking to process the files U and these files are images and videos so in the event driven architecture it responds to the events so when the file upload is happening in the Amazon S3 bucket there is a trigger in the Amazon S3 bucket which is configured and that trigger invokes a a Lambda function from AWS Lambda function it passes on the object URL of that Amazon S3 bucket to a service of of of AWS which is called Amazon recognition that uh is used to process the images as well as videos or do the analysis of those and once the analysis results are sent back to the Lambda function then those results are further stored in the Amazon Dynamo TP so if you notice in this architecture things are happening in the reactive manner it is it is even driven so um Ian has um now built the seress workloads as well and out of hundreds of thousands of the customer that currently are building their applications and workloads using serverless these are some of the common names that you might have heard of about and and this is a famous quote from our um CEO that if amazon.com had to start over today we will build it on serverless with that I'm going to pass the mic to my friend J and he will talk about the next step for strategic thank you s before we dive into this uh Amazon connect Solution please raise your hands if you heard about Amazon connect or you're using Amazon connect already W that's a lot awesome so this is one of the Strategic workloads I'm going to dive deep into also I'm going to dive deep into some of the Innovative workloads as well so before we dive into the Amazon connect let's see why we build this solution and why do we need Amazon connect so we listen to our customers and we observed from the existing cont Solutions there are some pain points from the customers perspective they basically said that there's an inconsistency and there's a lot of repetitive experience across these Solutions meaning that you call them you give some idea but then when you chat you to give the same information again and again people are feeling not very frustrated that's one then from the agents perspective Contex agents perspective they have many disjointed applications like for an example they might have CRM solution and then they have the Contex solution now they don't talk with you each other now they have a hard time like connecting those dots from the administrators or the supervisor perspective again they want to connect the data but now because of these disjointed applications all this data is disconnected limited and incomplete so they cannot make much sense out of it that's where Amazon connect comes in so what is Amazon connect again Amazon connect is a very simple self- serving UI you can configure in your own in your own organization and then under the hood oh in fact it delivers a dynamic and personal automated customer experience which is going to you which is going to help your agents life so much also the customer's life plus under the hood it bu Builds on AI model why is it important because it's going to learn over time what customers want what agents want So based on that it can make some recommendations on top of all these things it also provides some basic native analytics and some dashboard functionality so you can use that and understand how the context and the solution is working so far so this is a sample architecture looks like if you want to take a picture of slides please feel feel free again here what happens is customers make a call they can make a call through their cell phone mobile or they can chat through their computers and when they chat it's going to connect the Amazon connect first but then when it hits Amazon connect service it's going to called the Amazon connect connect flow what it does is it's an basically an orchestration service it's going to call multiples AI Services under the hood also the Lambda function but s of mentioned multiple times so what why why do you need all these Services imagine when a call we want to convert the voice to text so we have some AI services like for example Amazon poly that can do that imagine you're texting or typing you want to convert the text to voice that's where the Amazon Le comes in now if you call you want to understand what this customer really looking for what's their customer lifetime value looks like you want to find those things for that what will happen is the Lambda function under the hood again this same oration service can call the Lambda function it can go to a data store and pull the preferences customer preferences agent preferences so we can make a like valuable decisions based on that imagine we have done with all those things what is next step we want to take the C script and save that into the S3 bucket why do you need to St save it because when you save it there like Sor mentioned earlier maybe the event driven architecture or some other architectures we should be able to do some realtime analytics call analytics plus this is where the cool things going to happen imagine you have all this data available now you want to make some natural language questions like for an example last year which customer said statement did had most drop calls we want to find that you want to find that natural language question you can ask that using Amazon connect for Q again Amazon Q for connect I'm sorry again this is a service um we I'm going to dive deep into um Amazon Q very soon this is a generative a service it's implemented on top of Amazon connect to provide this ex extra experience or the best experience so these are some of the customers again 10 minutes tens of thousands of customers using Amazon connect and 10 million contact centers um 10 million contact center interactions per day it's happening so you all can use that as well talk with the S let's dive into Innovative workloads from Emma's perspective I know there are so many Ms here you raise your hands and um specifically ter I mentioned about generative AI generative bi I'm going to dive into all those things but before we dive into any of those things what is the most important asset is data why because data is growing exponentially and it's coming from many different sources also data has many different dimensions like for an example Vector data scalar data or um structured unstructured all kind of dimensions and then the data is used by multiple stakeholders customers or board of directors everybody want to know everybody want to know about data right so how do you manage the data that's where the data lake or Lakehouse architecture comes in what what it is basically in the center of the lake house architecture we have S3 simple storage service it has level n durability it can scale unlimited almost unlimited also very cost effective solution so from these all different sources we can bring the data into the S3 with our purpose built services and then we can build some governance controls around it Eric was talking about building some governance control how do we do that again we can tell who who can access what data at the row level or even cell level so how do you do that let's di an architecture so now you can see some of the data sources here like these are like a SQL no SQL and then some of the mobile data stores and the streaming data flat files all these data are coming in now first we need to ingest the data in that's where the purpose build in Services coming in like for an example if you want to bring the data from your relational databases you can use um database migration service what it does you can basically tell like hey dump all the data into the S3 or do it on a schedule basis or ongoing basis you can set your own schedule plus if you have streaming data like for an example you're getting some tweets data or something you want to get that that's where the streaming services comes in like Amazon kesa services or manage CFA or if you have some flat files I used to work for a Healthcare company we have this big big flat files we transfer for that we can use a just transfer family so with this all these Services we can bring the data in and put them into S3 bucket what is next we need to scoll through the data and see the structure so for that again we'll scoll through the data find the structure once we find the structure we need to save this logical structure somewhere that's where the AWS glue comes in it's going to do all this work for you and create a logical structure so you can ask questions plus imagine you want to bring some external data sources and you want to merge that data with the current data and ask some intelligent questions that's where the glue data Brew comes in so now you have all this data available you did The Logical logical structure now you want to do some analytics you can do it with canis analytics on the streaming data or if you want to run some online analytical like a long running processing queries Orab queries you can use EMR service or even if you want to create some nice dashboards for your board of directors anybody you can use Quick site or or as you are expecting if you want to run some gen application you want to store them in your vector data so like open search service you can do that so that said let's see what is Gen looks like pretty sure you all are talking about GNA quite a bit but I would have to dive deep and understand like what are the business use case for Genna there are four reasons we would use Genna One new experiences what does it means new experiences for the customers also new experience for the employees two productivity how to improve the productivity for the employees also other team members the third one is insights how do you find the Insight from this large amount of data finally creativity how to create new content it could be images or it could be movies or text how do you create this for all the these four different areas we can use gen let's see what are the possible applications can meet these business needs From aw's perspective or in general first let's think about the customer experiences that's where the chat Bots comes in pretty sure you all heard about when you go to any website something popping up you can ask questions but this chat SP a little bit different they answer you very intelligently based on all the contents ailable understand the context plus some virtual assistance conversation tools all these are coming under this category how to boost employees performance I don't know how many of developers around here I used to be a developer I I remember like many times I read a code and I made some mistakes and it take a long time to debug the code right but this code assisting tools what it does before you make the mistake it give you the popups also it writes the Cotes for you so it help us with our performance also productivity itself optimizing their business um processes again this especially for the lawyers accountant or anybody if they want to process like a long lot of documents this is where this comes in again J can help you with that offro detection detection this is also possible I'm going to dive you uh take you guys through a um simple architecture for gni again before I di and I want to tell you a story think about this ual are working for a big large railroad company and they have a 10 terabytes of data which is a policy document this is a real use case in fact I worked on that one and then the business person wanted to ask a question what is Blue Book rule means it's very much like when you see a green flag you can go or green light you can go very much like that in the railroads there a rule called Blue Book rule they want to find that so how to build a solution that can address this um this question so first of all we have this large document we need to take the document and bring them into create some embeddings what is embedding means we take this text convert that into numerical Vector with semantics meaning that we're not losing the the semantic meaning of the specific um paragraph or sentence but we still we are converting them into Vector but then we create we Chun them into Chun vectors and then we're going to store them into a vector data store once you have that document stored in the vector data store then guess what you can ask questions so when you ask the question again llm take that question and create an embeddings for that question what is Blue Book rule it create the vector conversion also the relationship between those words Blue Book um sorry U blue Flack rule meaning that it's it's a it's a it's a one word blue Flack it's one word and then it's a rule we would look for a rule specifically right it'll create this context itself but then it create the Chun vectors and go and ask the question the VOR data store then we'll get the relevant response back with the probability how relevant they are and then we can present the customers this is The Logical architecture looks like let's see how can we do with AWS again uh from your right side you can see that like for an example we have this um s maker what it can do it can take the documents and it can create the embeddings for you and it can also create the Chun vectors now you can store the all of them into a vector data store it could be Amazon open search service or Neptune or Amazon Aurora all these are vector data stor you can store or any third party tools you can sa save them there too now then you can create again you can go and ask a question using betrock API and then with API again it'll create the embeddings for you create the vector chance for your question and goes and ask the question the vector data s get the response back give it to the customers cool so we talk about gen AI quite a bit again ter I mentioned about gen bi too I I would love to dive into dive deep into this area too so before I dive deep how many of you heard about Amazon Q or using Amazon Q wow quite a bit that's awesome one of my favorite SES in fact so Amazon Q can go with many AWS services but for this specific use case or this specific business need we are integrating Amazon Q with our quick site this is another sample architecture but it does again in this architecture we talking about some document processing and creating some dashboards but again very very much like the previous use case I mentioned we have different sources of data that's coming in we have purpose build Services we taking the data and writing that into S3 our data Lake and then in even inside the S3 right we have this something access points what are those like directories we put them there now um Sor mentioned about the event event R architecture this is where the event Bridge comes in what is that it's an orchestration service it's going to call multiple Lambda function the seress functions sequentially or in parallel based on the business need and then it goes and read the read inside the bucket and do the processing imagine you have some image and that has some text you want to pull the text out of the image you can just textract or clo clot three and then once you get that one you you want to run you want to store the data into some data store again you can call a Lambda to save that into a data store here I put Amazon Kendra specifically because it is an implementation of open search service but it's fully managed so you put it there or Aurora and then you connect that to Quick side to create the nice dashboard guess what now the quicksite Q comes in you can ask some natural language questions such as for an example this is one of the examples right you pulled all this business sales data and you want to ask um what are the tops what are the top segments you ask this question so what do that means top segments could be top 90 percentile which means it can do some calculations behind the scene it creates a bar chart here for you it could be a pie chart if you want you can modify the chart also it it does some basic calculations as well that's one use case with genbi this is one of my favorite use cases in fact creating some stories storytelling is very important in the business World in fact this one aist storytelling in quick side is pretty cool very very powerful feature imagine you have all the sales data especially the campag data and you want to find how is the Campion performance over time also you want to find who are the uh top account managers also you want to ask how can you improve improve the Campa performance it s if you ask if you want to ask all these questions you can ask it out guess what it's going to create a nice story with a nice spider chart if you want you can refine that and then you can take this and present it to the board of directors so that said over 100,000 customers um using quicksite right now so can you play it out again there's some free versions available play play with your company talk with your essays they'll happily give you some demos and this is one of my favorite stories in fact BMW what they did is they exactly took the storytelling um functionality and they created the stories based on the sales data and such as their sales data and they created some outcomes so they were able to improve the business performance drastically so that it let's recap um quickly what we could do so for all the um Daves here or people who aspire to be Dave couple things I would recommend we would recommend one meet the AWS Partners ISU Solutions also your Solutions architect uh at the new to KI askk there's a Ki askk in the Expo Center plus attend NTA 316 resilience design um it's in a good session to attend to learn about Dr for Elsas here um if you want to understand the foundation technical Foundation again ask your sa team about um oh meet your sa team at the Kias and ask about the Resco workshops our team is doing a really good job and if possible talk with them about joining the Resco Workshop now or later on all good if possible attend NTA 46r again this is landing Zone setup processed from Zero to Hero also NTA 402r using A's budget to uh stay cost optimize who doesn't want that um for the Ian here again I'll meet you let's say at the new skas and ask your account team about map program this is an investment program for your company if possible at an NTA 314 it's an Amazon connect session for all the amas uh as she sit about ebcs EBAs working backward sessions also if possible at an N 315 building AWS generative AI application powered by anywhere data also n 304 or boost your kubernetes journey with eks blueprints for cdk and Amazon Q so that said to ask one please please please fill your survey or in your mobile app because that data is going to help us to improve our session over time two if you want to learn about the new tw's Amazon Journey scan this QR code this will take you to a URL that will give you some of the pre-build solutions all these things all free all these Solutions are available you can access them on behalf of our team Teresa Eric Soro thank you
2024-12-07