(upbeat music) - [Announcer] Please welcome to the stage, Yaara Tam and Barbara Baeyens. (upbeat music) - Hello, good morning, and welcome to AWS Public Sector
Symposium, Brussels 2024. I'm Yaara Tam, leader of the Multi-National
Organizations at AWS. - And I'm Barbara Baeyens, leading the public sector team in Benelux. - Thank you for joining us today for our sixth public sector
focused event here in Brussels. And thank you, in light of
those transportation challenges, even a greater thank you
for you this morning. So today is an opportunity for you to learn, to network and to
engage to help you innovate on behalf of your citizens,
students, customers and more. - But before begin, I would
like to take a few minutes to acknowledge our sponsors
because without them this event would not have been possible, so thank you so much to all our sponsors in
the room, our partners. And I would also like to invite you to visit the expo hall throughout the day, where you can learn more about what o
ur sponsors have to offer and how they can help you
in your cloud journey. If you've got questions
about today's sessions, then you can pass by
or ask an expert booth, where we have experts from AWS on site, on governments, on
education, on healthcare for whatever questions you might have, whether you're just
starting with the cloud, or you have more difficult questions on advanced technologies
like generative AI. And I would also like to feature
a digital sovereignty kiosk where we have a conti
nuity
of services kit available with an AWS Snowball, so very excited. And then last but not least, I would also like to invite you to join us for the networking reception
at the end of the day, also in the expo hall. Now we do have some interesting speakers and content lined up for you today, tailored for Europe and
specifically for your needs. With 20 breakout sessions
across four different tracks, we have sessions for everyone, whatever your role is or
your level of expertise. We have session
s on how to
build sustainable solutions, on digital sovereignty,
security, zero trust, but definitely also responsible
AI and generative AI, so check it out in the app. - As always, we are here today in Brussels because of our commitment to the public sector in the region. We know it's an exciting time for Belgium. As Belgium takes the presidency of the Council of the European Union. We also know and we understand that for our customers in the region, these are challenging times, not only becaus
e of the
economic challenges, but also because of the weight
of the geopolitical conflicts and as they make decisions around governance, policy,
and technology as well. The European Union has
been making initiatives, taking initiatives to
increase its sovereignty and its resilience to geopolitical events. AWS Cloud is sovereign by design. We are offering more
choice to organizations, with our advanced sovereign features on the AWS regions worldwide, with dedicated cloud servant solutions, infras
tructure solutions, and with our lately announced AWS European Sovereign Cloud. Dave is going to speak about those a bit more in a little while. - We are committed to the communities where we live and work. AWS provides knowledge and tools for all organizations of
any size or any industry to build solutions and implement solutions that meet their sustainability goals. AWS is offering today the most comprehensive
cloud solution in the world and also most broadly adopted. Used by millions of peopl
e, millions of users every
single day, they rely on it. Now, did you know that by
migrating your IT workloads from an on-premise data
center to the AWS clouds can reduce your energy usage and therefore also your carbon
emissions by nearly 80%? So that is five times
more energy efficient than the average European
on-premise data center. - We aim to reach net zero
carbon emissions by 2040 by investing in renewable energy and collaborating with partners
to broaden this impact. Amazon continues to b
e the single largest consumer
of renewable energy. Amazon is also on a path to reach a powering of our operations with 100% renewable energy by 2025. That is five years ahead of
our regional target of 2030. In fact, did you know,
that here in Europe, the electricity consumed by all of us of our data centers is already attributable
to 100% renewable energy? The cloud accelerates the
adoption of renewable energy, and I'll give you an example for that. AWS is teaming up with WindEurope to tackle ma
jor challenges
in the energy sector, mainly focusing on improving the process of how renewable energy projects are being approved and permitted. They have created on AWS a solution which is called EasyPermits, a user friendly, digital solution designated for authorities and developer. It is automating and digitalizing
the whole permit process. And it is expected to reduce
the administrative workload in up to 50%. It'll also allow the
agents to take processes of more than three times
admissions a
t one time. - We continue to work being
water positive by 2030. Now, what does that mean? This means that AWS will be returning more water to the communities
where we operate in then we use actually for our consumption, for our production, for our operations. So through our water stewardship
work that we've done, we've returned already 3.9
billion liters of water every year back to the community through replenishment projects that we installed and that are underway. And we are also committed to
the local communities
in the EMEA region. This, for example, has been done with the disaster recovery
work that we did in Greece, but also in Morocco. And on Morocco, for example,
the AWS response team, Disaster Response Team
helped to build out a map to help the rescue workers
actually with their work for the post-earthquake rescue efforts. And we also help to work to train the next
generation of tech talent. Now, did you know, that AWS already trains more than 1.2 million of people
in the Euro
pean Union alone on cloud skills since 2017? That's both free trainings
and paid trainings, but we continue to invest
hundreds of millions of dollars in organizing these cloud computing training skill sessions for people. And our goal is to reach 29 million people trained by 2025, and we're on our path to get there with already 21 million
people trained worldwide. - One of our programs to achieve this goal of
training, education goals is called AWS Re/Start. AWS Re/Start is a free to the learner
, full workflows development plan that helps jumpstart
careers in cloud computing where there's no technology
requirement or experience, prior technology experiences
is required for that. AWS Re/Start has been
operated here in Belgium for nearly four years. We have cohorts in the
Netherlands, in Poland, in Turkey, and here in Belgium. In Poland, for example, we are working to skill and re-skill
the Ukraine refugees. In Turkey, the efforts are more focused on educating those who were impacted by
the devastating earthquake
that happened in 2023. One of the biggest thing that people are interested in learning is artificial intelligence, AI. Late last year, we launched AI Ready. It is a new commitment
to train with AI skills 2 million people globally. To reach that goal, we are
launching new initiatives for adults and for young
learners, and we're scaling up our existing free AI skills training. Also, we are pleased to announce that the new and first EU
government learning plan is now live
. This plan allows more than
60,000 EU government officials the access to more than
30 hours of free education to get them knowledgeable
and introduced to AWS cloud and to generative AI concepts. You can enroll by this
QR, or you can speak to your account manager
about this new program. - That's amazing, Yaara. But now we're gonna hand it over to our Vice President of Public
Sector Worldwide, Dave Levy, who's gonna talk about some
yes, interesting customer cases, success stories from
your peers
in the region. And also some recent AWS
developments and announcements. We also have two great customers that are gonna be on stage with him. And we're gonna have a
demonstration on generative AI, so let's get it started. (audience applauds) - [Announcer] Please welcome
to the stage Dave Levy. (upbeat music) - Thank you. Thank you. Well, welcome and good morning. And I just want to thank
you, Yaara and Barbara and echo what said, which is thanks for making the journey
today and joining us. And b
efore we get started, I
do want to say that the group, the drummers we had this morning, I really wish somebody would give me their telephone number
and contact information. I'd love for them to help me
wake up my kids in the morning and get them ready for school. I could use that kind of energy
and enthusiasm every morning trying to get my kids out of the door. As I said, my name is Dave Levy, and I have the privilege of leading AWS's Worldwide
Public Sector business. And this is my first time
to
Brussels in that capacity. And, you know, European customers have always remained central
to our decision making at AWS. And from the European Parliament
to the European Commission, to the Bpost Group and the World Intellectual
Property Organization, we've worked with those
organizations and many, many more. And we're here today because
of our commitment to the region and serving you with
solutions and technology that are gonna improve the lives of citizens, students, customers, and all of yo
ur constituents
and stakeholders. And if there's just one thing I would want you to take away from today, it would be that it would be commitment. At AWS, we are profoundly
committed to the success of all the communities we
serve and all of our customers. And one of the things we
talk about most frequently is leaving those customers and communities better than when we
joined in serving them. And so our commitment is been enduring. We will be continue to be committed to Europe and the entire worl
d. And knowing that we'll take you on this journey hopefully today. Now, a little bit about AWS. You know, we built an
infrastructure from the ground up. And it's optimized for scale. It's optimized for
security and reliability. And that's everything from
routers to cooling systems to networking protocols. And it's all designed
to help you innovate. Our global infrastructure is fundamentally unique
from other cloud providers. In fact, what we like
to say is we're hoping to do the undifferentiate
d heavy lifting of the infrastructure for you. And we offer to do that. We offer the broadest and
deepest set of capabilities that matter to you, our customers. And it's like having three times
the number of data centers. You saw the statistics around how many data center regions
we have around the world. And we have three times the
number of data around the world compared to the next
largest cloud provider. And that makes a difference. That makes a difference in capacity, it makes a difference
in availability, it makes a big difference in reliability. And you just heard from
Yaara that we're five times, five times more energy efficient than the average on-premise
European data center. And, you know, you not only, so by moving to AWS,
you not only save money, but you also reduce the carbon footprint. And we offer over 60 more services and 40% more features than
the next cloud provider. Now, you're gonna hear
a lot about security. Security is our very top priority at AWS. And we know it
's your priority
and top of mind for you, too, but public sector
organizations are really this, what me and my team think
about almost day and night is about the security of
our public sector customers. And you can collaborate
with us to design solutions that meet security and
compliance requirements. And we're doing this all over the world. It's clear with our shared
responsibility model that we protect government
data, patient data, financial data, and other
highly sensitive workloads. Now, yo
u may have heard before, at Amazon, we are customer obsessed. And, what does that mean? That means that we work
backwards from your needs to invent with you or
invent on your behalf. It's a fundamental way that
we work with our customers. And we know here in Europe
and around the world that sovereignty is top of mind, so giving you our customers sovereignty has been a priority for us for well, since the very beginning
and when we were the only major cloud provider
that allowed customers to contr
ol the location and
the movement of their data. So 82%, 82% of global organizations are either currently
using, planning to use, or considering sovereign cloud solutions in the next two years. And that's according to the global marketing
intelligence firm, IDC. So we're committed to sovereignty. Without compromise, we're
investing in solutions and we've invested heavily already, but we're investing in solutions in that help highly regulated industries. We're investing specifically for national s
ecurity and defense, for highly regulated customers that have needs around
security compliance. And we've pledged, we've
pledged to continue this kind of ambitious investing program. And we've created a roadmap of capabilities for data residency that provide granular access restriction, encryption and resilience. Now, an example of that is NATO. NATO School Oberammergau needed to migrate their
website and learning platform. And they did that, quickly
and securely with AWS using our guidance in s
omething called a Trusted Secure Enclave on AWS. And by using this architecture and working with our partners, Catalyst IT Europe and Identify, E-to-E, they securely migrated
to AWS within two weeks. And you can hear more about this in one of the breakout
sessions later today. But what's important about that is we had architecture
ready for them to evaluate, and they could do that with speed, and they could have
confidence that it was secure. Now, we're focused on providing
customers with more c
hoice. And from our sovereign by design features in our AWS regions, almost worldwide, this includes dedicated sovereign cloud infrastructure solutions like our recently announced independent sovereign cloud,
European Sovereign Cloud. Now, the AWS European Sovereign
Cloud is being launched to provide customers in
highly regulated industries with more choice to help meet
varying data residency needs, operational autonomy, and
resiliency requirements. It's a new independent cloud, and it's operate
d within
the European Union, and it will have the same high-end security and
availability and performance that our customers get from
any existing AWS region today. Now, to address some of our public sector and regulated industry customers who need dedicated infrastructure and to help meet their regulatory and compliance requirements, we recently launched AWS
Dedicated Local Zones. Now, Local Zones are a
type of AWS infrastructure that is fully managed by AWS, but it's built to your standards,
f
or your exclusive use, where you want it and how you want it. And it's placed in your specified
location or data center. It's a dedicated local zone that, and they meet the same
AWS security standards that apply to all regions around the world and all local zones around the world. So there's a lot of complexity, there's a lot of complexity involved with navigating the evolving
digital sovereignty landscape, but you don't have to do it alone. It takes teamwork. Stop by the Digital
Sovereignty Zon
e at the expo and you can learn more. And we're here to help
and so are our partners. So in addition to all the
efforts we're leading, our customers can also
tap our global network of more than 130,000 partners worldwide. And they're specialized in various competencies
and industry verticals to get local guidance and services that can help you with things like leveraging integrated
offerings for architecture and configuring tools and controls. And so now, we talked a
lot about sovereignty, but l
et's move on to data. AWS works with millions,
millions of customers to solve some of the most
complex problems in the world. They're leveraging massive amounts of data to solve world problems, like finding cures to pervasive illness or preventing deforestation. And to get the most value from your data, you need three things, a
comprehensive set of tools. Tools that account for scale and a variety of data in
many ways it will be used. The tools needed to integrate with other services and combine
that data are dispersed all over your organization. And you need governance. You need governance to
provide the right balance of access and control. And so to make sure customers can store and analyze their data, we have a set of
comprehensive data services. Our customers never have to
compromise on price performance, speed, flexibility or scale,
and can instead simply choose the right tool, at the right
time, for the right job. But building an end-to-end data strategy doesn't just stop with a
broad set of capabilities. It requires you as customers
to integrate your data so that you can quickly
and easily access it and act on it no matter
what, where it resides. This is where customers need
to be able to connect the dots between the data sources across various departments
in various services. So to empower people
across your organization to use data to the fullest, you need both access and control. So with robust governance
your organization can give people access
to the data they nee
d, and only that data, no matter
in which database it resides. The Belgian French-speaking radio and television broadcaster,
commonly known as RTBF, is at the forefront of public service media
recommender systems. Through the use of AWS Data Services, RTBF offers personalized
content experiences across its varied media landscape including television, radio,
and digital platforms. The integration with AWS has enabled RTBF to use the efficiency of tools like Amazon Redshift and Amazon Glue. And pa
ving the way for sophisticated content
recommendations tailored, tailored to the preferences of over four and a half
million registered users. Another organization using a
wide range of AWS Data Services to innovate with their customers is the Belgian Post group, Bpost. And we've got Wendy Joos, their chief data officer
to discuss that journey. - [Announcer] Today, the Bpost Group is present in 13 countries, and has over 36,000 employees, delivering 650,000 parcels per day. A small organization
has come a long way. Bpost has not just grown, it has evolved, adapted, and diversified. Consumption patterns have
changed radically in recent years and consumers have new expectations, E-commerce, environmental issues, increased demands in tighter timeframes. To meet the needs of as
many people as possible, Bpost has surrounded itself
with the best partners and transformed from a local Belgium postal
company with subsidiaries into a group both, in
Belgium and internationally. Bpost Group. Whate
ver we do, whatever changes we make, we do so with one guiding principle, to connect people in a changing world. A world in which the needs of citizens, companies and institutions have moved far beyond the postal services, and in which our local success is inherently linked to
our global achievements. Because that's what makes us move. - [Announcer] Please welcome
to the stage, Wendy Joos. (audience applauds) (upbeat music) - Good morning, everyone. Today, I would like to share with you the stor
y of Bpost, of our data journey. And it's a journey that
began towards the end of 2019 when we decided that we were fed up with our traditional BI environment because it was old by that
time, costly as a result, but also built in silos, very much centered around reporting. And it gave us data in
a day plus one system, so we would run all of our transformation
processes overnight, didn't have our data by the next day. And what we really wanted
was to have real time data. We wanted to have a
platf
orm that would allow us to do more with our data in terms of business insights
and modern technology. Had to be scalable,
flexible, but also expandable because Bpost is also a group, and we wanted to go beyond just
Belgium as a data solution. We wanted to have something
that we could leverage for the companies within the group. So early 2020, the data team, or first version of the
data team, was created next to the IT team to give a proper focus to the data domain and separated from the
IT devel
opment activities that we were already doing. And this is still a situation today. Today, we still very much
are two teams separate, but obviously, very
closely tied to each other. We are very close collaboration. And we are both part of the responsibility of
the chief digital officer. Now, what did our journey look like? It was more or less something of a startup
approach that we did. So early 2020, we built a first version of the data platform, put some data on it, especially operational data.
We started off with a team that was basically just
me and 20 people in India. And we started testing,
trying to make use cases. We chose topics that we knew related to our internal stakeholders
because we would then also wanted to use these as
showcases to show people what the potential was
of using a data platform. And so from 2020 onwards,
we focused on new areas. 2021 was a big focus on real time because that was really important to us because of our operational activities. Then moving on to
data-as-a-service,
analytical models. And recently, I think like many companies, we've started to look into how we can apply AI
solutions to our business, so that it could make a
difference in people's lives. In the next few slides, I would like to give an example
of each of these domains that have really added
value to our company and were really relevant in
the solution they offered us. The first big added value was
already delivered in 2020, so the year that we started. In the end of year pe
ak, so the end of year peak
is usually the, you know, the busiest time of the year for Bpost and our other companies
throughout the world because it's when everybody starts ordering their Black Friday deals and then onwards to all of
their Christmas presents. But 2020, as I'm sure you will remember, was the full pandemic. There was a lockdown. Shops were closed. And we had unprecedented
volumes of parcels in 2020. So much that we really hit the boundaries of our operational capacity. And so espe
cially in pickup
points, parcel pickup points, we needed to extend our network. And we could do that with Decathlon, the Decathlon stores in Belgium, who were closed at that time. And so we started using those
locations to pick up parcels. The thing is that a few days in the government decided
to reopen the stores. And so Decathlon gave us a call and said, Dear Bpost, please remove all of your your parcels from our stores because we are reopening
in a few days time. We had thousands of parcels
s
tored in all of these locations and the problem was that we could only distribute
them from those locations. It was not an option to feed them all back to our central sorting activities. So we had to distribute and deliver to the homes of the receivers. Our IT system obviously did
not know these locations because they were meant to be pickup. And following the traditional
IT development cycle, it would take us two to three
weeks to adapt the system. And you can do the math, by that time the end
of
year peak would be over. Fortunately, we had been
playing our data platform with a few models on distribution rounds,
organizing distribution. And so we were able to reuse a model adapted in two days time. And make two specific models, one based on work time for the rounds
done by our own employees and one based on van capacity for rounds done by contractors. And in less than one week, all of the parcels were distributed to happy receivers of their parcels. So it was not only a quick win, but
I think also a lifesaver
for Bpost at that time. Next example, is digital twin. So this is an example of
the use of realtime data. It's a realtime replication of everything that happens
in our operational processes, sorting, distribution, transport, but also the warehouse
management activities of Radial, which is another company in the group. They're using it throughout
the warehouses in Europe. And also their business clients have access to this live tool so they can follow up on their operati
ons as Radial is fulfilling the activities. This makes a difference in the
live of managers, team leads who day-to-day have to deal with things that sometimes go wrong. And this really enables
them to be proactive and to correct problems before
they become a real issue, so a big improvement in quality. But also in work experience because it really gives them
a feeling of enablement, being able to do something
before it goes wrong. Two more examples on apps that we have. This is the first one,
an
app for our mailman. It's called Myworld because it's centered around
the world of a mailman, where we try to interact with a mailman, but also give him or her feedback on how he or she's doing in
terms of eco driving scores, NPS scores, customer complaints, but also a gamification
aspect of, you know, the number of parcels that it distributes. So we will send a message
out to to say, you know, you had the biggest
amount of parcels today, or you were done in the
shortest amount of time just to
also motivate them to
do well in terms of delivery. And that works, it works really well. So it's also a way to connect to people that are always in the field and to be able to reach
out to them in that way. And the last but not
least, the Mybpost app, that is our customer or end receiver app, the people who receive parcels
at home where we've been able to improve our delivery prediction model. We've had a delivery
prediction model for some time, but we've been really able to do, to produce a mu
ch better
model with a higher accuracy. And not only predict the time of delivery or the window on the day itself, but as of when one of our business clients creates a label to put on a parcel, we already predict the day of delivery, which is not necessarily the next day. So based on delivery patterns
of our own business clients, then we already give the
receiver the day that we expect that the actual delivery will take place, which is something our
competitors not yet do. Maybe just briefly, lo
oking back in the four years that we've had, I think the conclusion is
that having a data platform is a real game changer. We've been able to create solutions that we would not have been able to do if we had not invested in a data platform. Real time is an absolute necessity, but also data-as-a-service, feeding intelligence to an IT system, being able to avoid complexity in the development of an IT system because you do it on your data platform is a real game changer. And obviously being able to
work with machine
learning models and AI is very important as well. It was said before, governance is key because we have vast amounts of data. People who love to work with data, they love playing around
on a data platform. So you really have to keep a focus on what it is that you're
doing and why you are doing it. Also, who you give access
to data very important, clean it up from time to time. Looking forward, how are we
going to continue our journey? Well, definitely continue to exploring the
different data domains. Predictive remains a very important topic. Applied AI is something
we're looking into now, but we're also getting
out of the startup mode into a more organized and
structured approach now. So that means that also data governance becomes a very important topic
and also a data organization. So also in terms of
having the right profiles to work with data, not only the data team, but
also within our business teams, is a main focus for 2024,
and then going forward. And last b
ut not least, I forgot to mention it when
I showed you the slide, but we baptized our platform,
the global data nucleus, or briefly, the GDN. Global reflects our ambition in terms of the data team
to work for the group. We started out of the
Belgian business unit. We've started to work
for other subsidiaries, but it is really our ambition to serve all of our subsidiaries around the world with the data platform, that's
how it's been constructed. And so in the next few
years we'll be definitely al
so investing in realizing
that group ambition. So this is it for now, for me, but if you have any questions, do feel free to come up to
me today and and have a chat. I'd be happy to speak to you. Thank you. (audience applauds) - Thank you, Wendy. So, I can't repeat the
wonderful journey she's been on, but I did hear data, data,
data as the foundation. And getting your data in the right place and getting a governance model around it, or one of the foundational keys. Artificial intelligence, you k
now, AI has been around for quite some time. And with all of your data
we saw from the example, there's so much potential for innovation, including machine learning
and artificial intelligence. And AI adoption is forecast to unleash 600 billion euros of
growth in Europe's economy. And AWS has a broad and
deep set of AI services. Each of those are purpose built
to address your challenges. We also have a history of innovating for the last 25 years in
artificial intelligence. Now, that's through ou
r Amazon businesses like our retail business and Alexa. Now, according to a recent survey that we commissioned and released, today, as a matter of fact, most organizations will use AI by 2028. They surveyed organizations think positively about the impact of AI on their industries and their businesses. And nearly nine out of 10
surveyed employers and employees see benefits from using
generative AI in the workplace. So from data driven, we
expect that organizations will evolve into being AI driven
. The current generation of AI
is even more transformative. Generative AI, which you may have heard
a little bit about lately and is kind of buzzing
everywhere, is expected to become an integral part of the workplace by 2028. The top benefits cited
are in-task automation and supporting greater
innovation and creativity. Now, while AI is expected to offer the greatest value to
employees in the IT department, other departments and non-tech employees are also expected to
benefit from AI deployments
. While eight out of 10
respondents expect to see the benefits of generative
AI in IT departments, seven out of 10 workers
in non-IT departments are to benefit as well. And so generative AI will truly transform how
we work in the future. Now, we've worked to
democratize generative AI for organizations of all sizes
and all technical abilities. And we're doing this by providing the most comprehensive set of capabilities across three layers of
the generative AI stack. Now, this bottom layer is
the
infrastructure layer. It's the layer to train
large language models, and other foundational models, and produce inferences and predictions. Data is your differentiator. It's your differentiator for generative AI when you're building your
own large language model. When building your LLM, your
data needs to be up to date. It needs to be complete,
accurate, discoverable, and available when you need it. And a critical step to achieving that is moving the data into the cloud. Having an effective foun
dation
based cloud strategy will accelerate innovation. Now, the Technology Innovation
Institute in the UAE is an example of an organization that worked with us to build their models. They launched the Falcon LLMs, three foundational large language models to facilitate research in
healthcare, finance, education and even in some other areas. Now this next layer, the
middle layer of the stack, this is simple access to
all of your models and tools that customers and you will need to build and scale
generative AI applications with the same security or access
controls and other features that you expect from an AWS service. So since different models work better for different use cases,
customers don't want, what we hear from customers is they don't want a cloud provider who is beholden to primarily
a single model provider. What we've heard is that customers want a real choice of models. Amazon Bedrock is the easiest way to build and scale
generative AI applications with large language models
and foundational models. Bedrock will provide access to the broadest choice
of foundational models from leading companies
like Anthropic's Claude 3, family on Amazon Bedrock,
Amazon's own Titan models, and the recently announced
foundation models for Mistral AI that were developed in France. Bedrock is also designed to support enterprise
grade security and privacy. And as an example, no customer data is used to train or improve
the original base model Retrieval augmented generation, or RAG, wit
h knowledge bases is
now generally available. RAG with knowledge bases is another technique for
customizing models in Bedrock. It allows organizations
to have a foundation model and consider new knowledge
from their enterprise data when offering responses. It lets your organization tap
into multiple data sources like document repositories and databases. And finally at the top
layer of this stack, we've been investing in
applications in key areas like generative AI based
coding and Amazon Q. And
I'm gonna have Dom talk to you a little bit more in detail about Q, which we're very excited
about, and CodeWhisperer. Come on up. (upbeat music) - Thanks, Dave. Really
excited to be here today. In my role as Global Vice
President for Public Sector, I have the privilege and honor to be working with many
of the technical people that work with you to help you figure out how to best leverage a platform
to accomplish your mission. In this case, we're
gonna talk a little bit about how to use the AWS
platform and our new artificial
intelligence services, particularly generative AI, in order to do some really
unique and creative things around some of the activities
that you're trying to perform. One of the things I'd like
to do is mention the way that we're leveraging all of
these different LLMs and models from across the spectrum. Dave mentioned the
Mistral AI, which is one of the most recent AI models
developed here in Europe. And one of the things we're super excited
about with Mistral's A
I is it's one of the fastest, scalable, and most efficient models we've ever seen. One of the things we love about
our Bedrock is this ability to adapt and leverage
new models as they appear and make them available for you. But, you know, enough talking
about these different models, let's see a demo. I'm really excited to show off a demonstration that's been built by two of my solution
architects here in Europe. Maria and Inez have
put together an example of how you could build an
application us
ing generative AI to accomplish things in your mission. What we've got here is a demonstration that leverages some content that you would then potentially upload and then use for training. In this case, I've taken a video about generative AI building
generative AI applications. I upload the video to the system. As I page through the video
kind of stepping through it, you can see this is about how to build generative
AI applications on AWS. And you can see some
good pictures, bullets, all kinds o
f things that talk
about ways you can do this. And obviously, there's someone who's a lecturer or a talk track that's going through this presentation. What we've done is uploaded it
to our generative AI service. And we can see at the bottom are these little buttons
that have appeared that have processed this video so it can be used for benefit
and used for activities. Let's take a look first at the
transcript that was produced. So you'll see the demo, we'll move down, it'll click on transcript.
And now it shows you
using Amazon Transcribe the full content of the lecture. It has a lot of technical content between percentages, numbers and you can see it does
a quite a good job. Second, we have the summary, which produces learning
objectives and bullet points, the highlights of this presentation. Another nice thing about
thinking about ways you have to focus on and
say, what is emphasized? We go to flashcards. Created here are flashcards that can test the student to say, here's a question
about the content, here's the answer that you can use. Now, I built all this in
English, but, of course, my solution architects speak
other languages as well. And they said let's show it in French, so here are the flashcards
just automatically translated using Amazon Translate into French. And, of course, now we have sound as well. (program speaks French) So that uses Amazon Polly to make the the content
much more accessible for people who want to
listen to it as well. And then lastly, I have
s
omething called Assistant. Assistant Lets me ask
questions of the content as if I was watching the lecture and could raise my hand
and say, what was this? What was what this content said? A question I can start out in
French. I get a French answer. Even though the content of the lecture itself was in English, it
auto translates my question. Auto translate the
answer to that question, even though that may not
have been content in there, it generates the answer, translates the answer, and provides
it. I can switch back to English,
ask the same type of question. And then also get that
answer now in English, and then have this discussion
with the lecture content. So you could see how
powerful this would be if you're creating training videos that you want people to interact with and not just sit there
and click through them and have a test at the end. But give them assisted things like flashcards, the ability
to interrogate the content, and also get a summary
through the transcript. I think
it's a really powerful demo of how you can use our
foundational AI services to build real world applications that can be used inside your organization. You can imagine the possibilities, not only for government,
but for education, healthcare and nonprofits,
which is really exciting. So I mentioned, you know, how can generative AI be
used in the public sector? And we gave an example here
of a learning experience. We're also seeing use cases
like this in government, healthcare and non-profits to
say, how can I do
things like member drives, do things like summarizing legislation? All kinds of wonderful
activities can be built on top of these foundational
generative AI technologies. One of the most impressive
bits of technology that I'm excited to work
with and work with you on is something we call Amazon Q. Amazon Q is a broad name that describes a bunch
of different services that leverage all of our
foundational AI capabilities to give you a new capability on top of each of these
new la
nguage models. In many cases, Q leverages multiple models to provide answers for you. I'm gonna go through the way
that we've started to use Q in many of our services. We've already integrated it
into four of our services and you can expect to see many
more in the months to come. One of the things we're
gonna talk about at first is Amazon Q as a business enabler. One of the things that Amazon Q can do from a business perspective is reach out and connect
into your business systems. Amazon Q today
already has
40 different connectors that can talk to services
like S3, Salesforce, Microsoft 365, ServiceNow, Gmail, Slack, Atlassian for Jira tickets and stuff in your software
development cycle, and Zendesk. It respects all of your existing
controls on those systems, but allows some of your users and the people in your organization to basically ask questions
of your business systems by virtually the way
it's connected to them and provide answers that
then can take action through some of our a
gent capability to do things like create a ticket, open a support case, and document it. The second area where
we've applied Amazon Q is in building applications. This is an enabler for your
builders and developers. It was trained on over 17 years
worth of Amazon knowledge, the questions and answers
that were asked of us through our repost system,
as well as experts across AWS who helped train the model to understand technical questions, to provide technical answers,
giving you information basic
ally from the Amazon documentation and different code samples and snippets so that the knowledge base here is based on helping your
builders and developers. And all through to this is
something we call CodeWhisperer, which is an augmented
capability for developers to leverage generative
AI to be more productive. Our partner worked with
the European Commission's Director General for Maritime
Affairs and Fisheries. And they started to
generate high quality code, 30% faster development,
for each de
veloper. Huge efficiency gain, and
with more than 90% accuracy, lowering the in the incidents
of bugs and other problems that have to be solved
in writing software. Super powerful creativity,
but we didn't just stop there. We've also enabled Q for developers to do things like code transformation. This is super exciting for things where we have to keep systems
up to date and maintained. If you've ever had a system where you've had to go
from Java 15 to Java 17, you know there's a lot of work in l
ike looking at all the
signatures and the code, figuring out what needs to be changed, how to develop test cases. One of the things Q for
code transformation can do is explain the code to you. You can ask it, what does this code do? And it'll produce documentation about what function is being performed. You could then say generate test cases to make sure that this works well. So all of this kind of code coverage helps you maintain a system, which we know is over a life of a system, really takes
down the level of effort to keep a system live,
fresh, and going for you. Another place where we've applied Amazon Q is in our call center
product, Amazon Connect. This is a generative AI contact center agent
assistance capability. This allows your agents to
have real time conversations with your knowledge base that helps answer customer
questions in real time. It allows you to onboard agents faster because the essential
knowledge and training is built within the product and can be interactive w
ith
in a natural language fashion. This helps your customer agents respond more quickly to questions and helps onboard them that much faster. The last area so far that
we've introduced Amazon Q into is our QuickSight, or
business intelligence product. One of the things we've
heard from customers is I'd love the ability
to tell a story with data. People who really are
tired of static dashboards, pie charts and pictures that
they do in presentations. They want to be able to
ask questions of data a
nd quickly present things that are relevant and
motivating when you say, how can I show this data in a
fashion that exhibits a trend? And so now with Amazon QuickSight, you can use Amazon Q as a way to say, this is the kind of
dashboard I'd like to say. This is the story I'd like
to tell about my data. Please build a dashboard that evidences and shows off a data in the way I'd like to have it presented. This is a super powerful capability, to not only enable IT folks, like your developers with A
mazon Q, but end users and business analysts who need to produce
information based on data can leverage this generative AI capability to tell their data story. We're gonna be doing Amazon Q to many more services across AWS and hope these four examples show you that we're committed to
leveraging generative AI to improve all of our
services for your benefit so that you can be more innovative. But what if you'd actually like to get your hands dirty today
in a much more easier way? Now obviously, th
is demo that was built by my solution architects
is super impressive, and I can tell you it's
really easy to get started and you can go there as well, but if you'd like to really
get some hands-on capability, I recommend you check out
something called PartyRock. PartyRock is an easy way for you to build simple generative
AI applications. You literally just describe what you'd like the application to do, and PartyRock actually builds code to show you a sample application that achieves the result
that you're looking for. It could be something as
simple as writing poems, describing your music collection. One of our vice presidents actually use it to determine the ferry
schedule in Seattle. So I hope you take advantage of this. Check out PartyRock.AWS. And experiment yourself and see what creative
things you can build. Well, I hope that
exploration of generative AI was informative to you, not
only on what can be done, but also how we're embedding generative AI into all of our services so t
hey're easier for
you to innovate with. But, you know what? I'd
like to move on from that. And talk about how moving from the cloud, moving to the cloud from
on-premise infrastructure opens the door to this kind of innovation. Our next speaker is gonna share how their journey to the cloud led to a positive shift in
their entire IT culture. From World Intellectual
Property Organization, I'd like to welcome to the stage their Chief Technology
Officer, Francesca Duri. (audience applauds) (upbeat mu
sic) - [Announcer] Please welcome
to the stage, Francesca Duri. (upbeat music) (audience applauds) - So welcome, everybody.
Thank you for having me here. And as was just said,
I'm going to talk to you about what comes before
being able to exploit artificial intelligence and generative AI in an efficient way, and
that's the move to the cloud. And I'm quoting myself in
that in my long experience you find out that you made
the right technology choice when you get more benefits
and additional benefi
ts with respect to the
benefits you were expecting from that technology
in the very beginning. You get something that
you weren't expecting. So WIPO, is not just about
filing patents and trademarks. We believe that intellectual
property should be deployed for the benefit of everyone, everywhere. We serve 193 states. And our initiatives range from filing hundreds of
thousands of patents, but also helping creators protect their intellectual property. Helping the indigenous people protecting their
intellectual property. Supporting the Accessible Books Consortium to make sure that applications
are accessible for people that would be difficult to
access normal publications. We are digital, all of
our services are digital, and we deliver our services
through application that we develop for the
most part ourselves. And back in 2018, it became clear that in order to be more efficient, we had to move to the cloud. It's not been an easy journey. As I said, we started in 2018. And that was the ye
ar where
we build the foundation. Building the foundation meant
to adopt a cloud first policy, to build a landing zone, to learn about what architected framework and to put an internal
governance in place for whoever wanted to deploy or migrate
their applications to the cloud. Now today, we have more
than 70 applications deployed in the cloud. We have a data lake. And the cloud enables
us to build a data lake and a data analytics capability. But it was a long journey in between. And in 2022, we
faced a dilemma. What was the situation by then? In 2022, we were, you know, mature enough for all new applications
to be cloud native. And for our major
application, the application through which we deliver
most of our services, to be in the process of
being migrated to the cloud. But the pace of migration was not the pace we wanted really to be. And our infrastructure, our base services were
still mostly on-prem. And that was impacting,
not just the speed, but also some of the benefits,
such a
s example, reliability. Having, you know,
applications in the cloud, but the base services that were supporting them still on premise meant that we weren't able to get the reliability
that we were looking for. And that was a year where, you
know, it took a leap of faith or a little bit of, you know,
forcing the path forward, through a project that is not, what you will get from the
literature or the best practices because, you know, it's quite
clear that lift and shift is not the ideal way of
mi
grating to the cloud. But sometimes you don't have the time, or you don't have the
resources to refactor or, you know, rebuild everything. And the purpose of this move
was to get a big acceleration. And through this project
and through our migration, we got more benefits in reality
than what we were expecting. So by migrating to the cloud, and specifically with this
mass migration project, what were we looking for? Of course, we were looking for
more agility and flexibility. So being able to dep
loy, to build new applications in
a very limited amount of time. We were looking for improved
service to our stakeholders and the way when we measure our service, the quality of our service,
is with the Apdex index. So we measure the experience of all of our customers
all over the world. For us being really, you know,
treating all of our customers in all 193 countries the
same is very, very important. And, of course, we were looking for better business continuity
and that optimizes our cost. And
those were the benefits we realized. But, what were the benefits that we got as a bonus, let's say? New solutions to old problems. I'll just give you an example. An on-premise solution
that is a global database, whether we provide search on
a global database or brands, and this service gets attacks by bots. And the application teams, they were fighting to
counteract to those attacks. And we just simply very, very quickly were able to put the
web application firewall in front of the application.
So even if the application
was still on-prem, deploying the web application
firewall in the cloud delivered immediate benefits. The mass migration project meant that we had to have our hands dirty in all the applications, all servers. And that helped us a lot to
reduce our technical debt a lot. And if you are in IT and
you've been in it long enough, you know that, you know, tech,
getting rid of technical debt is probably one of the
most difficult things. That all of us have these
skeletons in t
he cupboard, we would really like to get rid of. We got cost avoidance. And that, you know, our new solutions are really very much pay-as-you-go. We don't build anything new. And we got better architecture practices. We had, as I said, to
put a strong governance at the very beginning of our journey. The result is that our whole architecture practices improved. All of our applications and architectures are now documented. They all undergo a formal approval, architecture approval process that impr
oves our security
of course as well. And that was something that got really a push from
the migration to the cloud. And last, but not least, involving also the operational teams. And not just the development
teams in with this, let's say, mass migration project, we upskilled them with real work, where they were really
learning on the job, side-by-side with the professional
services that helped us. And it was a little bit of
a team building exercise. So our, let's say, if I can summarize the fina
l result that we
got is moving really from, what was at the very beginning, a cloud first policy to
a cloud first mindset in that, you know, we think
about solving problems. So cloud first, that's immediate.
It's a mindset. Thank you. (audience applauds) - Well, if you're ready to get rid of some of those
skeletons in the cupboard, we are here to help you. I just really appreciated
the story and the journey that you just described. And it is a journey. And I want to say that we are here to help
you along that journey, on the decisions that you need to make on the time that it will take. My teams and I are standing by and our partners as well to help you. So thanks again for joining us today. We hope you learned something
new from our conversation, or we inspired you to want to learn more. And like Yaara and Barbara mentioned, we've got four tracks of great
content throughout the day, a robust expo, and so
much more to explore. And if you don't know where, excuse me, if you don't know w
here to
start, stop by the AWS Village, or the Ask the Expert booth in the expo. And we're here to help you
figure out your next innovation or work through any
challenge that you have. You can bring us your ideas. No matter how formulated they
are, we've got you covered. So let's get started experimenting.
Give PartyRock a try. I can share with you if you can find me what some of the things that
I've put into PartyRock, if you promise to keep it confidential. And hopefully, we'll get to share yo
ur story on stage next year. Enjoy the event today. And
thank you for joining us. (audience applauds)
Comments