Laura Spriggs
I'd like to welcome everyone to today's webinar focused on AI for the public sector, turning strategy into action. In collaboration with our colleagues from Microsoft. Today, we have hundreds of individuals from all levels of government and non-for-profit from across Canada, as well as from public sector organizations in other parts of the world. It's very exciting. This is certainly an indicator of how important AI is becoming and shaping our future.
I'm Laura Spriggs, Vice President of Public Sector Technology Consulting here at BDO Digital. I've been working with and helping public sector organizations with their transformation initiatives for over 30 years. And transformations today are unlike those that we've seen throughout recent history with the accelerated evolution of artificial intelligence. And we all know AI is changing the landscape of how services are delivered and how business fundamentally operates. In the words of our new Prime Minister, Mark Carney, AI is critical for our Canadian competitiveness as the global economy shifts. Canada is where some of the most important developments in AI have happened. It's key to unlocking productivity, higher paying jobs, and new prosperity that's gonna benefit all of us. For government, AI is how government organizations will improve service delivery, keep up with the speed of business. It's how government will maximize efficiency and reduce cost. And in fact, time savings and increased efficiency were the top priorities that emerged from the small survey that you would've responded to when you registered. And many of you are only at the beginning of your AI journey, 74% based on that survey. So we hope that this webinar will provide you with helpful information as you embark upon your AI journey.
I'm joined today by both BDO and Microsoft colleagues. Olivia Neal, Microsoft Canada's director of government strategy is going to kick things off with some opening remarks focused on current trends and the public sector's AI journey. Haya Elaraby, a strategic leader in BDO's AI, data and automation services, is then going to walk through some practical approaches and considerations for your AI journey, as well as some potential challenges that you might face in operationalizing AI, as well as talk about how you can get ready to start your journey and progress your journey and embed AI adoption within your organizations. Eugene Zozulya, Microsoft Canada's National Digital and Application Innovation Leader is going to talk about some of the AI foundations that you can build upon. And finally, Bill Syrros, our national AI leader and head of BDO Innovations, Exponential Labs, is going to close our session today with some insights regarding the role of AI, as we head into the future. We will welcome your questions, which you can submit using the comments feature, and we'll get to as many as we can at the end of the presentation. However, if we run out of time, we will be sending the responses to your questions to everyone who is registered here today.
Just before we jump in, I thought I would share a little bit about BDO Canada. So in addition to our audit tax and accounting services that we are very well known for, we also offer a wide range of consulting services. This includes our management consulting strategy and transformation services, risk advisory services, and of course our technology consulting services, under our BDO digital business line, we have a public sector client base of over 850, which includes all three levels of government, crown corporations and agencies. We're a $1 billion firm in Canada and we've been in business for over a hundred years. We operate out of 95 offices across Canada with a workforce of over 5,000, and that includes our technology consulting workforce of over 700 employees. Olivia, perhaps you'd like to introduce Microsoft Canada to our attendees.
Olivia Neal
Thanks Laura, and thanks everybody for joining us on the webinar today. So I'm gonna go into a little bit more detail on trends that we're seeing in AI in a moment, but just first of all, just to introduce Microsoft in Canada, like BDO, we are deeply embedded across Canada. We are working with governments at municipal, federal, and provincial and territorial levels. So really looking forward to sharing some of those lessons and some of those examples through the conversation today. Similarly, again, to BDO, we've got over 5,000 employees based all across Canada, from coast to coast to coast, and data center investments, which are really bringing value into a number of areas. So really looking forward to sharing those lessons from some of the 40 years of operation that we've had here in Canada so far with many more to come I hope.
Laura Spriggs
Wonderful, thank you so much Olivia. At BDO, we are extremely proud of our longstanding partnership with Microsoft and we're honored to have received the 2024 Microsoft Canada Partner of the Year award. But equally as proud are we of the numerous awards that we've received over the years, including the 2023 AI impact award in recognition of BDO's global leadership in this domain. We also hold all six Microsoft solution partner designations and each of those designations has a dedicated practice here at BDO with leadership at the partner level in addition to 15 advanced specializations that we hold. And we believe that this decades long industry experience working in the public sector ensures that our team understands the unique needs and challenges that many of your organizations face, blended with our technology consulting services, we're able to leverage best practices and innovative cutting edge cloud data and AI solutions adapted for the public sector. And central in this model, through our partnership with Microsoft, we're able to offer tailored solutions that ensure both regulatory and policy compliance and that enhance service delivery efficiency and most importantly, foster innovation. We believe that this is a winning formula that enables our public sector clients to modernize their operations, improve citizen services, drive impactful digital transformation, and achieve long-term sustainability all while maintaining security and cost effectiveness. So on that, I would like to turn things over to Olivia who's going to talk about trends in Canada and the public sector AI journey. Over to you Olivia.
Olivia Neal
Thanks so much, Laura. Well, Laura already mentioned some of the recent direction that we've been hearing from the Prime Minister and I think what we've heard loud and clear from the federal government level is a real focus on productivity and using that increased productivity to build a strong Canadian economy. And so those ambitions, those priorities are looking at how AI is used across the Canadian economy in all industries, but also within government as well. How to make government more productive by deploying AI at scale. And I think that at scale element is an important thing for us to consider as we're going through these conversations today. We're seeing a lot of people start with proofs of concepts and pilots, which is fantastic.
Now we need to think about how do we expand those? How do we bring everybody into this conversation? And while that's a federal level direction, I think this is very representative of what we've also been hearing from across municipalities, provinces, territories as well. So this is really a one Canadian approach and one Canadian direction. And so when we think about what does that productivity mean, Microsoft did a study last year which looked at if AI was adopted at scale across the Canadian economy, what type of value could that bring in? What type of impact on GDP could that have? And our study showed that that could bring in $187 billion a year in terms of added growth in the Canadian economy. So a really important factor. And actually that study was based on increases in productivity of around 30 minutes per day per worker, which is not a hugely ambitious goal.
And it's actually something that we've seen be reinforced by some findings recently, which have come out of the UK where they have introduced just one tool, Microsoft Copilot to their government workers. And they found that workers in just a short pilot saved on average 26 minutes a day, which ended up being 13 days a year. So 13 days of increased availability for tackling different tasks.
But this really is thinking about that productivity space. This is not about how do we reduce jobs. I think we're all feeling the pressure of increased activities, to do increased backlogs to tackle increased volumes to deal with. And that's borne out by a recent study which we conducted, which says 76% of workforce, all of us feel that we don't have enough time or enough energy to do our jobs as they are right now, so something has to change. And then also, similar percentage, again, 76% of leaders within that survey said one of the areas that they're really looking at is how they will use AI agents to meet that demand for more workforce capacity and agents is something that my colleague Eugene is going to talk a little bit more about in his session because this is the new area of AI where we are really starting to see those real benefits for productivity. So listen out for some more on AI agents and agentic AI to come.
I just wanted to start with a little bit of recap for anyone who's not familiar of how have we got to where we are today because artificial intelligence isn't something new for government. It might be new for lots of us in our roles, in our conversations that actually this is an area where governments have been investing in using these technology capabilities for many decades. So going right back to the 1950s, that's when AI research, AI investment really started to happen. And then in the seventies and eighties we started to see governments use those capabilities for things like weather modeling or space exploration or in the medical profession for diagnosis. So going back a really long time now, then as we went through the seventies and eighties up to the nineties, that's when we started to see obviously the internet come into play, and machine learning, and machine learning allows machines to learn from existing data and improve upon that data to make predictions. And so in a government sense that really came into play looking at areas like fraud detection for tax systems, for benefit systems, use those types of AI capabilities, those machine learning capabilities to look at those large patterns and make predictions. Then as we came into the 2010s, deep learning was the next stage of AI and this is really thinking about what we would call neural networks, which can process data to make decisions and then a government capacity, those neural networks could be used for things like managing traffic flows, understanding how those work, environmental conditions and predicting natural disasters. So AI is not something new and it's in these capabilities that we are using every day.
If you use Teams for example, you're already using AI, if you put a blurred background in, these are not new capabilities, but what is new, what has come in the last few years, which has really generated this excitement in this conversation is generative AI. So if you use use ChatGPT, if you've used Copilot, if you've used any of these new tools that are out there, then you've been using generative AI. And what this means is that we can be using our own natural language, the words that we use every day to give prompts, to give instructions, to create new capabilities, to create written material, to interrogate data, to create visual material. And what that means is that these AI tools are in all of our hands. They're not just in the hands of the developers or the data scientists, although there are absolutely generative AI tools for those specialists as well.
So that's what has really kind of ignited this conversation and really driven the focus on productivity because now this is something that can be used by all of us. So as I said, that really means that this is something that we can think about for each one of us as individuals. How might we use these types of tools that are available? What is out there? How can I use it for my work? How can I use it for my team? What does that mean for my department as a whole or for my government as a whole? How can we rethink some of the processes that we're working on? And there's a lot of questions that you're going to be thinking through as you're looking at these areas. And through this session today, there's going to be a number of options of kind of guidance in terms of how you can think through those strategies, how you can think through that approach. At Microsoft, we're seeing a number of drivers of getting to the value because it's all very well having AI capabilities, but we've got to get the value from them. And I think what's important about this is that this is not just a technology strategy. So absolutely there is a technology strategy here and one of the really important parts of that is having the right access to data and having the right technical infrastructure, having the cloud capabilities to set up.
But this is not a technology conversation, this is a conversation for an organization as a whole. So this is not just something that should be in the IT team and the digital team. This is something that has to be thought about really from the highest levels and across different organizations as well. So we have to think about the operating models that are there, support from all the right levels, what outcomes we're trying to get to, because AI is a tool, it's not going to be the right tool for every job. It's going to be an incredibly effective tool and sometimes, but we really need to understand what are those outcomes that we're trying to get. We need to think about the training, the skilling, and absolutely we have to think about the governance as well. Are we using the right tool at the right time with the right controls and processes in place? So you can read more into that approach and we've posted the link into the chat. I'm happy to answer any further questions in the chat as well after I finish speaking. But as we're seeing this in a government lens, we're really starting to see governments focus in a number of different areas in terms of how generative AI can be used.
So we are seeing people use it for delivering more personalized experiences. So you might have used a generative AI powered intelligent search on a website. Alberta.ca for example, has got a really great intelligence search feature on it. And I would encourage you to go on there and check it out and just have a play with that. And in that intelligent search feature, you can see how they badge that very clearly as something which is experimental but is really helping people get to the answers that they need. We talked a little bit about productivity already, seeing a lot of examples of generative AI being used in contact centers to empower contact centers or agents to get the information that they need more quickly. We talked about fraud detection, how it's being used in there to analyze data and then absolutely in the government operations space. When we think about some of the risks around government technology, one of the things that we think a lot about is aging infrastructure and aging applications. There are really powerful tools for developers to use generative AI to modernize code bases and help make those systems more resilient and help move things forward. And then absolutely in the cybersecurity space, we are seeing cyber attackers using generative AI capabilities, which means that our cybersecurity experts are also using generative AI capabilities to detect those threats and respond to those threats as well. So really important security element in generative AI considerations as well. And in some cases, you know, these are really happening all around the world.
There are examples in Canada, but these are happening at all levels of countries, all geographies. So there's so many opportunities in India, we are seeing a big focus on how generative AI can support people in different languages. There are hundreds of different languages spoken across India. How can people get better access to their government services using those language capabilities, helping them find the right information as quickly as possible. In Qatar, their postal service are using generative AI for document analysis and optical character recognition to help analyze and upload those documents and take away that kind of really day-to-day data entry work and allow people to focus on higher values tasks. In Laval just outside of Montreal, their municipality, generative AI is being used within the 311 service to help people get the information that they need more quickly. And in Brazil, lawyers are looking at how do they use AI copilots to really understand the historical jurisprudence of case history to help 'em get to information more quickly and to help them with drafting when they've got large amounts of information to consume.
So at Microsoft, as I mentioned earlier, we take governance really seriously. Responsible AI is something that is at the heart of how we think about operating with AI. Whenever we're working with a government on a generative AI use case, we have a very thorough process where we review that use case and we review it against our AI principles. And these are something which absolutely align with the government of Canada's faster AI principles as well as other international standards. And it's really important to know as well that when you're using an enterprise Microsoft tool, your data is your data and it's protected by the same security tools as all as your other Microsoft products. So we are not training any of those large language models which power this generative AI with your data that remains within your control. So I will end there and I will hand over to Haya, but I just wanted to leave us with the idea that this was from the CIO of the city of Kelowna. You don't have to be a huge enterprise to be innovating with AI. This is something that we all need to understand. This is going to be coming, this is here now, how can we start to take advantage of this? How can we experiment and then think about how to take those things to scale responsibly, safely and securely so that we can get to those productivity gains and serve Canadians better. Haya, over to you.
Haya Elaraby
Thank you, thank you Olivia. So for this section here, I'll speak about how we can successfully adopt AI. To start off, how to successfully adopt AI really depends on where you are in your journey. And as Laura mentioned about over 70% of the respondents for the pre-event survey are really kind of in the early stages in the basic or exploratory phases, meaning still exploring AI, what it can do or you know, what it can do kind of in the early stages. Where we see organizations really realizing ROI and material ROI from artificial intelligence is really when they start getting into the, you know, the latter part of the spectrum around the strategic, the strategic area or the strategic level of maturity. So what the strategic level of maturity is about, it's around, you know, having some AI use cases in production, having clear AI op model in terms of who's going to be doing what. And you know, with solid governance in place, 'cause governance is absolutely key when it comes to artificial intelligence as well as assessing optimization and scaling, to Olivia's earlier point. And most organizations by the way are in that early maturity phases, but where organizations will see the most value will be towards kind of the strategic and the AI DNA part of the spectrum.
Now how to actually get started or potential starting points. So potential starting points will depend on the goals and objectives of the organization. Some are listed here on the slide. So some organizations it's better to start with a strategy and plan. And some triggers to that are things that, you know, if you're hearing it within your organization such as AI is an enterprise priority, there are many AI opportunities, we don't know where to start. Perhaps a business case is required to get buy-in and this is where creating a strategy and a plan or a roadmap would be necessary. And we've partnered with several organizations to really help 'em on that, on the journey to have that business case solved business case in terms of how AI initiatives can help them get to return on investment.
The second set of triggers are around it's reporting. BI analytics, a priority for you. We have poor data quality, we don't feel that the data that we have is reliable and this could be a part where, you know, potentially looking more into the information to the insights into getting data-driven decisions, whether AI could be the potential starting point. Another set is around, you know, we're starting to experiment with AI but the initial AI education or AI literacy is lacking. And you know, we're investing in strategies to be discussed. But you know, we've identified Copilot is a quick one and this is where, you know, looking at more experimental and tactical wins, building potential kind of proof of concepts. So that can be scaled from there. And also focusing on the AI enablement and AI adoption, AI literacy would also be a potential starting point.
Now as Laura mentioned, we actually wanted to spend some time here talking about what are the challenges that we've seen working with organizations in operationalizing AI and driving ROI and they're really kind of divided into three main buckets as shown at the bottom of the slide, the planning problem, the data and technical problem and adoption problem. And I'll go into each one of them in more details and many organizations are often stuck at the, you know, the MVP or the minimal viable product or the proof of concept stage or have a list of AI opportunities, but they don't know really how to go from there. So that's where these three problems, if tackled can actually help organizations deliver value with AI sooner and in a more scalable manner.
The planning problem is going back to the idea of organizations can see a lot of, you know, AI tools out there, not necessarily knowing where to start, not having a methodological way to kind of get started. So this is where planning or a strategy would really come in and help kind of put that structure around that governance around how and where we can leverage AI if and where it makes sense for our business to drive ROI.
The data and the technical problem is around, there could be a ton of, you know, a ton of AI use cases out there, but determining which one is actually the right one for the organization. And then also determining what are the prerequisites from a data readiness and a technical readiness perspective as well. And then the adoption problem, which is absolutely key, which is as Olivia had mentioned, this is not just a technology problem, but it's a broader, it's AI is not just focused on technology but it's focused on people, process, data and technology.
So the adoption problem is all focused on the people. So often some organizations would go in and implement an AI tool and not necessarily build the necessary AI literacy or the AI adoption capabilities or enablements to enable the people and the process to ensure a sustainable deployment and implementation of the AI tools. Now we'll go into each one of those into a little bit more depth to explain a little bit more. So the first one, to tackle the first problem, which is the planning problem, the first thing to do is to enable planning. How to do that is to do it in a business first practically planned approach, in a business driven technology enabled approach.
Starting off with the business, which is looking at answering kind of four key questions along the way. One is what do we need to accomplish and can AI help? What I mean by that is what are the AI use cases that can actually help our organization and determining what's the feasibility for each one of them? And then after that, going to the next question, which is the why will my organization benefit from AI? And this is where we can help spend a bit of time to build the business case. What is the expected return on investment and what does actual return investment means in the realm of AI as well to help get buy-in to investment artificial intelligence.
Then after that, going into the third question which is how do we get ready to leverage AI? And this is where we look into AI readiness across not just technology, not just data, but also across people and process and preparing the processes and policies for artificial intelligence as well. And then last but not least, it is answering the question of when do we implement and what is our path to execution? And this is where we start looking into what is the roadmap going to look like? What are the high priority items that we need to tackle right now and what are some items that could potentially, you know, are more long-term strategic initiatives that can be pushed a little and defining the sequencing and the costing as well.
And I know here in the chat we have posted as well a link in terms of BDO's view on practical AI solutions. So please feel free to reference that for more details as well. In terms of the second, in terms of the second problem which I had mentioned was the data and technology problem. One thing that, you know, to tackle that it is to ensure technical readiness across data and technology. And actually a recent briefing that BDO attended with one of the provincial governments noted that 85% of POCs, while they are great ideas, have not gone past the POC state. And that could be for several different reasons. And we also have over here kind of a conservative view from Garner as well that says that through 2025, 30% of the gen AI projects will be abandoned after the proof of concept due to poor data quality. And that speaks to really the data and the technical problem.
So it's very, very critical to look at technical readiness, data readiness. How do we actually prepare our data? Do we have the right infrastructure ,do we have the right data quality? Are we looking at AI ready data to enable our AI initiatives. And data readiness, just want to kind of highlight this for AI, it's not something that is built once and for all. It's really something that's iterative. A continuous improvement type of mindset would be necessary. You can start and then kind of take it from there.
The third problem that I had mentioned was the adoption problem. So how to actually tackle that is by adopting and embedding adoption throughout. That's from, you know, from the strategy, from the design all the way to the deployment and the monitoring phases. And what you'll see on the left hand side here is drivers of results. Actually 10% is the algorithm, 20% is technology, but 70% lies with the people and the processes. So it's so, so important to look at people and processes. How are we going to transform our processes to best leverage artificial intelligence and how are we going to best inform our people not only from an AI awareness in terms of what AI can do, but also from an AI literacy perspective in terms of what integration like a more in depth understanding of the technology is also necessary. And the way that we kind of view artificial intelligence is that AI is like a vehicle, not the driver. So you can have the best car out there, but if the driver, the people, don't have a driver's license or don't know how to drive it, you're not going to be getting the best potential out of the technology. So people and processes are the core of a successful AI implementation.
Lastly here, I want to just wrap up here with some practical offerings for various starting points in terms of how you can actually get started and how BDO can support you on that journey. We have a list here of just initial jumpstart offerings that can, that we've, you know, we've worked on with a few clients, we've delivered to a few clients as well in the past. Starting with that can tackle those three key problems across planning to the data and the technology problem as well as the adoption problem to enable each one of those three. One is around AI strategy building kind of a holistic AI strategy to, you know, to plan what are the right use cases, what's the prioritization, what's the business case and support the delivery planning. There's also a data governance and management component 'cause going back to the idea of data enablement is so key to have the right data set in there and to make sure they have the right data quality and the right data governance in place. AI quotient and literacy, which is all just, you know, another word to say AI readiness across people, process, technology and data.
Microsoft Copilot Care Plus programming, collaboration with Microsoft, which is how do we actually best enable Copilot and how can we, let's call it, turn on Copilot and get the best use of it and define different use cases for different business units as well. End-to-end process review and optimization. So as I had mentioned, process is key along with people. So how do we actually look at how not only to modernize our processes, right? Which is just replacing one part of the process with AI, but how do you actually transform how we do things to get the most benefit out of artificial intelligence. And then the last but not least, jumpstart offering that we have is that AI bootcamp, which is focused more on the AI literacy component, on the people component ideation of what are the use cases that can be leveraged. We're happy to talk through more of these. If you have questions, please feel free to reach out to us after. And with that, now that I've given you the starting points, I'll pass it over to Eugene who will, you know, show us how to build on that, building on that AI foundation and what's to come.
Eugene Zozulya
Thank you Haya, can you hear me? All good. Hi everyone. Let's be more practical and let's talk about the practical use of that AI agents. And as we all know that generative AI has been a hot topic for the past couple of years and interesting enough that AI is evolving faster than the traditional technology. We've seen how fast ChatGPT was growing as a app. We've seen that the usage of that AI apps everywhere now and two main trends we've seen in the enterprise adoption are use of the multiple models and challenge in scaling this to production, Gartner predicts that by 2028, at least 15% of daily work decisions will be made autonomously by agentic AI. Just imagine that I read recently 15% of the automation and 30% of year to year of the human productivity will increase in the next few years.
We're just getting to this level of discussion and around that agentic AI the new frontier for the AI. So this concept is transforming how we interact with the technology in the future. And rather than simply automated tasks, AI represent the intelligent agents that can handle more complex workflow as we can imagine in the past. And this evolution, moving beyond that limitation of the traditional automation. I would like to spend some time to explaining what are agents and how they work. AI agent come in the different types ranging from simple to highly advanced. We're all familiar with large language models and we know that limitation, we know that they can answer questions, they can generate the natural language text, they're good in summarizing everything that we do and all the meetings and everything. We are using that Copilot and assistant to get this level of generative.
And we see that examples, images, audio, video that we see the examples that everywhere in the social media that we're leveraging as generation by agents. Then the next level of complexity is actually retrieval agents also based on the foundational models. But it's a next step to get that agent grounded in your data to get the answer your questions and what you need in the right time, in the right information. And the next level is the task execution, the actions. Because we know that limitation of the chat engines that they just can chat, can generate the text, can generate that video audio, but they cannot execute the task.
So the task, it's automation of the repetitive tasks, it's autonomous agent is the next level. And I would say that if we talked about that spectrum of those execution agents, starting with the simple task execution and then transforming to the multi-agent environment. And I will explain this a bit later and how this works, but the organization are at different points at this AI journey. Some are building powerful autonomous agents, but most of us, especially what we've seen in the public sector are on the initial stage trying out doing the POCs, started with the simple chat bots, but then how you execute, how you build this knowledge, how to use the tools that's already available for you. And the simple way to start is to start with Microsoft Copilot.
Copilot, not just a product, Copilot, it's a metaphor. It represent how AI can empower people to achieve more, helping each of us to be more effective, more productive and closer to our full potential. That I see that the biggest advantage right now even in the enterprise world, you have to have this in the protected environment. You have to ground this to your data, understand your challenges, understand your environment, and then when you enhance this, you use Copilot as AI for the agents. And Copilot is a personal assistant grounded in your work content like emails, meetings, everything related to the Microsoft Office and agent on other hand go beyond. They can act on behalf of teams, departments or entire organization. They can automate tasks, retrieving the information and even making their own decisions. Very interesting concept and let's understand, I love this example from the healthcare how that evolving of clinician assistance in the healthcare and Copilot is playing that crucial role of the user interface for the AI agents.
This is the real example and the real behavior of the drug in Copilot that brings the vision to life by combining the trusted voice dictation of proven drug and medical one. And with the ambient AI capabilities of DAX Copilot. In just the past month, it has supported 3 million patient conversations across 600 healthcare organizations in America. So let's deep dive in terms of how this works. Just imagine that you're clinician, you are taking a lot of patients, especially in the hospital environment with everything hectic, everything has to be very fast. So, and then clinicians simply can speak during the patient visit. Copilot will listen and automatically creates a structure of clinical notes. It can even capture the conversation with multiple people in multiple language, and turns them into the accurate summaries. And behind the scenes multiple agents work together to create real time notes, generate referral letters and provide after visit summaries so that seamlessly connect with all the Microsoft technologies as well. And I believe that the huge impact of this first network of agents that working in the healthcare, it's huge, it's voice first, AI powered and built to save time while reducing the burnout.
The recent survey in those clinics and hospitals that the average clinician is saving five minutes per encounter and most of reporting that reducing burnout and fatigue. So imagine that impact for the future. Let's take a look on what the agent looks like. And I believe that different organization creating the wide different agent varieties and you're trying to experiment, build up the first agent. And the most simple is use the foundational model, OpenAI models, Microsoft models or any others, Meta models and Google models. But you're starting to create those agents from the simplest to the most advanced that can act autonomously and they can self-learn in this journey. So they all comes with the different shapes and sizes and different complexity that those agents trying to solve. But one thing in common is that each agent is made up in the same components .So this is the knowledge that is grounded, data. Then the skills, the tools that agents are using, different actions, different triggers, different workflows. And then the next level is autonomy. How to do the self-learning, how to ground this into that trustworthy framework and make sure you control the outcome of this autonomy. And then exceptions planning and building up the multi-agent workflows and frameworks related to this. As the next level illustration of this at Microsoft we believe that AI isn't just a cool experiment, it's powerful tool to drive the real impact. And to help organization get the most of their AI investments, we're focusing on the four key pillars. Enrich the employee experience, is AI making your people more productive and engaged? How you measure the outcome and you know, happier employees become you get the better retention and the better productivity to accomplish tasks.
And simple use cases examples could be AI HR Copilot agents or knowledge retrieval agents. Then the next level, when you interact with your citizens, with your customers, reinvent customer experience and engagement, are you connecting with more customers in more meaningful way? AI helps to personalize experience at scale. And examples could be the citizen services agent always on support agent or even simple multi-language, multi level of document assistance in different parts. How you reshape your business processes, are you thinking how work get done better in your environment? AI can streamline your operations and automate repetitive tasks.
Some examples that we work with our customers, RFP tender review agents, imagine that how many tenders you run, how many tenders you run in the past, how to analyze all this information. And then most important part I think for public sector that I had discussion with is compliance monitoring. Just imagine like how many compliance policies and rules you have, you need to monitor all this. And then the next level, when you have this environment ready, how you bend the curve of innovation? Are you using AI to build and launch faster? The use cases, public safety civilians for example, how you monitor that video feed, the audio feed, how you monitor the compliance part.
Recent announcements from Microsoft, we did, Microsoft build is agentic DevOps and I'm working with some central public sector customers that we just reinventing like how DevOps works in the future, how to modernize the AI applications, how to enable the innovation at scale in these organizations. So we explore how that agents drive overall real impact. And then if you start with the simple productivity, what's the out of the box available to you? So with Microsoft 365 Copilot, Copilot Chat and Copilot Studio, you can boost the productivity and build custom AI agents using the low-code, no-code tools for the citizen development for adjustments. So these agents are automating tasks, retrieving the information and act across the Microsoft environment, across the different tools, Teams, Outlook, SharePoint, OneDrive, and grounded in your organization, data protected.
But just imagine that we're going beyond that part and we building that out of the box agents, for example, the IT help agent, IT help desk agent, employee onboarding agent, case management agent, team navigator agent. We're using that Copilot for sales internally as well. And then when you go to beyond those out of the box or low-code, no-code agent, we have all the tools available for the native application building up. So those tools are Copilot Studio to adjust your configuration part. Out of the box Copilot, out of the box GitHub Copilot as well. When you are using that more going into that, how to develop and deploy this at the huge scale in production. And at this point you're starting to rely on the platform. Azure AI Foundry is actually that the foundation that will help you to build those agent services. And with Azure AI Foundry and AI agent service, we are delivering the AI application at scale. So this is like application server for the future for the AI and this is enterprise ready. So your company data is secure, governed and well managed. We offer connectors to all your systems and breaking down the silos in your organization. Then you get the latest model and foundation to orchestrate different agents and different workflows in your organization.
And most importantly, you build a secure AI infrastructure around this. And as I mentioned, that Copilot is playing that user interface for your AI agent. And all of this built with the zero trust approach. So your AI agent are secure by design, from data to deployment side. And this is transferring the next speaker, Bill will talk about the actual application and maturity models for the future.
Bill Syrros
All right, thank you everyone. You know what, let's zoom out. I think a lot of people are hearing about the different tools and they're hearing about AI and how you should be using it and all the wonderful things around how powerful it is, but I just want to focus on like, we're just going to zoom out a bit. Here's what you really need to know. The tools out there are amazing and Microsoft has a ton of them. Their approach is second to none, but that's not, what that doesn't mean is that, that that's the only tool they're using. I know at BDO, and I know in the government people want to use AI tools and you're hearing a lot about ChatGPT and I know for sure people are using them on the side for both professional and personal reasons. And so using AI is really, now the way you have to see it is, it's a professional standard.
So as a national AI leader, I go across the country and I speak to a lot of people, a lot of organizations, a lot of associations, a lot of government organizations. And I say, listen, this is all about a professional standard that we now have. And it's not an arduous standard, it's not something we should see as negative, we should see it as incredibly positive. The average IQ for a human being is a hundred. When we start using these powerful tools, it's like we have augmented our intelligence by an extra 70 or 80 points And I think it's important 'cause as people are talking about AI around you both in government, in private industry, at the dinner table with your children, people are getting smarter.
There's no such thing as not having reference points and solutions and sources for information and being able to answer all those things you've always had. And we talk a lot about applying AI in the workplace, but you know, I think as we said earlier, adoption is where we need to be 'cause adoption helps bring on the productivity gains that this country needs. And so we are seeing adoption, I'm super optimistic because I'm hearing about it from all my clients and I'm certainly seeing it from the government. I see things like Can Chat and Agpal and the translation bureau's work, and I meet with government organizations and I see that, you know, it's certainly not, they're not scared of it. They're actually being forced to take it super seriously because people, like everyone knows how powerful it's becoming.
And so as they start using it and they start figuring out that this isn't Google, right, this is a human like natural language process where when you do speak to it, it responds to you like a human. And so when you're able to speak to your files and speak to your proposals and speak to your presentations and ask it for advice and ask it for improvements, it really changes the game for all of us because guess what? It's the first piece of technology that we've been given that actually saves us time. So I think what you've heard a lot today is a lot of different tools, and I'm sure maybe you're thinking, oh my god, more complication to my life. But the reality is when you use Copilot and you use other AI tools, it ultimately augments your capabilities and it augments it to the tune of 80 plus points. So, you are working faster, more efficiently, more productively than anyone. And so what you're going to have is you're going to have a situation where there are people around you using AI and then people that aren't. And there'll be a significant difference in how they're doing their job in the future. And you're going to really start seeing that over time.
At BDO, I'm just going to skip through things because I don't have a lot of time, but at BDO we were like customer zero. Like we rolled out Copilot across our entire organization and then we rolled out some ChatGPT licenses and then we developed data platforms for our clients and then we started building on top of Azure products to provide AI tools for all our staff. And now we're taking it to the next level, which is how do we build a scalable AI infrastructure with governance and regulatory policy implemented and security and all these things that we're going to need. So that all of the federal government and all of BDO, when I say BDO, we've got four lines of business, 5,000 people, 128 different services. How do we scale up so that AI is used efficiently? And it's the same message, it'll be the same message for you in your organizations.
The tools are easy to use, people are going to demand them. And as we start building these AI agents that you've been hearing about, there's two types of agents. There's the assistants, you know, which are the ones that are kind of helping you build your presentations and create your emails or whatever it is you're using for, by the way, don't feel guilty about it because it's making your emails better, it's making your presentations better. You're ultimately going to have the last say on things, but it's important for you to realize that it just allows you to do things faster. So as you sort of continue to use, you're going to start seeing use cases throughout your organization because you're becoming more familiar with it. So what I want to end this with is, for me personally, it's my desktop. It's my workspace, it's what I do everything in. And it's an incredibly exciting piece of technology. More exciting than the internet. How terrible it was when we were doing, when we had to access it by dial up. Imagine, we have access to the world's data and knowledge all at our fingertips. So I'm going to leave it at that. I want to thank you all for tuning in today and I'll hand it off to Laura.
Laura Spriggs
Wonderful, thank you. And thank you all for the wonderful presentations that you've made. We have just a couple of minutes, so what I think I'll do, there were a few questions that came through to some of the presenters. I thought that I would maybe throw a couple of them out and I'll look to some of our presenters today to respond to those questions. And Bill, the first one, I'd love to send your way. And the question was, if you were advising a senior government official, a deputy minister, assistant deputy minister who happens to be skeptical about AI, we know the Prime Minister is really keen on moving forward with it, is promising it's going to deliver efficiency and productivity, but the senior official is still quite skeptical. What would you say to them to help to shift their mindset?
Bill Syrros
Well, I think the adoption rates are increasing dramatically. So as those adoption rates increase dramatically, you are starting to see this evolution of how AI is starting to provide productivity to organizations. And it's important for, you know, we cannot delegate down as leaders. AI isn't something that is to be delegated. If we as leaders don't learn, understand, and use it professionally as a standard, then it's going to be hard for us to break that barrier of skepticism because AI isn't anything for the skeptics, really. This is real, this is the real deal.
Laura Spriggs
Yeah, and it's big and it's here to stay and it's going to continue to evolve. I've just been sent a note that we're unfortunately we're going to have to wrap up. That is all the time we have for today. I'd like to thank everyone for being here and also thank my BDO and Microsoft colleagues. This session was recorded. So for all of you who have joined us, you will receive directions on where to find the link to the recording. You'll notice that there is a QR code on the screen. You can use that if you'd like to enlist a demo of some of these wonderful capabilities that Bill had talked about that we have created. And in the email that you receive with those directions, you'll also have the contact details for all of our presenters today. And we are at two o'clock sharp now. So I would like to wish you all a wonderful rest of day and thank you very much for joining us.
Eugene Zozulya
Thank you.
Bill Syrros
Thank you.
Haya Elaraby
Thank you.
(calm music)