Introduction

well good evening all many thanks indeed for having me I've been coming to the mindstone events for uh on and off for for a year now so it's particularly exciting and particular privilege to uh to join you this evening uh yes so my

name is Pete I'm a principal consultant uh in the advisory team at digital consultancy called um

ffsp and since um coming on a year now I've been working with a number of different clients anything from sort of small private organizations to really large kind of uh public sector organizations

uh on launching co-pilot for Microsoft 365 um and i' just like to take this time to to share some of those experiences some of the benefits of the technology uh but also some of the challenges challenges that it pre

presents um but yeah just before I begin

just kind of quick overview of my my own background because it's this kind of work I'm particularly excited on because quite a good sort of good coalesence of all of my different experience over the years CU I I started my career out as a developer focused on kind of the Microsoft stack before moving into

Overview of Microsoft 365 Co-pilot

start off by just kind of a quick introduction to co-pilot for Microsoft 365 now many of you would be aware that

Microsoft offers a wide number of different co-pilot and generative AI um Solutions and I I work I've done work across a number of these uh for instance

you've got co-pilot for Dynamics 365 as you'd imagine a feature of Dynamics 365 that proide provides assistance across business functions such as sales

marketing and supply chain uh just recently started a project on rolling out Microsoft uh security uh co-pilot uh which as you imagine is is a um security

solution and really kind of helps to detect and investigate and respond to to cyber threats and then I imagine a lot

of you here are familiar with GitHub co-pilot obviously a feature of GitHub uh that helps developers to write code

but for this presentation uh very much been focusing on where my attentions has been focused recently which is around co-pilot for

Microsoft 365 now that becomes a bit of a mouthful after a while so I'll just use co-pilot going forward uh but hopefully you know you you'll know what I

mean yeah so in a nutshell um co-pilot is an AI powered Pro productivity tool that helps with various different tasks across the Microsoft 365 applications

which you might might use as part of as part of your daytoday such as writing email oh oh yep thank you uh such as

writing emails creating reports scheduling meetings and generally kind of generating insights across your organizational

data and really this can be thought

Key Components of Co-pilot

about in three key components so firstly there's the large language model and in this case obviously is open ai's GPT

models um GTP GPT for Turbo believe is uh at this stage but they're constantly upgrading it as as as they as they come out and be made available to to

Microsoft and aure and as we understand as we know this is the type of AI that can understand summarize and predict and generate content one of the kind of big

kind of key differentiators for co-pilot Microsoft 365 versus more kind of consumer facing tools is co-pilot uses

grounding in the graph and as process of connecting the large language model to your organizational data which is

available in the Microsoft graph and that could be your emails your chats and teams your documents anything that you have kind of permission at to access and

that provides much more relevant personalized uh suggestions based upon that Contex and context and your

context um and then thirdly co-pilot is integrated within the Microsoft 365 apps so as I mentioned the the apps that you use uh as part of as part of your daytoday

and it's the combination of these three things that provide kind of great um potential benefits for users and and and

organization so things like increase productivity so automating time consuming tasks uh enhancing cre creativity suggesting ideas inspiring you with new

examples and then one key one was around learning support So discovering new topics skills and interests or deepening your existing knowledge and expertise in

particular areas and then one I think quite key uh benefit for a lot of users and organization is better discovery of of data so understanding your natural language queries then it's able to much better retrieve information from from different data sources in Microsoft 365

Co-pilot in Action

so for instance as you can see here uh this is co-pilot in like in situ in word and you can use it to draft a document

based upon a rough outline or from data in another document or spreadsheet so say you have a spreadsheet that's got a

report and you want to kind of pull that into a written word word document it gives you really good stter for 10 in

those kind of use cases in teams you can use it to catch up on a meeting by taking the transcript

gives you a summary that then can ask questions based upon the transcript of that session whether it be after the

session or indeed during the session and I think for many users that I work with and as we start to kind of roll it out

and provide trials this is is really one of the kind of the first key kind of benefits that we see across many user

bases uh and then M365 chat this is available in the browser but also teams but it's a sort of central place to

access and interact with co-pilot and really helps you to find and create and content and generate in

insights soorry let me go back a moment so for instance you've got an up a upcoming meeting you can ask it to

prepare based upon the information that you've been providing for that meeting or any other relevant context if you've

got a particular bid management process and you want to understand that or what are the kind of the key roles and

responsibilities it's really good at kind of finding that relevant information and pulling it together into

into a useful uh response again sort of one of the key kind of differentiators from um the

publicly available co-pilot or chat GPT is then you can look into your files and pull out key themes and then you can go

further and say okay based on those key themes relating to a particular uh senior Management's accountability and

then summarize that into an email and then provide guidance around making it short and suitable for Senior Management

and then you've got your email ready ready to go and I think yeah in the uh in the in

the context of of Outlook and I've just recently come back from a good long three we holiday for instance and

hundreds and hundreds of emails really useful at that point to be able to kind of summarize the key points and help me

prioritize which which emails to to to respond to and when I do respond to it it's able to take the context of my

email Fred and provide a relevant response and or provide coaching around a better better

response in PowerPoint um really good and useful to create a kind of starter starter for 10 Again based on a kind of

given Word document or a sort of a prompt around a particular topic I didn't use it to create this

presentation um it's the design sometimes left uh isn't isn't isn't brilliant but it does give you that kind

of good started for 10 to then build off of and then you can ask it to add in relevant stock

imagery and then there's been some users that have had some sort of good experiences in uh co-pilot in Excel it

is in preview so it's definitely one of the more immature areas but and I think it provides a lot of support for those

users who aren't necessarily expert in in in Excel but it helps you to analy and explore data and then pull it out

and visualize it in in uh in your charts so um as well as kind of helping customers to uh to to roll it out across

their organization I myself had co-pilot for um uh coming up to a year now and I thought i' flash up just a few kind of

examples across my across my day where I might use it across um across the different across the different

applications

Impact and Benefits of Co-pilot

okay um so Copa has actually uh as of a couple of weeks ago I believe been um been available uh for a year not it um

to to the kind of the Early Access program but at the end of last year Microsoft published a rather kind of comprehensive report on the impact that

co-pilot was having on elements like productivity creativity and the time saving employees and some of the key highlights

was around sort of 70% of users reported they were more productive about the same said they saw an improved qual quality

of work and overall they're about 30% faster in a series of tasks I just recently compl completed a business case

for expanding it across uh an organization and um had some really good sort of uh responses where people were saying they were saving up to up to 7

hours a month through some of this kind of function functionality so a really good um return on investment for that

organization and those particular users so by empowering employees to embrace a generative AI organizations can really help to drive Innovation and

competitiveness and this kind of brings me on to another kind of key benefit of of of co-pilot and and as not so recent now but a few months ago I think sort of

sales FS P published this um report which highight highlights that companies really need to provide employees with that secure sandbox to explore and adopt

generative AI otherwise they will just go off and likely find a way that is tools that aren't properly governed and could potentially put their business and

and their IP IP at risk and I think as such co-pilot enables your employees to embrace and explore the potential of AI within that

trust boundary within that Microsoft 365 tenant and I think that's what's kind of excited me so much about this work is

for a lot of organizations this is their first 4A into generative Ai and it's because of these these mechanisms and these this trust boundary that they're

able to that they felt they were able to do that so yeah as this high level architecture flow highlights the information with your prompts the data

they retrieve the generated responses are all within that Microsoft 365 service and indeed none of that information is gone off to train any

foundation llms including the ones Microsoft maintain maintain themselves however and if you've explored kind of co-pilot yourself that

your organiz or your organizations have uh this will be one of their key considerations I'm sure um but for any Implement organization implementing

co-pilot it relies heavily on the quality and the quantity of the data that's available within your Microsoft 365 tendency therefore it's really

important to ask um have you got good data management and governance strategy in place um and ask yourself questions like do you know where your data lives

do you have uh can you identify old data if it brings back stale content uh it can be kind of it can be a real Challenge and crucially can people

access sensitive data that they shouldn't and I think that last one is a kind of particular concern for a lot of com lot of

companies um because co-pilot doesn't give you access to content um that you don't have permissions to but it makes it that much

easier to kind of get to get access to it um however there are lots of kind of controls around security and privacy of your data that are available at every

stage of your co-pilot usage uh and organizations to really review these tools and kind of get them in place it's a big part of kind of what we've done

the work with different organizations and I think also just kind of depending on your data governance posture you want to be quite sort of

selective around the candidates you choose for trials as as as you roll it out so on the subject of data um out of

Implementation and Data Governance

the box co-pilot is limited to the data that's available within your Microsoft 365 tency now however most organizations

and certainly the ones that I'm working with use separate file stores different tools as part of their workflow and potentially really limiting the opportunities for some of those benefits that I highlighted

so but through co-pilot studio and plugins and graph connectors you can bridge the gap between co-pilot and its native functional functionalities and

extend it out to kind of external data sources and thirdparty platforms and really create that kind of centralized place to kind of access uh this data and

and this functionality and just to kind of give you an example this is uh a ticket from JRA being pulled into that Microsoft 365 chat that I that that I showed before

so there's a number of kind of basic technical requirements to consider but however outside of that data governance

piece that I mentioned deploying and activating co-pilot for users is really quite

straightforward but with like like with so many te Technologies and I think it was kind of highlighted before in the in the first talk to kind of really realize

the benefits and build trust in this technology requires quite a thoughtful and responsible approach to driving adop adoption and proficiency this is true of I said like

lots of Technologies but co-pilot and and AI War its own particular set of considerations I think also currently at $30 per user per month for a lot of

organizations that's quite a significant investment so making sure you get a good Roi on that investment uh and getting it to the right employees and ensure

they're making the most value out of it is is really

important how long have I got

wrap it up okay cool yeah so many of you familiar with the different kind of change and adopt adoption approaches and

I just wanted to kind of highlight some of the kind of key learnings from my own experience when rolling out uh rolling out out

Strategies for Successful Adoption

copilot and I think when you kind of consider your strategy and your vision for co-pilot it's really important to

assemble a diverse and cross functional working group comprised of individuals different expertise and yes that might be AI but also ethics that's legal compliance and bu business strategy as I

was saying before for so many organizations this is the first time they're really rolling out this technology at scale so kind of getting these good practices in place is is

really important I think similar and governance as well but and similarly by being quite selective and who you're kind of you're providing it to uh

keeping HR AB breast of of these decisions is is simly really important yeah likewise like Liv with with your Champions pulling from a diverse cross functional user base uh so

that they can unearth and understand their use cases and then share them again uh really

important um yeah just quickly on kind of training um accessing co-pilot is actually kind of pretty straightforward

and there's a lot of training content out there um so focusing your attention on really building proficiency around prompting I I think is one of the kind of key the key learnings and ensuring

that your users don't feel uh lost with their first prompt but continue to are continue to ask and prompt

co-pilot um for for what they're looking for and once you've got those prompts and found found really great useful prompt ensuring that you share it ac

across your other teams and I think when it comes to your comms channels making use of the Microsoft 365 tools that they

already have such as team create ask me anything channels use your SharePoint sites to hold your resources and use use via

engage yeah and I think finally just kind of when it comes to sort of measuring success uh one of the key elements we've used is uh is user surveys and we've got some really good

templates if anyone is interested are happy to happy to happy to share those uh which creates a really good Baseline

to then kind of track adoption and benefits as you progress through uh through your

trial um you can also use uh Microsoft provide a co-pilot dashboard which is available via Viva insights that is just another way of ensuring you track usage patterns and potential Time Savings and

productivity

Conclusion

gains yeah so just to finish up um

certainly from my experience as a personal user and rolling out across a number of different organizations can be a really has the potential to be a really really powerful tool for employees and organizations improving productivity innovational learning if implemented and used effectively

requires a really sort of strong kind of data Readiness and governance um program to ensure quality and security um and it also does require um quite a careful change and and adoption strategy to ensure that your users are using it to the best of its abilities

just finish up by saying thank you ever so much um we are hiring um for a senior AI consultant an ml engineer and a data architect so please do get in touch if if that's of interest to you I don't

know if there's time for questions but I'm very happy to answer them now we'll take them off in the break if we can take the break would be great absolutely thank you very much