Unlocking Microsoft Copilot at Scale
Key Points
- The CTO of a 6,000‑person firm realized they’re spending six‑figures on Microsoft Copilot yet only using it for email, prompting a deep‑dive guide on unlocking its full potential.
- The video outlines practical use‑cases, required organizational shifts, and an overview of all 12 distinct Copilot products so teams can move beyond basic tasks.
- Despite 90% of Fortune 500 companies adopting Copilot, most users aren’t tapping its broader capabilities, likening the situation to driving a Ferrari in first gear and missing out on intelligence‑on‑tap benefits.
- Real‑world case studies—including Vodafone’s rollout to roughly 68,000 employees—demonstrate concrete workflows, specific prompts, and the potential to save several hours per week per employee.
Sections
- Deep Dive into Microsoft Copilot Rollout - A comprehensive guide that outlines how enterprises can move beyond basic email assistance to fully leverage all 12 Copilot products, detailing use cases, necessary organizational changes, and strategic implementation steps.
- The Copilot Naming Chaos - The speaker outlines the confusing array of Microsoft‑branded Copilot products—free Windows Copilot, paid Pro tiers, Microsoft 365, GitHub, Security, Dynamics, Power Apps, and Copilot Studio—with wildly different purposes and pricing, illustrating how the identical branding leads to user bewilderment.
- Microsoft 365 Copilot Feature Overview - The speaker outlines how the Pro plan offers individual perks like AI‑generated media, faster writing assistance, and student priority, then contrasts it with the enterprise‑focused Copilot that accesses workplace data and provides Outlook email summarization and draft replies.
- AI Autocomplete Accelerates Repetitive Coding - The speaker argues that tools like Copilot boost productivity on repetitive tasks by up to 55%, while noting a rapid split in the developer community between those who augment hand‑coding with AI suggestions and those who rely entirely on AI‑driven agents.
- AI‑Driven Email Management & Writing - The speaker shows how AI can triage unread emails into actionable prompts and rapidly generate documents such as blog posts, dramatically cutting the time needed for both tasks.
- Copilot Enhances Meeting Summaries & Research - The speaker explains how Microsoft Copilot can automatically generate meeting minutes, summarize key discussions, and conduct web‑based research, while stressing that fully exploiting these features requires a cultural shift within the organization.
- Reimagining AI‑Driven Workflows - The speaker explains how viewing AI as a pervasive intelligence—rather than a single‑purpose tool—enables end‑to‑end automation of tasks like email management, blog creation, and report generation, dramatically cutting time and effort.
- Creating a Shared Prompt Library - The speaker outlines how teams can establish a collaborative hub for reusable AI prompts—organizing, tagging, and pinning concise prompts for marketing, sales, and legal tasks to streamline workflow.
- Marketing-Product Alignment via Copilot - Marketing leverages Copilot to analyze social‑media feedback, prioritize themes, and dynamically align them with the product roadmap, streamlining collaboration with product managers and reducing friction.
- Enterprise AI Adoption and Job Fears - The speaker outlines how a large company is rolling out AI through internal support sites, role‑specific training, and CEO‑led messaging that addresses employee concerns about job loss, citing expert evidence that AI has not yet caused widespread attrition.
- Leadership, AI, and Copilot Vision - The speaker stresses the crucial role of strong leadership in adopting shared machine intelligence and highlights Microsoft Copilot’s Vision feature, which lets users screenshot errors or PDF tables to receive plain‑English explanations and automated conversions like Excel output.
- Copilot Success Stories in Law and Healthcare - The speaker highlights how Microsoft Copilot dramatically cuts task times—turning a week‑long contract review into hours and shrinking doctors’ discharge summary work from 30 minutes to minutes—by using precise prompts tailored to legal and medical contexts.
- Urgent Call for Copilot Adoption - The speaker urges businesses to quickly adopt and master AI Copilot—highlighting its fast‑becoming status as a table‑stakes competitive advantage, the need to measure ROI, and offering immediate workflow guidance.
Full Transcript
# Unlocking Microsoft Copilot at Scale **Source:** [https://www.youtube.com/watch?v=cW9eaEcx6OY](https://www.youtube.com/watch?v=cW9eaEcx6OY) **Duration:** 00:43:59 ## Summary - The CTO of a 6,000‑person firm realized they’re spending six‑figures on Microsoft Copilot yet only using it for email, prompting a deep‑dive guide on unlocking its full potential. - The video outlines practical use‑cases, required organizational shifts, and an overview of all 12 distinct Copilot products so teams can move beyond basic tasks. - Despite 90% of Fortune 500 companies adopting Copilot, most users aren’t tapping its broader capabilities, likening the situation to driving a Ferrari in first gear and missing out on intelligence‑on‑tap benefits. - Real‑world case studies—including Vodafone’s rollout to roughly 68,000 employees—demonstrate concrete workflows, specific prompts, and the potential to save several hours per week per employee. ## Sections - [00:00:00](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=0s) **Deep Dive into Microsoft Copilot Rollout** - A comprehensive guide that outlines how enterprises can move beyond basic email assistance to fully leverage all 12 Copilot products, detailing use cases, necessary organizational changes, and strategic implementation steps. - [00:03:08](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=188s) **The Copilot Naming Chaos** - The speaker outlines the confusing array of Microsoft‑branded Copilot products—free Windows Copilot, paid Pro tiers, Microsoft 365, GitHub, Security, Dynamics, Power Apps, and Copilot Studio—with wildly different purposes and pricing, illustrating how the identical branding leads to user bewilderment. - [00:06:36](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=396s) **Microsoft 365 Copilot Feature Overview** - The speaker outlines how the Pro plan offers individual perks like AI‑generated media, faster writing assistance, and student priority, then contrasts it with the enterprise‑focused Copilot that accesses workplace data and provides Outlook email summarization and draft replies. - [00:09:49](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=589s) **AI Autocomplete Accelerates Repetitive Coding** - The speaker argues that tools like Copilot boost productivity on repetitive tasks by up to 55%, while noting a rapid split in the developer community between those who augment hand‑coding with AI suggestions and those who rely entirely on AI‑driven agents. - [00:13:10](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=790s) **AI‑Driven Email Management & Writing** - The speaker shows how AI can triage unread emails into actionable prompts and rapidly generate documents such as blog posts, dramatically cutting the time needed for both tasks. - [00:17:16](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=1036s) **Copilot Enhances Meeting Summaries & Research** - The speaker explains how Microsoft Copilot can automatically generate meeting minutes, summarize key discussions, and conduct web‑based research, while stressing that fully exploiting these features requires a cultural shift within the organization. - [00:21:05](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=1265s) **Reimagining AI‑Driven Workflows** - The speaker explains how viewing AI as a pervasive intelligence—rather than a single‑purpose tool—enables end‑to‑end automation of tasks like email management, blog creation, and report generation, dramatically cutting time and effort. - [00:24:47](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=1487s) **Creating a Shared Prompt Library** - The speaker outlines how teams can establish a collaborative hub for reusable AI prompts—organizing, tagging, and pinning concise prompts for marketing, sales, and legal tasks to streamline workflow. - [00:28:17](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=1697s) **Marketing-Product Alignment via Copilot** - Marketing leverages Copilot to analyze social‑media feedback, prioritize themes, and dynamically align them with the product roadmap, streamlining collaboration with product managers and reducing friction. - [00:32:05](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=1925s) **Enterprise AI Adoption and Job Fears** - The speaker outlines how a large company is rolling out AI through internal support sites, role‑specific training, and CEO‑led messaging that addresses employee concerns about job loss, citing expert evidence that AI has not yet caused widespread attrition. - [00:35:33](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=2133s) **Leadership, AI, and Copilot Vision** - The speaker stresses the crucial role of strong leadership in adopting shared machine intelligence and highlights Microsoft Copilot’s Vision feature, which lets users screenshot errors or PDF tables to receive plain‑English explanations and automated conversions like Excel output. - [00:39:10](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=2350s) **Copilot Success Stories in Law and Healthcare** - The speaker highlights how Microsoft Copilot dramatically cuts task times—turning a week‑long contract review into hours and shrinking doctors’ discharge summary work from 30 minutes to minutes—by using precise prompts tailored to legal and medical contexts. - [00:42:29](https://www.youtube.com/watch?v=cW9eaEcx6OY&t=2549s) **Urgent Call for Copilot Adoption** - The speaker urges businesses to quickly adopt and master AI Copilot—highlighting its fast‑becoming status as a table‑stakes competitive advantage, the need to measure ROI, and offering immediate workflow guidance. ## Full Transcript
Last week, I had the CTO of a 6,000
person company send me this DM. We're
paying six figures for Microsoft Copilot
licenses and almost everyone on my team
uses it for writing emails. That's it.
What the hell are we missing? So, this
is a deep dive on Microsoft Copilot.
This is a complete guide to how to roll
it out and make the most of it. I have
been asked for this particular guide for
months and I have taken my time because
of how complex it is. Strap in, stay
tuned, get out your notebooks. This is
going to be a long and complicated video
and it is totally worth it because you
know why Microsoft is in almost every
enterprise out there. The chances are if
you can't use Chad GBT, if you can't use
Claude, you got to use Copilot. So what
do you do? I get that question a ton.
This is the answer. I want you to walk
away with a clear set of use cases you
can adopt. A clear sense of the
organizational changes you need to go
through to enable Copilot to get beyond
the email use case and a clear sense of
which of the 12 different Copilot
products to pick. I'm not kidding you.
There are 12 and we're going to go into
each of them. Okay, here's where we
start. Something like 90% of the Fortune
500 has copilot. 90%. That's why this is
such an important video. That's why I've
taken my time on it. It's got to be
correct. Microsoft's own data though
shows that most people aren't using the
full capability set of Copilot. Now,
look, if you're not using the full
capability set of Microsoft Word, which
I know I don't, that's okay. It's
Microsoft Word. It's there to help you
type stuff. But if you're not using the
full capability set of Copilot, it's a
big deal because Copilot is literally
intelligence on tap. And so if your
workforce isn't using it, if your teams
don't feel good about using it for
things that are more than email, you're
the ones missing out. Companies are
literally paying for a Ferrari and they
are driving it to the grocery store in
first gear. That is what is going on
with Copilot. So I've spent months
looking into case studies. I've spent
months talking to actual folks who are
working on C-Pilot in these different
enterprises that adopt it. And I've
interviewed users who are individual
contributors. Users in the engineering
departments, interviewed folks in
product management, sales. I want to get
a sense of how Copilot is actually being
used and then how it could be used. And
this is the fruit of all of that. This
is me finally sort of sharing a little
bit of what I've been learning so you
can benefit. So you're going to know
exactly which copilot you need. You're
going to save you'll be on track to save
several hours a week if you go through
this video and actually apply it. These
are going to be real workflows. I'm
going to give you specific prompts and
I'm going to show you some case studies
including how Vodafone rolled this out
to tens of thousands of people. I think
it was 68,000 people successfully. Okay,
with all of that for intro, let's jump
into part one. I want to talk about the
12 different flavors of Copilot because
it is completely confusing and I get why
people don't understand it. Frankly,
Microsoft did something absolutely
insane. They took the name Copilot and
they slapped it onto everything. I'm
going to read you the real actual
co-pilot list and I want you to tell me
which one is the best co-pilot.
Microsoft Copilot. This is free. It
comes builtin to Windows. Copilot Pro.
It's a $20 a month power user version.
That doesn't mean it's actually for
professionals in a work setting. Copilot
for Microsoft 365.
It's a $30 a month one and it looks like
it's tied into companies, right?
Microsoft 365 copilot chat. Somehow this
is different. and free for business
accounts. Both of these are tied to the
Microsoft 365 business product line.
GitHub Copilot. This one's for
developers and tied to the GitHub
product line. Everybody got the C-pilot
branding. Are you getting the idea?
GitHub Copilot for business. Different
from regular GitHub, about double the
price. Security C-pilot. This one costs
a lot of money. I don't understand how
you can name the product the same thing
and have it cost so much different. You
can have something that is free in
Windows and the same name. Copilot also
can cost $35,000 a year minimum. Copilot
and Dynamics 365 for sales. It's built
into your CRM. Copilot and Power Apps.
It's for building apps. Do you get the
idea why it's taken so long to make this
video? People say, "How do I use
C-Pilot?" In my head, I'm like, "Which
of the 12 are you using?" Windows
Copilot. It's different from Microsoft
Copilot somehow. Copilot Studio. It's
for building your own co-pilots. So, we
can have more. And there's more. There's
variants for customer service, for field
service. There are companies that have
bought C-Pilot Pro thinking it was the
business version because it makes sense
because it says pro, but it's not the
business version. There are companies
that have bought Microsoft 365, but they
don't have the right base licenses. So,
Copilot doesn't work properly. There's a
dev team. Dev teams have bought C-Pilot
for business when they just need
individual licenses. The product surface
confusion is as bad in Microsoft with
Copilot as it is with chat GPT and
picking model names. They're both wildly
harming the capability of the product
and the ability of the product to
deliver value by being unable to name it
correctly. So everybody's confused, but
smart companies can still get massive
advantage because fundamentally it is
intelligence on tap. I tease them for
the naming and the naming is confusing.
But we're going to go through each of
those products and we're going to talk
about what they do so that you
understand them and can figure out what
you need. So let's dive in. Co-pilot
decoder. Like I want you to get through
this and basically figure out which one
is right for you because the difference
is thousands of dollars per employee per
year. Like it's a big difference if you
get this right. So let's jump in. Number
one, we talked about Microsoft Copilot
Free. So it's built into Windows and it
is designed for consumers. Probably if
you're in business, you're not using
this. As an example, you can tell it to
turn on focus mode. Play jazz music.
Summarize this web page in Edge. Help me
write a birthday message for mom. This
is the variant of Copilot that is ending
up in the ads because this is the one
that average consumers have on their
Windows machines in the booknook at
home. Explain this screenshot. Right?
That's another example. It doesn't work
files. You should not put confidential
information into it. It does not have
prioritized GPU access. So, it gets slow
during peak times. If you have Windows
11, it's already there. Okay, let's move
up a little bit. Copilot Pro. This is
for people who are individuals who are
not professionals at work, even though
it's named Pro. It's roughly on par with
the sort of chat GPT plus offering. So,
it's like 20 bucks a month or so. So,
you get priority access to GPT4 Turbo,
which is it's an okay model. It's not
that great. You get a 100 image
generations per day versus 15 on free.
So, if you want to make I guess a fancy
childhood book, I don't know. Uh, it
works inside office.com web apps, which
is a real help. Uh, and you get early
access to new features. So, as an
example of things you can do in Pro, you
can generate your social media images
there. If you want faster and better
responses across long documents as a
writer, Pro can work well for you. If
you're a student, you get priority
access during finals week when people
are using it, right? And so, it it can
it can make sense, right? Like if you
were working on final drafts for a
document and it saves you a few hours a
week cuz it can handle the load. That
adds up to real money. It easily pays
for itself. But let's keep going.
Copilot for Microsoft 365. I believe
this is 30 bucks a month and it's aimed
at the at at at the business layer,
right? At enterprise. So the key
difference is this one will see your
work data. It's tied into the Microsoft
365 ecosystem. That is the point. And
this is often the one that people are
talking about when they talk about
copilot at work. Not always. There's a
few others we'll get to. So specific
capabilities by app for this particular
flavor of copilot. Copilot for Microsoft
365. You can get the entire intelligence
experience in Outlook. I tease that
people mostly use it for Outlook, but
I'm starting there because that is where
people start. Summarize emails from John
from last week about project X. You can
easily do that. Get a draft a reply
accepting the meeting and proposing
Tuesday instead. You can get that in
five seconds, right? Very easy. If
you're in Word, create a project
proposal for X topic based on this
template, which is a DOCX file. Rewrite
this section to be more executive
friendly, right? That's a very PM thing
to say. I can say that because I'm a
product manager. Add a risks and
mitigation section using data from
risks. Risk.xlsx, right? Like let's
assume it's an it's an Excel's file. And
it does work in Excel as well. What are
the top three trends from the sales data
is something you can ask. It will tell
you like it grew this much etc. As long
as the data is clean. Create a pivot
table showing sales by region and month.
Instant pivot table which by the way as
someone who's had to suffer through
Excel the fact that people can now
create instant pivot tables. I we live
in a blessed world. I would I hated
creating pivot tables when I was in
marketing. It was just so much pain. So
it's really nice to be able to have
pivot tables that I can just type in and
speak and and up they come. Right? It's
very handy. PowerPoint create a fly five
slide deck from report.docx. This will
not be gorgeously formatted, but it will
do it. It will have some design. It will
have some slide transitions, which you
may hate or you may not hate, but it
will have them. And you can tell it to
add speaker notes based on the content.
If you're in Microsoft Teams, you can
say, "Hey, what did we just decide about
the budget?" And you can actually like
type it in or talk it in during the
meeting. And like you can get some kind
of a response from Copilot. You can ask
Copilot to generate meeting minutes with
action items, summarize the channels,
discussion. You will notice I am just
going through co-pilot 365 which is not
the fanciest co-pilot. I am already
giving you so much more than just email
here and we are just barely getting
started. Okay, let's go to the developer
side of the house. GitHub copilot a $10
a month individual. I think it's 19
bucks a month for business. The idea is
very similar to Cursor, very similar to
Windsurf. It writes code alongside you,
right? And so depending on who you talk
to, you're going to get people saying
that this makes them much faster, right?
is 55% faster on repetitive tasks.
That's the number I saw. I don't care if
it's 55 or 30 or 25 or or 75. The point
is it's a repetitive task, right? That's
the thing to call out. If it is helping
you to autocomplete your code faster,
then it is going to give you that speed
up. And in that sense, like you bring
the design, you bring the knowledge of
the code, you bring the ability to write
the code, and all copilot does is it
acts like a smart autosuggest and it
runs really fast. And so for folks that
write code by hand and that's the way
they do it, that's great. For folks that
are using more of the agentic
capabilities in cursor co-pilot like
this is not really going to work for
them. And so this is actually one of the
areas where I think the developer
ecosystem is changing extremely rapidly.
You still have developers that write
code by hand and they are finding it
better to write with an accelerated pen
so to speak or an accelerated keyboard
because they can get the co-pilot
generated auto suggestions but you also
have people who are moving away from
writing code at all and depending on AI
more maybe through vibe coding or maybe
they're real developers and they're
using agents to basically run pull
requests through or maybe they're
writing some of the code themselves and
then they're using agents for the rest.
The developer ecosystem is fracturing
really quickly and developers if you've
ever talked to them have strong opinions
about this. I am not going to have the
which development stack makes sense and
is future aligned conversation in this
particular video but we will have it
soon. For now that's what you need to
know about copilot for GitHub. It
basically helps you write code you
already know how to write faster.
Security C-pilot. This one is something
like four bucks an hour for compute. It
costs roughly $35,000.
It's for I know I'm not over it either.
This is like it's free. It's 10 bucks a
month. It's $35,000.
Really? That's the pricing? Anyway,
we'll leave the branding aside. So,
security teams with 10 plus analysts
might use this. And basically what it
does is it helps you to analyze
suspicious user behavior across
different logs. It helps you to write
incident reports. It explains malware
code. If you are at a certain scale, I
tease them. But like this is absolutely
worth the money. If you can explain
malware code in plain English and write
incident reports and you're operating at
true enterprise scale, pays for itself
like that. You can do something like say
investigate if user John Doe's account
is compromised and it will really run an
investigation. Or you can ask it to sort
of file write up and describe an
incident that comes across your desk and
accelerate the time to resolution. And
if you think about the millions of
dollars lost on incidents in secure
environments, it the ROI is off the
charts even at $35,000. And that is one
of the things that makes large language
models and AI hard for someone like me
to talk about. I have to talk both about
how individuals are using this in their
booknook with Windows 11 and also how a
security team at a Fortune 100 is using
a $35,000 edition with the same name
that is enabling them to get back online
two or three hours faster, saving the
company $10 million. Those are the same
named thing. All right, we got through a
bunch of the different co-pilots. I want
to talk about some specific workflows
that can help you deliver a win. So yes,
yes, I'm not kidding. We really are
going to start with email. If you've
never done Copilot before, this is
great. If you've done Copilot and you
can do email, we'll breeze through it
fast enough that you can get to the
other cool stuff, too. So, open Outlook
with Copilot. You can click on the
C-pilot icon, ask it to summarize your
unread emails. That's a great one.
Everyone has unread emails. Do you have
a zero inbox? I rarely have a zero
inbox. And so, you'll get answers. You
know, Ellis had a budget revision
request. Carol, she wants her client
meeting moved to 2 p.m., etc. Now, you
know what needs your attention. That's
value right there. If you have a long
complaint email from a client, just say,
"Hey, draft a professional response
acknowledging concerns, offer an x%
discount, and propose a new timeline."
You can do that kind of like business
engagement with your email inbox. And it
does save time. And so as much as the
CTO that started off this whole thread
in this video was complaining about the
fact that his team only uses it for
email, even in that case they were
saying they were saving, you know,
multiple hours a week. It wasn't
nothing. Let's move on to another use
case, document creation. Take take a
task like a blog post. Let's say you're
you know back in Nate's former chair,
you're you're a marketer. We need to
write a blog post about sustainable
packaging. You can just say, "Hey, give
me five trends in sustainable packaging.
Give me a sentence each." Okay, great.
Now create a detailed outline for a
thousandword article. Okay, great. Write
the first paragraph. Help me see your
hook. Okay, let's twe twe tweak it.
Tweak it. Okay, now I like it. Write
section one. Write section two. You're
going to be done with that blog post in
15 minutes versus 2 hours.
Traditionally, it's a huge savings for
SEO type document creation. I will
hasten to add it is extremely good at
document creation, which may lead you to
think it is extremely good at strategy.
Those are different things. If you want
a good strategic partner, that is not
the workflow to go through. You still
need to bring your brain if you are
doing strategic thinking. What co-pilot
is doing, it is accelerating your
ability to execute once you have
clarity. I'm just going to underline
that like three times. Okay, let's go
over to the Excel. Let's look at the
data analysis piece. Let's say like me,
you don't love all those Excel formulas.
You can be as simple and as high level
as you want. You can say, "Hey,
co-pilot, analyze this data." And you're
in a spreadsheet. You can say, "Tell me
what's interesting. Sales increased 23%
year-over-year. July was an anomaly with
a 45% spike. The western region is
underperforming." Honestly, it will give
you stuff like that. You might follow
up. Hey, why is July so high? Well, and
then it will show you what it knows,
right? C-Pilot might say, "Hey, column J
shows that there's a summer promotion
running for July entries." And then you
can say, "Great. Okay, create a chart.
Show monthly trends with a July spike
highlighted." and you get a professional
chart. That entire workflow used to take
me an afternoon when I was working
through marketing data anomalies. It is
now doable in just a couple of minutes.
This is an example of something that is
actually it's no code. It's not that
hard, but I hear actual orgs using
Copilot with Excel a lot less than I
hear them using it with DocX and with
Outlook. I don't know why, but Excel
seems to be under reppresented. Don't
sleep on Copilot in Excel if you have
C-pilot enabled. Meeting productivity.
Let's get into the meeting side of
things. Everybody has meetings. One of
the things that I learned actually is
that all of our meeting burden increased
after 2020. We had I think something
maybe 2 3x depending on how you measure
the baseline number of meetings and that
hasn't really gone away. So all that to
say meeting productivity matters more
than ever. You could try this prompt,
right? Let's say you're in premeating
prep. You type into co-pilot, I have a
client meeting about project X in an
hour. Give me an agenda. I want status
update. I want challenges. I want next
steps. It's going to spit out an agenda
with time limits and give you a sense of
those time limits. And then you're going
to say, hey, what questions should I ask
to understand the concerns that that the
client is going to uh express a little
bit better? And it's going to spit out
questions that you can actually dig
into. You can get co-pilot with teams in
a live meeting. And so you can say, what
have we discussed so far if you joined
late? and co-pilot will give you an
update without you interrupting Betty
and saying, "Hey, Betty, what are you
know what are we talking about?" You've
all been there, right? Like somebody
joins the Zoom or somebody joins the
team's meeting, you know, 6 minutes, 8
minutes late and like they're a highly
paid important person and so they get
the update and everybody has to sit
there and let you know the update
happen. You know, you can just ask
Copilot for it. Now, I'm not saying that
we humans will still do that, right?
Like again, there's a gap that we're
going to cover between what the tool can
let you do and how humans behave. Using
Copilot to the max demands cultural
change of your business, and we're going
to get into that later in this video.
Okay, so C-pilot searches your
transcript. Copilot produces summaries.
Post meeting, you can say, "Hey, create
meeting minutes. Give me attendees, key
discussions, decisions made, action
items with owners, and you'll get
formatted minutes super fast that you
can send to everybody." Now, I will say
like meeting minutes by themselves are
not particularly a thing. You can get
them from Granola. You can get them from
Otter. You can get them from I think
OpenAI does them. Now, the thing that
Microsoft has always sold is
distribution and bundling. Right now,
it's all in one place. Your your
co-pilot here will take care of it.
Okay. Day five, research assistant. You
want to look at
how you can use co-pilot beyond just
operating on documentation you already
have. So web research is a great example
of discovering new information. And so
you can give co-pilot a prompt and you
can say hey can you find say current
pricing for top five CRM software
companies. I'd like to create a
comparison table. Copilot searches it
returns this is what Salesforce costs.
This is what HubSpot costs. Here are two
or three others etc. Creates a nice
formatted comparison table. By the way
if you are in sales this is an example
of how people are actually collapsing
the buying funnel. people are making
high consideration software purchases by
talking to Copilot, by talking to chat
GPT.
Okay, so let's say you're not looking on
the web. Let's say you're looking
internally. You remember if you're on a
certain flavors of Copilot, the 365 one
for example, you can search all of your
documents for mentions of Q4 budget and
summarized findings because Microsoft
365 has has access to it and it will dig
it up and come back and say, "Hey, I
found three documents. There's budget
draft XLS, there's budget draft v2 xls."
You know, we all have our weird names
for our docs. I'll spare you the rest of
it. And it will tell you a little bit
about what's inside. It mentions 2
million allocated. It mentions a 15%
increase was approved. I think the thing
I will call out here is that this is
actually a feature that works because of
how powerful Microsoft has been for
decades. There are other companies like
Perplexity that are working on internal
document search for your PC, for your
Mac that have have got a hard road to
hoe. They have to build a lot more. and
you have to trust and remember
perplexity to use it and you're still
only an individual. And the advantage of
Microsoft really shines through here
because they have your data, you built
the data in their tool, they know where
the data is and they can just slap
Copilot over the top. This is why a lot
of people are using Copilot and hence
this video. Let's go from the basic
workflows with basic tools to advanced
individual workflows. We are past, hey,
can you draft this response to Sheila
from sales? We're going to go and do
something more fun. Let's do a batch
processing prompt in email. You can do
that. You know, can you please group
this list of 50 emails by topic? Get
project X 12. You can get budget
questions five. Meeting requests 8. It
really will group them, which makes it
easier to batch process your email. That
was always very hard to do before, but
my brain works by topic, so it's very
helpful. You can create prompts for
common emails. You can then like invoke
those prompts really quickly and you can
quickly answer and knock out those
batches. You can even have end of day
prompts. Hey, which emails from today
need follow-up tomorrow? And it will
give you that summary for the day. Okay,
let's get out of email. Everyone does
too much email. Let's go to the blog
post workflow. We talked before about
drafting it. Let's go farther. Let's do
research. You can find five recent stats
about remote work productivity. You can
outline and create a blog structure with
an intro of three main points in
conclusion. You can draft it, you can
polish it, you can say, "Hey, suggest
five SEO friendly titles for the post
and social. Create three Twitter posts
promoting this article." None of this is
really surprising. It doesn't take code.
But what it does is it challenges
someone who's only used to, hey, draft
this response as an AI capability or,
hey, write this blog post as an AI
capability to think of AI as
intelligence that underlies your entire
workflow. What could you do with your
time if intelligence was pushing on all
of the different pieces of your
workflow? And so the email example I
shared, what makes it advanced is that
you're thinking about your inbox
differently. You're thinking about it in
terms of topics and you're asking the
intelligence you got with Copilot to
attack it differently as a result. Same
with the blog post. You're asking the
intelligence that Copilot brings to help
you through all of the stages of
drafting and sharing, not just the
drafting and writing. you'll see that
theme pull through, that sort of end to
end theme pull through with most of
these advanced workflows. So, report
writing summarize all of our project X
documents from this month. That alone is
worth hours by the way. If you have
clean data, that is incredibly valuable.
Create a monthly report template with
standard sections. Creating a template
is something that used to take me a long
time because I had to think about
whether it was best practice. I had to
look up whether it was best practice. I
had to sort of agonize over every
section. Not anymore. create a monthly
report template and you know by
definition because of how co-pilot is
trained through reinforcement learning
on business data it is going to be best
practice. Now you can add to it but you
know you're getting a pretty decently
good template which is fantastic. Okay.
Now fill in the section with the
summarized information. Give me a
onepage executive summary and suggest
three charts. You can actually if you
have again if you have good data if you
have an opinion on what you want to call
out you can write an exact report like
that very very quickly. Previously that
would have taken a day or two. I will
emphasize again it does not mean that
you are letting the co-pilot do the
strategic thinking. I I think in a sense
Microsoft has branded it well as a
co-pilot because it means you are still
the pilot. You are still the driver.
Okay. Another advanced workflow. Let's
go to the data analysis problem. You can
actually let's go end to end with sales
analysis. We're not just going to say,
"Hey, tell me what's interesting." We're
going to say, "Hey, raw data, 10,000
rows of sales transactions. What's the
overall trend in the data?" And it will
say, "Hey, sales is growing 3.2% month
over month." Okay, great. Which products
are driving growth? You know, product A
is up 45%, product B selling to
developers is down 12%. I'm kidding. I I
love you good developers. You're great.
Are there any concerning patterns? Well,
returns are increasing for product B
people, especially, you know, in our in
our southern region. We're we're having
license expiration issues. We're having
churn issues, whatever it is, right? I'm
giving you an example that helps you
understand the kind of prompting you can
do with 10,000 rows, right? You can do
this with software. You can do this with
physical goods. Okay, great. Create a
pivot table. I want to see this by
region broken out. And then think with
me about actions that we can take. And
this is where I want to call out that
this is an advanced workflow partly
because of the you have to exercise.
Because at the end of the day, if you're
asking for the actions that we should
consider, co-pilot's going to give you
best practice actions, it will not
necessarily give you creative actions.
You need to think about the creativity
that you want to bring to this. And you
need to think about what you want to
accomplish strategically and make sure
that any recommendations co-pilot
generates are in line with that overall
strategy. Again, you bring the brains to
the exercise. Okay, we've talked a ton
in this video about individual
productivity. But one of the things the
CTO wrote me about was that he wanted to
see team productivity change. It's a
vexed issue, especially if you're in
seuite. How do you upshift teams? How do
you move teams from buckets of
individuals who do work with copilot and
say they save three or four hours a week
on surveys and you don't know where it
went to actually getting teams to move
forward. Let me give you a few examples.
Number one, super simple. Build a shared
prompt library. A lot of people are
doing this. They'll have email templates
like marketing team examples or like
draft a customer win announcement email.
Write the webinar invitation. These are
obviously not full prompts. I just gave
you three or four word prompts. If you
know me, you know I am known as the
prompt guy and I do long multi-page
prompts. So let's assume shorthand that
these are longer prompts that you are
now sharing across the team. That is a
good example within a team of
collaborating effectively. Generate five
LinkedIn post ideas for a white paper.
That's another one. Create an email
nurture sequence for new leads. That's
another one. Analyze campaign
performance. Suggest improvements.
There's another one. This inside team
baseball, this inside team improvement
doesn't just work for marketers. You can
do it for sales. Hey, draft a follow-up
email for the demo call and sales can
customize that to particular product
lines, right? Create a proposal
executive summary from our meeting notes
to send to a key stakeholder. Again,
that can be something you share. Write a
contract amendment for a scope change.
legal can weigh in there on the prompt
and give you the right framing so that
you get good draft language they can
later approve. This is not that hard.
You can create a team's channel called
co-pilot prompts. You just pin
successful prompts. You can tag them by
use case like # email #content and just
do a fun little monthly brown bag,
right? Like that's a very simple way to
start to socialize prompts across the
business. Let's say you don't want to
just look at productivity within teams
but between teams, right? Sales to
implementation. Let me give you an
example here. Sale can sales can create
a deal summary using copilot and it can
say extract all technical requirements
from this contract so our engineering
team can understand it. Copilot will go
through it will extract the technical
requirements. I would recommend reading
it again just to make sure but it's
going to give you a technical summary
and engineering will then get a clean
requirements list. Engineering can then
use copilot to say hey can you create an
implementation timeline based on these
requirements? engineering in most cases
will look at that implementation
timeline, swear at it, go to lunch, come
back, have a beer, and then write the
actual one down. But I find even if
you're upset with a co-pilot estimate,
it still moves the ball forward because
it gets you off the blank page. I find
in most cases when we're having like
agile estimation meetings with
developers, a lot of what we talk about
is that inherently estimating software
production is hard and starting with an
estimate somewhere helps us to move the
ball and shape our understanding. it it
helps the conversation to move in the
direction we want it to move because
we're no longer focused on getting onto
the page. We have an estimate now we can
argue about it. And that's one of the
things that Copilot helps you with. In
this case, the key value co-pilot brings
is it helps to translate hard contract
language into technical requirements
engineers can get. That by itself can
save you days of meetings. I have seen
complicated enterprise contracts founder
because the engineering team was not
brought in early enough. Most of us who
have been around software long enough
have seen stuff like that. Co-pilot can
help. That's an example of something
that literally doesn't just save time.
It can save an entire deal from
churning. Let me give you another sort
of cross team example. Marketing to
product. Marketing can say analyze the
customer feedback we're getting from
social media this month. Now, you're
going to have to get social media data
into an Excel spreadsheet if you want a
clean data set, or you're going to have
to trust that your signal is loud enough
on social channels that you give
co-pilot that you will get responsible
feedback. I will say if you just trust
it to go out and look at your social
accounts, it tends to be recency biased
and it tends to do things like sort of
cut off its investigation after a few
tweets or after a few Reddit threads.
And so, this is a case where marketing
needs to be responsible, understand how
the tool actually works. it's a little
bit token lazy and pull in all the data
that you need ahead of time. But
regardless, let's say you've done the
good stuff. You've got the data in place
in Excel. Analyze the customer feedback
from social media. You'll get themes.
Hey, there's UI confusion, X number of
mentions, feature X requests, uh, and
feature Y frustration. Great product can
then take that analysis that you get
from marketing. Maybe you throw a chart
in there, right? Copilot can help with
that. And it can then say, "Hey, can you
help me prioritize these feedback
themes? How do they align to our
existing roadmap?" And it can invoke
like roadmap.docx and it can put them
together and they can say, "Hey, this
could slot in here. This could slot in
here." That creative streamlining is
something that PMs have historically
been very frustrated by because it means
revisiting a road map and changing it
all the time. This makes some of that
pain easier. It means that marketing has
more of a voice with product. It is one
of the advanced workflows that
ultimately pays for itself not just in
sort of the time and alignment and lack
of frustration but in the right feature
getting built for customers more often.
Is it perfect? No. RPM still going to
argue about the road map 100%. But it
gives you a sense of what Copilot can do
beyond just help me with my email. Okay.
I want to go into an example of the
Vodafone rollout which I promised at the
beginning of this video. It's an
enterprise rollout. I believe they have
something like 68 70,000 employees
across a couple of dozen countries.
Obviously across any teams that size,
you have different technical skill
levels. You have multiple languages.
You've got various job functions. So
what did they do? They started with a
small pilot. This is all publicly
documented by the way. They started with
a small pilot. Only a few hundred users
in the UK mixed group. They actually
deliberately mixed, right? Sales,
support, IT management. And they had a
sixe pilot and they had regular
check-ins. They measured everything. the
time, the quality, the satisfaction.
They wanted to understand if this was
actually helpful. It was the pilot with
a small number of users that sold
leadership on the value of co-pilot at
scale. What they saw was it saved
something like 3 hours a week per
person. They saw that most people wanted
to keep using it. Most people said their
work quality improved. Customer service
in particular said response time was cut
dramatically. I think it was something
like 40%. Uh sales said proposals went
out days faster. it saw ticket
resolution times improve. When you start
to see in controlled environments that
are small like that, real gains and
people wanting to keep using it, it's an
indicator that this is a generally
available technology that can uplevel a
large range of enterprise job functions
if you let it. And what I love about the
Vodafone example is they took the time
to make sure in that controlled
environment that everyone had access to
the tool, the training, that everyone
understood the expectations, and that
everyone knew that this was special and
they would be focusing on this to make
sure they adopted it successfully. I bet
you this would not have been as
successful if it was just rolled out
casually across everybody. if they
didn't have training, if they didn't
measure everything, if they didn't do
weekly check-ins. The way you roll out
the intelligence helps you see if the
intelligence works or not and helps
people know how to use it. Otherwise,
you're just going to have Barbara
summarizing emails with Copilot and
writing, "Well, I save about an hour a
week, right?" And then CTO's email me
and say, "What are we doing wrong with
Copilot?" And the answer is probably a
culture change piece. There's a rollout
piece. So, phase one, foundation months
one to two, executive announcement from
the CEO. like when you are rolling out
intelligence. I'm assuming here that
this is a from scratch roll out. It's
not often actually the case because 90%
of you know the Fortune 500 or whatever
have co-pilot somewhere. But let's just
pretend that for clarity in the YouTube
video this is a clean roll out. So
initially couple first couple months CEO
announces this is coming. They've
identified a couple of dozen co-pilot
champions across different divisions.
I'm assuming this is a big company.
Again this is our focus is on enterprise
right now. They've built an internal
support site where you can get your
questions answered. They've developed
role specific training for different
departments and those champions are
owning that training. You can see how
you're starting to seed the change in,
right? And the CEO is setting the pace
and saying this is good, this is
positive, allaying fears about co-pilots
stealing your job, which anecdotally
talking to individual contributors, you
have to talk about the job thing if
you're going to talk about the AI thing.
People have been so primed by the news
to be afraid of AI stealing your job
that they won't listen to anything else
until you address it. And that must come
from the leadership team and preferably
from the CEO. And by the way, there's
not a ton of evidence that AI is
actually stealing jobs in aggregate
right now. The it's not just me saying
that. Ethan Mollik is one of the best
known AI uh academics out there. He
studied AI extensively. He tweeted as
recently as like early July 2025 that
there was just not much evidence that AI
is actually responsible for significant
job attrition isolated instances or
expectations changing for job roles 100%
changing job descriptions as scaling up
hiring for AI specific roles definitely
seeing that but we're not seeing
largecale attrition with AI and I think
part of why is the very reason I've had
to make this video guys If AI was as
good as it needs to be to take our jobs,
I wouldn't have to make this video
because you could ask Copilot and
Copilot would tell you how to roll
itself out. But it doesn't. And so we
need to do this culture change. Okay,
I've gone through my little soapbox
thing. Months one to two, we've set the
foundation. We've allayed the fears
around jobs. Phase two, now we're
getting into scaled out deployment.
Let's say you have a six-figure employee
count company. You're trying to get
10,000 users a week on, right? That's
the pace you can do with the like teams
channels for Q&A with a success story
sharing with the daily co-pilot tip
emails. You basically have some
operational support for these things
that kind of caps out at 10,000 users
and you want to see how it's working. By
the time you get to full adoption, you
know what? Maybe you have 100,000, maybe
you have 200,000, maybe like Vodafone,
you have 68,000 users enabled. You'll
have department competitions for best
use cases. You'll have integrations with
existing workflows. you're going to have
continuous measurement and optimization
to see if things are working. And you
know what? If you take your time and
roll it out like that, if you make it a
serious organizational change thing, it
gives you space to celebrate and talk
about your success factors. So you can
talk about the fact that the CEO and the
CTO personally championed this, that
that skin in the game mattered. You can
talk about how the focus of your roll
out was on helping each employee to do
their job better, not to just say use
AI. You can talk about and celebrate the
champions in your departments. You can
have weekly dashboards. You can even
have bonuses tied to impact. You can
have updates, success stories, tips
shared in dedicated teams channels
shared in brown bags, shared at all
hands. Listen, the the larger lessons
learned are not too hard here. You want
to have a phased roll out. If you're an
enterprise, you don't want to just do a
big bang. You want to have champions
that are individual contributors on the
front lines. They translate the tech to
business value. And your job as a leader
is to celebrate them. You need to
address fears upfront. You need to be
clear that this helps you and it doesn't
replace you. You need to measure what
works and what doesn't work and be
really honest about the places where it
doesn't work so you can fix it and
celebrate the wins. It really matters.
By the way, none of this is new. If you
have been in the culture change business
and leadership, this is what we've been
saying for a long time. And it's not
different with AI. In fact, it's more
important to do this with AI because AI
feels more personal. We as a human
species have never had to do shared
intelligence at work. We are at the
Henry Ford generation for figuring out
how to work with machine intelligence at
work. We need good leadership to make
that successful. And like it or not,
even if you love Chad GPT, and I talk
about Chad GPT a lot, co-pilot is where
a lot of that is going to be worked out
in the workplace because of Microsoft
distribution advantage. Okay, let's just
call out a few of the advanced features
people miss as we sort of transition
toward toward the end of this lengthy
video. Thank you for staying with me. I
told you it'd be worth it. There's a lot
of specifics. So, Copilot Vision is a
really interesting one. People don't use
it way beyond email. Now, you can take a
screenshot of an error message and you
can say, "Hey, can you explain this
error and how to fix it?" Copilot will
read your screen. It'll explain it in
plain English. You can take a screenshot
of a PDF table. Hey, can you convert
this to Excel? Unless the table is
really small with tiny print, that will
often work. Uh, and you'll get properly
formatted data and you can have it ready
to paste. You can take a screenshot of a
complicated piece of software or
onboarding and say, "Hey, I don't know
how to onboard. I don't know what option
to pick. How do I export data?" I will
tell you, I have absolutely done this
with complicated software that is poorly
documented. And I get so much help from
AI doing that. Like I can get buttons,
you can get menu paths out of Copilot.
It's really, really helpful. Okay.
Another one is prompt chaining. I've
talked about prompt training elsewhere.
This is how you do it with Copilot. As
an example, think about five prompts
that create a business plan. Number one,
outline the business plan for an AI
consulting firm. Number two, expand a
particular section that has goes deeper
on market analysis. Number three, add
financial projections for the next three
years. Number four, create an executive
summary based on the sections. And
number five, suggest five potential
risks. individually. These are prompts
that you might think of for drafting,
but when you take them together, you're
basically going on a chain of thought
with the AI. And I said it really fast,
but you're going to write it slower. And
the reason why is that you need to bring
your own thinking to each piece in
prompt chaining so that you're actually
partnering with AI. And it almost looks
like a walk through the conceptual space
you're exploring together with machine
intelligence. Each step in the prompt
chain helps you move your own thinking
forward. This is a sneak peek into how
people are doing advanced drafting with
AI. Let's look at another example.
Report generation. You can do multiple
different tools here in Excel. Analyze
the Q4 sales data. Copy those insights
to Word. Then talk to Word and say,
"Hey, Copilot, create a professional
report using these insights. Add an
introduction. Then you move over to
PowerPoint Copilot. Hey, create a five
slide presentation from this Word
report." Then go to Outlook. Hey, draft
an email to leadership with the report
attached. Bam. Bam. Bam. Bam. So fast.
Again, you have to bring your brain. You
have to know what you want to say. But
it is possible to jump from dock to doc.
And this is another example where
there's no code needed. And co-pilot can
help. And most people aren't doing that
jump. Here's another one. You can build
your own bots. HR onboarding bot. An HR
onboarding bot built by HR without code
can save HR several hours a week by just
answering what's the vacation policy. I
need to add my spouse to insurance. Walk
me through setting up my phone. If
you're at scale in the enterprise, you
get lots of those queries every single
day. Having a bot is really helpful. IT
help desk bot, similar thing. You're
going to get a lot of queries at the
enterprise scale. Build it yourself. It
can run password reset workflows. It can
run software installation guides. It can
run escalation to a human when needed.
It might deflect half of your tickets,
60% of your tickets. It can be very
helpful. All right, I don't want to
leave you without a few more success
stories with Copilot. DWF law firm, this
is mentioned uh Microsoft talked about
them actually as a case study. They use
C-pilot to reduce the time it takes to
review contracts. I haven't talked a ton
about how legal is seeing tremendous
changes with AI. But even though legal
has a very high bar on accuracy, they
cannot mess anything up. You can still
use something like Copilot to get an
early read and reduce the time. I think
the time taken was like 7 days to 7
hours and they got more consistent
quality because they uploaded the
contract and critically they asked the
right prompts. Identify obligations.
Identify deadlines. Flag unusual terms
compared to our standard, which requires
Copilot to know your standard. Create a
summary in plain English. It's all about
the kinds of prompting that you bring.
And you see how they're prompting in
ways that give the AI room to succeed
here. Healthcare is another example. You
can use Copilot for patient discharge
summaries. Doctors might spend 30
minutes writing discharges. Now, Copilot
can just pull from patient records,
create the discharge summary, including
medications, and follow up and be done
with it. It can result in a dramatic
drop in the time spent doctors on
doctors spend on paperwork and more time
with patients, finance, and I know this
sounds like an AI advertisement, right?
Like that my point is not to advertise
copilot partly because I don't need to.
Everybody has it. The point is to
articulate that people can do a lot more
and people don't realize it. Finance is
another one. You can feed market data
into Excel and you can actually get a
second perspective on market anomalies.
Retail, you can upload a product image
and you can say write this product
description. That one gets done a lot.
Doesn't just get done with Copilot.
Okay, if you want to do this, it is not
that hard. You can get co-pilot in. You
can try basic prompts. You can get into
your email. You can get into Word and
Excel in the first couple days. And if
you want to get into Copilot Vision, if
you want to get into chain prompting,
some of the advanced stuff I talked
about, just try it. This is one of those
things that I keep emphasizing with AI.
Just try it. You will learn if you are
willing to try. If you're with teams and
you're trying to sort of not teams the
software, teams the people like people
teams, let your team try it. Celebrate
the wins your team is getting with
co-pilot. Encourage your team by showing
them that you value their time saved.
And you will not just pile more work on.
You will not just tell them you must not
be that good at your job because you did
it that fast. you'll actually celebrate
the win. Okay, here's the reality. I'm
going to go right back to the beginning.
The CTO wrote me and was like, they're
only using email, 6,000 person company.
These companies are are using 10% of a
six figure investment in some cases,
whereas the ones that are using it
fully, they're getting extra more return
on investment. The swing that I see here
is insane. I talked earlier about the
security edition that costs like $35,000
a month. I talked about the ROI you get
if you can reduce the time of the
incident because you're actually better
at triaging. That by itself delivers 20x
ROI. Another way to get to that 20x ROI,
get these cross team workflows I talked
about going. Get marketing talking to
product with co-pilot. Get sales working
with engineering more successfully.
You're going to hit that ROI. advanced
co-pilot usage like I am describing is
going to be table stakes relatively
quickly. If you don't adopt it, you're
going to be behind. And that's another
reason I'm choosing to make this video.
I could have sat on this and said, "I
need more information. I need more
information. It's not perfect because no
information set ever is." I'm making
this video now because you need to adopt
Copilot quickly if you have it in your
business and you need to adopt it
successfully and waiting won't help you.
Start with a workflow I outline here in
the video. Measure the time you saved
this week. Share the video with your
team. Co-pilot mastery is a competitive
advantage in the workplace by definition
because co-pilot matters so much because
copilot is in 90% of enterprises.
Everything I'm describing is going to
look like a basic requirement in about a
year. So start now. Okay. I have a full
written guide up. You know where to find
it. If you want to see me make more
videos on co-pilot, let me know. This
has been a long time coming. There was a
ton to cover. I'm happy to dive deeper
on specific industries. I'm happy to
dive deeper on specific flavors of
Copilot. Wanted to summarize it and give
you the overall introduction first. I
hope you've enjoyed it. I know that uh
I've been excited to actually have
meaningful AI conversations with people
who use Copilot so we can move from this
world where people are just using it for
email to a world where people actually
can take advantage of the fact that we
have taught the rocks to think and now
they're at work with us in Microsoft
Word and they can help us do smarter
things faster. You still bring the
brains to work, but co-pilot will help.