00:00:00
I think the notion that business
00:00:04
applications exist that's probably where
00:00:07
they'll all collapse right in the agent
00:00:09
era Microsoft's CEO Satia nadela said
00:00:12
apps as we know them are going away in
00:00:15
favor of agents and this is a huge
00:00:18
statement he basically put a flag in the
00:00:20
ground showing where the future of
00:00:22
software development is going no more
00:00:24
applications which means no more SAS
00:00:27
which also means likely far fewer
00:00:30
developers so let's watch the video and
00:00:32
I'm going to break it down for you so
00:00:34
this is on the BG squared podcast I'll
00:00:36
drop a link to the full interview down
00:00:37
below SAS applications or Biz apps so
00:00:40
let me just speak of our own Dynamics
00:00:42
thing the Approach at least we're taking
00:00:44
is I think the notion that business
00:00:50
applications
00:00:51
exist that's probably where they'll all
00:00:54
collapse right in the agent era because
00:00:56
if you think about it right they are
00:00:58
essentially
00:01:00
crowd databases with a bunch of business
00:01:03
logic okay I'm going to pause there
00:01:05
because what he just said might sound
00:01:08
like no big deal but it is actually a
00:01:10
stunning proclamation of the future what
00:01:13
he essentially said is that SAS and that
00:01:15
means software as a service is
00:01:17
essentially just a thin UI layer on top
00:01:21
of a database and when he says crud that
00:01:24
means create read update destroy those
00:01:27
are the base units of interactions
00:01:29
between between an interface and a
00:01:31
database so create create a new record
00:01:34
read extract information from the
00:01:36
database update make a change to an
00:01:38
existing record and Destroy delete so
00:01:41
all he's saying is there's going to be
00:01:43
the underlying database the fundamental
00:01:46
data of what you need and then it's
00:01:48
going to be agents that are interacting
00:01:50
directly with the database and if you've
00:01:52
been watching this Channel at all you
00:01:53
know I've been talking about this for a
00:01:55
long time I truly believe the entire
00:01:58
application stack is going to disappear
00:02:01
there just really is no need for it when
00:02:03
you have artificial intelligence
00:02:05
interacting with the core grounded data
00:02:08
that sits within a database and again
00:02:11
what does that actually mean for the
00:02:13
entire SAS industry what does that mean
00:02:15
for application developers I don't know
00:02:18
exactly but I have a feeling it's going
00:02:20
to look vastly different than it looks
00:02:22
today and I've been in SAS my entire
00:02:25
career I have built multiple SAS
00:02:28
companies I've worked for SAS companies
00:02:30
and if I'm being transparent I probably
00:02:32
wouldn't start another SAS company right
00:02:34
now and I probably wouldn't invest in
00:02:37
any SAS companies right now either it's
00:02:39
just too unknown now that isn't to say
00:02:41
there won't be software companies in the
00:02:43
future it's just going to look
00:02:45
drastically different let's keep
00:02:47
watching crowd databases with a bunch of
00:02:51
business
00:02:52
Logic the business logic is all going to
00:02:56
these agents and these agents are going
00:02:59
to be multi- repo crud right so they're
00:03:02
not going to discriminate between uh
00:03:06
what the back end is they're going to
00:03:07
update multiple databases and all the
00:03:10
logic will be in uh the AI tier so to
00:03:15
speak I'm so excited to show you this
00:03:17
video this 3 and a half minute video
00:03:19
might be the clearest depiction of what
00:03:22
the future of software engineering is
00:03:23
going to look like now when he's talking
00:03:25
about business logic he basically means
00:03:27
taking that core data found in a
00:03:29
database and actually doing things with
00:03:31
it whether it's updating a CRM record or
00:03:33
sending out emails based on different
00:03:35
criteria found in the database or
00:03:37
anything any single piece of software
00:03:39
that sits on top of a database that is
00:03:41
what he's talking about all of that
00:03:43
business logic is going to be compressed
00:03:46
into the agent the agent will be able to
00:03:48
accomplish anything and all you need to
00:03:51
do is tell it what you want give me a
00:03:54
graph of my top five customers by
00:03:57
Revenue okay now write emails to all
00:04:00
five of those customers telling them I
00:04:02
want to do business with them in 2025
00:04:04
all of this will just automatically be
00:04:07
done where previously it would take
00:04:10
hardcoded business Logic on top of the
00:04:12
database now that's not necessary you
00:04:15
just tell the agent what you want and it
00:04:17
will either know how to accomplish it or
00:04:20
write tools to go accomplish it on your
00:04:23
behalf and when I say tools I just mean
00:04:25
it's going to write its own code to
00:04:27
interact with the database to send it an
00:04:29
email and so on I want to thank the
00:04:31
sponsor of this video chatbase chatbase
00:04:34
is the ultimate platform for creating
00:04:36
and managing intelligent AI agents for
00:04:38
your business they are designed to
00:04:40
streamline customer support generate
00:04:42
leads and enhance business operations
00:04:44
with Integrations across websites social
00:04:47
platforms and other digital tools
00:04:49
chatbase helps businesses provide fast
00:04:51
consistent responsive and personalized
00:04:53
interactions at scale and it is powered
00:04:56
by Advanced AI models from open AI
00:04:59
anthropic and others you can train these
00:05:01
agents on your own unique data so
00:05:03
they're customized for your business and
00:05:05
know to give the right answers at the
00:05:07
right time and they just launched a
00:05:08
brand new feature AI actions and if you
00:05:11
watch this Channel at all you know I'm
00:05:13
all about agents taking actions as the
00:05:16
future of agentic workflows so now
00:05:18
chatbase allows these agents to not only
00:05:20
respond to your customers on your behalf
00:05:23
but actually take realtime actions based
00:05:25
on what they're looking for some of the
00:05:27
actions include getting data real-time
00:05:29
data from their account or even allowing
00:05:31
them to update things in their account
00:05:34
based on the conversation that they're
00:05:35
having with that agent so check out
00:05:37
chatbase today thank you again to
00:05:39
chatbase for sponsoring this video and
00:05:40
now back to the video and once the AI
00:05:43
tier becomes the place where all the
00:05:46
logic is then people start replacing the
00:05:48
back ends right we people you know
00:05:50
that's what you know in fact it's
00:05:51
interesting as we speak I think we are
00:05:54
seeing pretty high rates of wins on
00:05:57
Dynamics backends um and the agent use
00:06:01
and we are going to go pretty
00:06:03
aggressively and try and collapse it all
00:06:05
right whether it's in customer service
00:06:07
whether it is in you know uh by the way
00:06:10
the other fascinating thing that's
00:06:12
increasing is just not CRM but even our
00:06:14
what we call finance and operations uh
00:06:17
because people want more AI native Biz
00:06:20
apps right another thing that he clearly
00:06:22
says is the backend doesn't matter the
00:06:25
agents aren't specialized in one type of
00:06:27
database they really don't care so
00:06:30
really it's about building the database
00:06:33
that works most efficiently with the
00:06:34
agent is cheapest is the right for The
00:06:37
Right Use case so as an engineer or a
00:06:39
data provider you really just need to
00:06:40
figure out what's the best database to
00:06:42
use and then the agents do not care and
00:06:45
so it's going to be interesting on how
00:06:46
companies choose the database that they
00:06:48
decide to use that means the biz the
00:06:51
logic tier can be orchestrated by Ai and
00:06:56
AI agents so in other words copilot to
00:06:58
agent to my business application should
00:07:02
be very
00:07:03
seamless now in the same way you could
00:07:06
even say hey why do I need Excel like
00:07:09
interestingly enough one of the most
00:07:10
exciting things for me is Excel with
00:07:13
python is like GitHub with co-pilot
00:07:15
right that's essentially so what we've
00:07:17
done is when you have Excel like this by
00:07:21
the way will be fun for you guys right
00:07:22
which is you should just bring up Excel
00:07:25
bring up co-pilot and start playing with
00:07:27
it because it's no longer like oh you
00:07:30
know it is like having a data
00:07:33
analyst uh and so it's no longer just
00:07:36
making sense of the numbers that you
00:07:38
have it will do the plan for you right
00:07:41
it will literally like how get up
00:07:43
copilot workspace creates the plan and
00:07:45
then it executes the plan it this is
00:07:48
like a data analyst who is using Excel
00:07:51
as a sort of R column visualization to
00:07:55
do analysis scratchback so he is saying
00:07:59
why do you even need Excel and he
00:08:01
probably has thought a lot about this
00:08:03
because really what is Excel it is a
00:08:05
thin UI layer on top of a database and
00:08:08
of course you can add a bunch of
00:08:09
business logic and analytics but here's
00:08:11
the thing what he's describing is
00:08:13
essentially if AI can write python code
00:08:16
to analyze the data why do you need
00:08:18
Excel at all and he might be right he
00:08:21
probably is right why do you need Excel
00:08:24
if I need an analysis of a huge amount
00:08:27
of data I just say what I need I just
00:08:29
need to describe the end product I need
00:08:32
a graph I need a chart I need a 3D
00:08:35
visualization whatever it is you just
00:08:37
describe the end result you have the
00:08:40
core data the raw data and then the
00:08:44
agent writes the python code to actually
00:08:47
extract the right data and put it in the
00:08:49
right format for what you asked now
00:08:51
let's keep watching so it's got a tools
00:08:53
you so the co-pilot is using Excel as a
00:08:56
tool with all of its action space
00:08:58
because it can generate and it has
00:09:00
python interpreter that is in fact and
00:09:03
by the way when he says python
00:09:04
interpreter that just means the ability
00:09:06
to execute python code fact a great way
00:09:09
to reconceptualize Excel and at some
00:09:11
point you could say hey I'll generate
00:09:13
all of excel uh and that is also true
00:09:15
after all there's a Code interpreter
00:09:16
right so therefore you can generate
00:09:18
anything um and so yes I think there
00:09:20
will be disruption but so the way we are
00:09:23
approaching at least our M365 stuff is
00:09:25
one is build co-pilot as that organizing
00:09:28
layer UI for AI get all agents including
00:09:32
our own agents you can say the Excel is
00:09:34
an agent to my co-pilot word is an agent
00:09:38
it's kind of specialized canvases which
00:09:40
is I'm doing a legal document let me
00:09:43
take it into pages and then to word and
00:09:45
then have the co-pilot Go With It uh go
00:09:50
into Excel and have the co-pilot go with
00:09:52
it this is a fundamental change in how
00:09:55
Microsoft applications work and really
00:09:58
any soft Ware as a service business this
00:10:00
is a complete restructuring of the
00:10:03
entire industry and for somebody who has
00:10:06
been in SAS for so long it is so
00:10:08
fascinating to watch and companies are
00:10:11
either going to adopt it and kind of
00:10:13
evolve with the times or they're going
00:10:15
to die there is no in between and it's
00:10:17
clear that SAA from Microsoft is
00:10:19
thinking well ahead he is already very
00:10:22
invested into open AI he already has the
00:10:24
best model he's already thinking about
00:10:26
co-pilot whatever you think about
00:10:28
co-pilot and it's quality today at least
00:10:31
the structure the architecture of how
00:10:33
it's going to work in the future is
00:10:36
clear and right in my opinion and so
00:10:39
that's the whole video and it's again
00:10:40
only 3 and a half minutes but clearly
00:10:43
lays out what the future of software
00:10:46
will look like all he's talking about is
00:10:48
enterprise software business software
00:10:50
but at the same time that can be applied
00:10:52
to Consumer as well what's the
00:10:54
difference where a consumer it's just a
00:10:56
database and you're getting a different
00:10:58
type of on top of it and so if you had
00:11:01
an agent why do you need the interface
00:11:02
at all the future of software
00:11:04
development the future of application
00:11:07
development the future of interaction
00:11:09
with data is going to be drastically
00:11:11
different than it is today and I'm so
00:11:13
excited for it if you enjoyed this video
00:11:15
please consider giving a like And
00:11:16
subscribe and I'll see you in the next
00:11:18
one
