3 reasons why every real-time application needs AI
By
Bryan
Kirschner,
Vice
President,
Strategy
at
DataStax
Imagine
getting
a
recommendation
for
the
perfect
“rainy
Sunday
playlist”
midway
through
your
third
Zoom
meeting
on
Monday.
By
Bryan
Kirschner,
Vice
President,
Strategy
at
DataStax
Imagine
getting
a
recommendation
for
the
perfect
“rainy
Sunday
playlist”
midway
through
your
third
Zoom
meeting
on
Monday.
Or
a
receiving
text
about
a
like-for-like
substitute
for
a
product
that
was
out
of
stock
at
your
preferred
e-commerce
site
10
minutes
after
you’d
already
paid
a
premium
for
it
on
another.
Or
arriving
late
for
lunch
with
a
long-time
friend
and
being
notified
that
“to
have
arrived
earlier,
you
should
have
avoided
the
freeway.”
We
all
expect
apps
to
be
both
“smart”
and
“fast.”
We
can
probably
all
call
to
mind
some
that
do
both
so
well
that
they
delight
us.
We
can
also
probably
agree
that
failures
like
those
above
are
a
recipe
for
brand
damage
and
customer
frustration—if
not
white-hot
rage.
We’re
at
a
critical
juncture
for
how
every
organization
calibrates
their
definition
of
“fast”
and
“smart”
when
it
comes
to
apps—which
brings
significant
implications
for
their
technology
architecture.
It’s
now
critical
to
ensure
that
all
of
an
enterprise’s
real-time
apps
will
be
artificial-intelligence
capable,
while
every
AI
app
is
capable
of
real-time
learning.
“Fast
enough”
isn’t
any
more
First:
Meeting
customer
expectations
for
what
“fast
enough”
means
has
already
become
table
stakes.
By
2018,
for
example,
the
BBC
knew
that
for
every
additional
second
a
web
page
took
to
load,
10%
of
users
would
leave—and
the
media
company
was
already
building
technical
strategy
and
implementation
accordingly.
Today,
Google
considers
load
time
such
an
important
positive
experience
that
it
factors
into
rankings
in
search
results—making
“the
speed
you
need”
a
moving
target
that’s
as
much
up
to
competitors
as
not.
The
bar
will
keep
rising,
and
your
organization
needs
to
embrace
that.
Dumb
apps
=
broken
apps
Second:
AI
has
gotten
real,
and
we’re
in
the
thick
of
competition
to
deploy
use
cases
that
create
leverage
or
drive
growth.
Today’s
winning
chatbots
satisfy
customers.
Today’s
winning
recommendation
systems
deliver
revenue
uplift.
The
steady
march
toward
every
app
doing
some
data-driven
work
on
behalf
of
the
customer
in
the
very
moment
that
it
matters
most—whether
that’s
a
spot-on
“next
best
action”
recommendation
or
a
delivery
time
guarantee—isn’t
going
to
stop.
Your
organization
needs
to
embrace
the
idea
that
a
“dumb
app”
is
synonymous
with
a
“broken
app.”
We
can
already
see
this
pattern
emerging:
In
a
2022
survey
of
more
than
500
US
organizations,
96%of
those
who
currently
have
AI
or
ML
in
wide
deployment
expect
all
or
most
of
their
applications
to
be
real-time
within
three
years.
Beyond
the
batch
job
The
third
point
is
less
obvious—but
no
less
important.
There’s
a
key
difference
between
applications
that
serve
“smarts”
in
real
time
and
those
capable
of
“getting
smarter”
in
real
time.
The
former
rely
on
batch
processing
to
train
machine
learning
models
and
generate
features
(measurable
properties
of
a
phenomenon).
These
apps
accept
some
temporal
gap
between
what’s
happening
in
the
moment
and
the
data
driving
an
app’s
AI.
If
you’re
predicting
the
future
position
of
tectonic
plates
or
glaciers,
a
gap
of
even
a
few
months
might
not
matter.
But
what
if
you
are
predicting
“time
to
curb?”
Uber
doesn’t
rely
solely
on
what
old
data
predicts
traffic
“ought
to
be”
when
you
order
a
ride:
it
processes
real-time
traffic
data
to
deliver
bang-on
promises
you
can
count
on.
Netflix
uses
session
data
to
customize
the
artwork
you
see
in
real
time.
When
the
bits
and
atoms
that
drive
your
business
are
moving
quickly,
going
beyond
the
batch
job
to
make
applications
smarter
becomes
critical.
And
this
is
why
yesterday’s
AI
and
ML
architectures
won’t
be
fit
for
purpose
tomorrow:
The
inevitable
trend
is
for
more
things
to
move
more
quickly.
Instacart
offers
an
example:
the
scope
and
scale
of
e-commerce
and
the
digital
interconnectedness
of
supply
chains
are
creating
a
world
in
which
predictions
about
item
availability
based
on
historical
data
can
be
unreliable.
Today,
Instacart
apps
can
get
smarter
about
real-time
availability
using
a
unique
data
asset:
the
previous
15
minutes
of
shopper
activity.
‘I
just
wish
this
AI
was
a
little
dumber,’
said
no
one
Your
organization
needs
to
embrace
the
opportunity
to
bring
true
real-time
AI
to
real-time
applications.
Amazon
founder
Jeff
Bezos
famously
said,
“I
very
frequently
get
the
question:
‘What’s
going
to
change
in
the
next
10
years?’
…
I
almost
never
get
the
question:
‘What’s
not
going
to
change
in
the
next
10
years?’
And
I
submit
to
you
that
that
second
question
is
actually
the
more
important
of
the
two—because
you
can
build
a
business
strategy
around
the
things
that
are
stable
in
time.”
This
sounds
like
a
simple
principle,
but
many
companies
fail
to
execute
on
it.
He
articulated
a
clear
north
star:
“It’s
impossible
to
imagine
a
future
10
years
from
now
where
a
customer
comes
up
and
says,
‘Jeff,
I
love
Amazon;
I
just
wish
the
prices
were
a
little
higher.’
‘I
love
Amazon;
I
just
wish
you’d
deliver
a
little
more
slowly.’
Impossible.”
What
we
know
today
is
that
it’s
impossible
to
imagine
a
future
a
decade
from
now
where
any
customer
says,
“I
just
wish
the
app
was
a
little
slower,”
“I
just
wish
the
AI
was
a
little
dumber,”
or
“I
just
wish
its
data
was
a
little
staler.”
The
tools
to
build
for
that
future
are
ready
and
waiting
for
those
with
the
conviction
to
act
on
this.
Learn
how
DataStax
enables
real-time
AI.
About
Bryan
Kirschner:
Bryan
is
Vice
President,
Strategy
at
DataStax.
For
more
than
20
years
he
has
helped
large
organizations
build
and
execute
strategy
when
they
are
seeking
new
ways
forward
and
a
future
materially
different
from
their
past.
He
specializes
in
removing
fear,
uncertainty,
and
doubt
from
strategic
decision-making
through
empirical
data
and
market
sensing.
About Author
