Businesses can effectively leverage generative AI by shifting focus from automating entire jobs to automating specific tasks within those jobs, using a framework that assesses technical feasibility and business value.
Mind Map
انقر للتوسيع
انقر لاستعراض خريطة الذهن التفاعلية الكاملة
many businesses think of a large or a
small company say have many people doing
many different TS there's a framework
that had originated in economics due to
Eric Brunson Tom Mitchell and Daniel
Rock for analyzing the work task for
possible automation using Ai and this
framework turns out to be useful not
just for Economist to understand the
financial or economic impact of AI but
also for businesses identify specific
specific opportunities to use generative
AI let's take a look at how to do this
whereas there's been a lot of discussion
for example in the media about will AI
automate jobs it turns out that from a
technical and business perspective it's
more useful to think of AI not as
automating jobs but as automating tasks
and it turns out most jobs involve a
collection of many tasks let's look at
an example a customer service
representative will do a number of
different toss including maybe answer
inbound phone calls from customers
answer customer chat queries via text
rather than a voice or phone interface
they may check status of customer orders
keep records and directions and assess
the accuracy of customer complaints and
if you work in a company with say many
customer service Representatives the
first step to analyzing the potential
for using gender AI would be to
understand for your business what are
the tasks that the representatives in
your company do after that we can then
take a look at these different tasks and
try to assess their potential for
generative AI to either hope or augment
or to automate these thoughts for
example for generative AI to pick up the
phone and have a long conversation
that's still pretty difficult so we
assess that to be a lower potential
opportunity but answering text chat with
customers that might have a higher
potential maybe checking status of
customer orders is medium whereas
keeping records of customer interactions
could be high and assessing accuracy of
customer complaints may be low all of
these examples in the rightmost column
are hypothetical and the actual impact
on your business will be different and
would depend on the specifics of your
business but after an analysis like this
and I'll go in a little bit into the
specifics of how to carry out this
analysis you might then decide that
answering customer chat queries and
keeping records of customer interactions
have The Highest Potential and therefore
Focus your efforts on those two tasks
now the opportunity for Gen of AI could
be either augmentation or automation by
augmentation I mean we can use AI to
help a human with a Tas in the customer
service representative context we might
have genbi recommended response for a
customer service agent to edit or
approve but not fully automate the
sending of a message back to the
customer so if we're not sure yet if the
Gent of AI would give good answers then
recommending a response could speed up
the people doing the work but not fully
automated and this would be an example
of augmentation and automation would be
if we have an AI system fully
automatically perform at all so if we
were to automatically transcribe and
summarize records of customer
interactions um that could be an example
of automation what I see in many
applications is that businesses will
sometimes start with augmentation to
maybe let a human double check or
finalize the output before this use but
then as you gain trust and gain
confidence in the output of the Gena of
AI then the user interface can be
adapted to make the process more and
more efficient for humans and to then
gradually shift toward higher and higher
degrees of augmentation and perhaps
eventually to full automation now given
a list of tasks like this how do you
come up with this column on the right
how do you evaluate the different tasks
for G AI potential the potential for
augmenting or automating a task depends
mostly on two things technical
feasibility and business value so
technical feasibility refers to can AI
do it and also how costly is it to build
an AI system to do it and with regard to
using an LM I found the framework we
discussed last week of asking can a
fresh college graduate following the
instructions in the prompt the task that
could give you a first guess an
imperfect not necessary fully accurate
guess but it gives you a way to think
about whether a certain task may be
doable or not and sometimes if you're
not sure if an LM can do a certain task
I would encourage you to try prompting
an LM to see if you can get the LM to do
that task and this would be an
experiment that you might be able to do
quite quickly so long as you're not
revealing confidential information if
you you take some prompts for say
answering customer chat queries and
paste them into a large language model
you can maybe quickly get a sense of how
good RS response is and this could help
you relatively quickly assess technical
feasibility of using generative AI for a
particular cost and an AI engineer can
also help you assess if more Advanced
Techniques like rack retrieve alment
generation fine-tuning or other
techniques can help and also give you a
sense of perhaps the complexity and
therefore the cost of building an AI
system to tackle a certain task in this
course I'm focusing mainly on technical
feasibility using gen of AI technology
if you or your team is familiar with
other AI tools such as supervised
learning you can also assess the
technical feasibility of using other
tools as well for augmenting or
automating different tasks other than
technical feasibility the second
criteria I urge you to think through is
the business value so how valuable is it
to use AI to either augment or automate
a particular task and so the questions I
would ask to frame up my thinking on
business value would be things like how
much time is spent on this task so how
much time savings can we actually
realize second I'd also ask does during
this task significantly faster cheaper
or more consistently using AI create
substantial value while it may seem like
augmentation and automation helps lead
to cost savings we'll see later this
week as well that when you automate a
TSS sometimes the benefits are much
greater than just cost savings because
it also lead to rethinking the workflow
around that toss but if what I'm saying
doesn't make sense yet don't worry about
it we'll see some specific examples of
this later this week before wrap up this
video there's one more resource I want
to share that maybe us useful for your
analysis of how to break job RS down
into task which is that there are online
occupation databases that you can look
up to see what are the tasks that
comprise a certain roow here's a
screenshot from a website called onet
which is a US government funded website
that for the customer service
representative row lists lots of
different tasks including confir of
customers by telephone or in person key
records of customer interactions and so
on a found that occupation databases
like this tend to be General and not
necessarily specific to your company and
so I wouldn't recommend just using the
results from say this onet database and
assuming it's accurate for your company
there'll usually be some entries there
that you read and feel like no this
doesn't seem like it applies to my
company but I found that this is useful
resource to take a look at just for
ideas and to help make sure that maybe
you haven't missed anything when
thinking through what are the tasks done
by people in different job roles in your
company onet is a little bit us Centric
but has a nice easy to use user
interface so that encourage you to play
with it and there are some other
countries as well that have some other
country or region specific databases
that you may able to find online as well
but I found that for many job RS onet is
may be a reasonable initial starting
point so that's how you can look at
different job roles and start to break
them down in the TSS and analyze the IND
indidual task for potential for
augmentation or Automation and I hope
you play with the onet website and get a
feel for what different tasks and
different job roles look like in this
video we went through the customer
service representative example I'd like
to go through you a few examples of
other job roles as well so let's go take
انقر على أي نص أو طابع زمني للانتقال إلى تلك اللحظة في الفيديو
مشاركة:
معظم النصوص تصبح جاهزة في أقل من 5 ثوانٍ
نسخ بنقرة واحدة125+ لغةالبحث في المحتوىالانتقال إلى الطوابع الزمنية
الصق رابط YouTube
أدخل رابط أي فيديو YouTube للحصول على نصه الكامل
نموذج استخراج النص
معظم النصوص تصبح جاهزة في أقل من 5 ثوانٍ
احصل على إضافة Chrome
احصل على النصوص فوراً دون مغادرة YouTube. ثبّت إضافة Chrome للوصول بنقرة واحدة إلى نص أي فيديو مباشرةً من صفحة المشاهدة.
