0:03 there are many ways that J of AI is
0:06 leading not just to cost savings but to
0:09 revenue growth but there's so many ways
0:12 for general purpose technology like
0:14 generative AI to create value to talk
0:17 about all of these ways but in this
0:18 video I'd like to take a look at a
0:20 couple of the emerging or the more
0:22 common paths I'm seeing toward this type
0:25 of growth let's consider the
0:28 hypothetical example of a surgeon
0:30 preparing to car out a surgical
0:32 operation and then actually carrying out
0:34 that operation without generative AI
0:37 tools they may need a long time to do
0:39 background research to understand the
0:42 medical procedure and how to carry that
0:44 out in the context of a specific patient
0:47 and then they will go into the operating
0:49 theater to carry out that surgery if gen
0:53 of AI maybe with custom rag or retrieval
0:55 augmented generation tools can help with
0:58 the research portion of this work then
1:00 perhaps the time and effort need needed
1:03 to deeply research the medical procedure
1:06 could shrink and then they still need to
1:09 go in to carry out the surgery after
1:12 that so this would be an example of how
1:15 gen of AI could hypothetically help a
1:17 surgeon or some of job row where you
1:19 need to do some research and go and
1:21 carry the task with this diagram
1:24 illustrating how the workflow or the
1:26 effort needed might change let's look at
1:29 the second example a lawyer reviewing a
1:32 comp Lex legal documents might without
1:34 generative AI first have to gather
1:36 information from the client so sit down
1:38 with the client to really understand
1:40 what's the purpose of this contract who
1:42 are the people signing it what are the
1:44 key business terms and then they might
1:47 review the documents and then finally
1:48 sit down with the client to give them
1:52 feedback with Gen of AI maybe gathering
1:54 information takes just about as much
1:57 effort as before but if gen of AI can
2:00 help read over the document perhaps this
2:03 step can become shorter but then maybe
2:04 the lawyer still needs to sit down with
2:06 the client to give them detailed
2:08 feedback again this is a hypothetical
2:10 example based on what I'm seeing
2:13 companies do but the details of your
2:16 workflow may be different in contrast
2:18 this is what changes the workflow might
2:20 look like for reviewing simple as
2:23 opposed to complex legal documents for
2:26 comparatively simple legal documents say
2:29 a typical non-disclosure agreement or NDA
2:30 NDA
2:33 there may not be much information to
2:34 gather and so the lawyer could just
2:37 review it and then give the client
2:39 feedback now with generative AI maybe
2:42 the review of the document can become
2:45 much faster and if we use generative AI
2:47 to create a document to summarize the
2:49 issues that the lawyer can send to the
2:50 client and discuss with them more
2:52 efficiently than the process of giving
2:54 client feedback can also become maybe
2:57 more efficient what I'm seeing is that
2:58 companies might explore building a
3:00 system like this but after having built
3:02 such a system might decide to further
3:05 change the workflow where after the
3:07 reviewing of documents may decide to add
3:10 an extra fast human verification step in
3:12 the middle just to double check that
3:14 what the Jif AI said about the document
3:17 is comprehensive and correct before then
3:20 having J of AI and the lawyer together
3:23 try to go to that last step of giving a
3:26 client feedback and so this type of
3:29 re-engineering of workflow is quite
3:31 common once both genitive AI or other AI
3:34 tools are inserted into a system because
3:37 when one task is automated or augmented
3:39 it often makes sense to wethink what are
3:41 the other tasks that need to be done
3:45 together in order to deliver a valuable
3:47 work product such as a surgical
3:49 operation or a review of a legal
3:52 document let's look at the third example
3:55 where the redesign of the workflow can
3:58 become even more sophisticated if a
4:02 marketer has to write copy for a website
4:04 that often is very time consuming and
4:06 requires a lot of thought and then after
4:08 that they might push the new website
4:10 copy to the website and that's how a
4:13 marketer might Market a new product but
4:15 with Gen AI tools let's say that the
4:17 marketer finds a way to significantly
4:19 accelerate and make more efficient the
4:22 writing of the website copy and once
4:24 writing the website copy is so efficient
4:27 it may be worth also investing in better
4:30 processes of software to make pushing to
4:32 the website more efficient and we could
4:34 end here if we do this then we just have
4:36 a marketer that's is able to do their
4:39 work much more efficiently than before
4:41 but now that we're able to write copy
4:42 and push the website much more
4:44 efficiently here's how we could rethink
4:47 the workflow which is when a marketer
4:49 wants to Market a new product they could
4:52 write some copy push the website but
4:54 then maybe write a different variation
4:57 on the copy and push that second version
4:59 to the website so they can start to
5:02 launch an AB test where you can now have
5:05 two different versions of a website copy
5:07 and start to measure which one works
5:09 better and in fact rather than just
5:11 testing two versions if we are so
5:14 efficient at writing copy and pushing
5:16 copy the website maybe we can now
5:18 generate four different versions so it's
5:21 not testing two but now four versions
5:23 and after collecting data from how all
5:25 four versions of the website copy
5:27 perform we might end up sitting down and
5:29 actually spending quite a long time
5:30 analyzing ing the campaign performance
5:34 to see which website copy works the best
5:36 and then based on that understanding go
5:39 back to further improve the website copy
5:41 Maybe by improving the prompts that we
5:45 use to generate that copy and so in this
5:48 example even though it seems like G of
5:51 AI can be used just for cost savings to
5:54 help the marketer be much more efficient
5:55 in the process of writing and pushing
5:58 copy we can also take advantage of this
6:00 efficiency in a totally different way
6:03 and write and push and test many more
6:05 versions of the website copy and this
6:07 leads to other changes as well to the
6:10 workflow of the marketer so that this
6:12 way hopefully the marketer can deliver
6:15 much more compelling marketing campaigns
6:19 now in this video we've seen a few ways
6:22 that gen of AI can not just save costs
6:25 but lead to significant growth through
6:28 new types of service offerings or
6:30 products in addition to searching for
6:33 growth opportunities by looking at tasks
6:35 done by the employees of a company
6:37 there's one other framework for analysis
6:40 that found useful that frankly isn't
6:42 done as much today but may be worth
6:45 considering for your work which is
6:48 rather than analyzing employees task
6:51 consider also analyzing the task of your
6:54 customers so for example let's say you
6:57 offer a product or service that helps
6:59 your customers to build websites
7:01 then I would urge you to consider
7:03 listing out what are the tasks that your
7:05 customers have to do not the tasks your
7:07 employees have to do but the tasks your
7:09 customers have to do in this case maybe
7:12 your customers have to first select a
7:13 template for the website to get the
7:16 initial design then they have to write a
7:18 title for the website then select images
7:20 write a copy for the homepage and then
7:23 optimize the copy for search engine
7:25 optimization to make it easy to find vir
7:27 search engines and maybe other TS as
7:31 well and if you were to analyze the geni
7:34 potential for your customer task then
7:37 you may also identify opportunities in
7:39 this case writing the title and SEO
7:42 search engine optimization as places
7:45 where gbi could help your customers and
7:48 I found that this framework for analysis
7:51 and brainstorming can sometimes lead
7:53 companies to build very different
7:56 product or services to deliver to the
7:58 customers and this can be another recipe
8:01 for getting lots of happy customers and
8:03 therefore pursuing growth for your
8:05 business so there are many ways that
8:08 gent of AI is creating value and can
8:11 lead to growth in terms of brainstorming
8:13 Frameworks I found that looking at how
8:15 you can automate or augment tons by
8:18 employees or ultimately augment TS by
8:20 your customers to be useful ways to
8:22 approach thinking about how to create
8:25 that value but if you have some other
8:27 idea that isn't arrived at by some of
8:28 these brainstorming Frameworks that's
8:30 great too and I hope you explore
8:31 building those
8:35 too now after coming with an idea of how
8:38 to use gender VII in your business some
8:40 probibly could be done via web user
8:42 interface but some may require building
8:45 a custom software application in the
8:47 next video we'll take a look at common
8:49 team structures and some best practices
8:51 for building J of AI software
8:54 applications the good news which I
8:56 alluded to in the first week as well is
8:58 that it may take less resources than you
9:00 might think because gent of AI allows
9:03 for very efficient building of AI
9:05 applications compared to some earlier
9:07 ways of AI technology but let's go take
