June 5, 2023 • 41 mins
💥 Could the dawn of artificial intelligence (AI) revolutionize the way we conduct business?
In this episode, we say our goodbyes to the Business Growth Accelerator podcast and give a hearty welcome to Leveraging AI. We unpack the untapped potential of AI, delve into strategies for business success in the AI era, and explore how AI can radically transform business operations and strategies.
💡 Topics We Discussed:
🎯 The Role of AI in Businesses and How to Navigate its Challenges: Understand the pitfalls of AI, like the potential for "hallucinating" data and how to avoid it. Learn how to formulate guidelines for good and bad AI practices.
🚀 The Need for a Continuous Education and AI Innovation Committee: Learn why staying updated about AI trends is crucial and how an innovation committee can steer the ship in the right direction.
🎩 Why Leadership is Vital in AI Strategy: Understand the significant role of leadership in AI decision-making and why it's important for leaders to be hands-on in the process.
🌐 Reevaluating Your Business Strategy: Learn why business strategies must be reassessed in the face of rapid global changes and how your business can adapt to future trends.
🔮 The Three Different Paths for Business Success in the AI Era: Discover the three main avenues to drive your business to success in this new era, from proprietary data to human relationships.
📚 Educating Your Business on AI: Understand the importance of educating your staff on the capabilities of AI and the value of implementing AI across the company.
💼 Finding Low-Hanging Fruits for AI Implementation: Learn how to identify tasks in your business that can benefit from AI for increased efficiency.
👥 Nurturing Human Relationships: Understand the importance of building human relationships in the era of AI where technology can't create a lasting moat.
As we bid adieu to the Business Growth Accelerator podcast, we're thrilled to take you on a new journey with Leveraging AI. This new podcast aims to enlighten, engage, and inspire you to harness the power of AI to push the boundaries of your business potential.
Let's dive into the future together. 🚀
Hi, It's Isar the host of the Business Growth Accelerator Podcast
I am passionate about growing businesses and helping CEOs, business leaders, and entrepreneurs become more successful. I am also passionate about relationship building, community creation for businesses, and value creation through content.
I would love it if you connect with me on LinkedIn. Drop me a DM, and LMK you listened to the podcast, what you think and what topics you would like me to cover 🙏
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Isar Meitis (00:00):
Welcome to the Business Growth Accelerator.
This Isar Meitis, your host, andthis is a a bittersweet episode
for me.
I have been recording and airingthe Business Growth Accelerator
podcast for three and a halfyears now, and I'm going to take
a break.
And today, episode is uniquebecause of that, because it's
gonna be the last show for awhile.
(00:20):
It, I'll probably come back torecording some new shows in the
future, but as of now, I'mtaking a break and I'm taking a
break because I'm focusing andI'm seeing a lot of amazing
results with my new podcast thatis called Leveraging AI.
So first of all, if you've beenlistening to the show and you
like the content and you likethe way I present it, just open
your podcasting app right nowand search for leveraging ai.
(00:43):
You'll see the same face thatyou know from Business Growth
Accelerator, different colors,obviously a different brand
name, and just subscribe to thatso you can keep on listening to
content that I bring over there.
And the content leveraging AI isstill hardcore business
practical content that you canimplement to grow your business,
but it's focused on how to bringAI capabilities into your
(01:06):
business in order to make itmore efficient, grow your
business and advance yourcareer.
But this is going to be the lastepisode of the Business Growth
Accelerator, at least in theimmediate future.
Today's episode is anintroduction to artificial
intelligence, introduction toai.
So if you have been hearing alot about artificial
(01:26):
intelligence, so maybe playingwith it a little bit and you're
trying to understand how you cantake it and implement it in your
business, this is gonna be anamazing episode for you.
If you haven't heard anything,it's time that you hear about it
and you get a betterunderstanding because it's most
likely gonna have a dramaticimpact on your business, your
industry and the business world,and the world as a whole.
(02:48):
So let's begin.
I will start by saying that youhave been using ai, whether you
know it or not for a while now,because many tools and
applications that we useregularly use AI behind the
scenes.
When you talk to Alexa or Siri,that's what actually is running.
It is an AI algorithm thatreplies on their behalf when
(03:08):
you, listen to songs on Spotifyand it suggests the next song.
That's an AI function.
When you get recommendations forproducts on Amazon or for movies
on Netflix, that's an AIfunction.
When.
You have navigation on whatevernavigation platform you're
using, and you get suggestionson where to turn and how to
turn.
There's an AI algorithm definingall of that, and so on and so
(03:29):
forth.
So you've been using AI and weas, the human race, have been
using AI and been using itregularly, daily for a while
now.
And you've also been using toolsin your work probably that has
AI infused to them.
If you've been using tools likescript for editing, videos and
audio, if you've even been usingsimple web apps like Remove bg,
(03:51):
that removes backgrounds forpictures, that's an AI tool.
So you have been using AI toolsboth on your personal life and
your business life for.
A while, so why the hell issuddenly everybody going crazy
about ai this AI that AI gen,what the hell happened?
(04:12):
What happened was on November oflast year, open ai, which is a
company that's been around for awhile.
Has released one of its modelscalled ChatGPT, to the world in
a simple chat interface, andthat started a wildfire of ai.
And the reason for that, theybrought in a mature, very
(04:35):
capable model in an interfacethat was an extremely easy to
use user interface.
It's basically one line and youcan chat back and forth with it.
And it made advanced AI languagecapabilities accessible to
basically anyone very quickly.
And that started the snowballrolling.
But what is ChatGPT?
(04:55):
ChatGPT is what's called a largelanguage model or an L L M, and
what it basically means, thinkabout somebody who's an expert
on everything that doesn'tforget any information it sees.
That is amazing at analyzingdata and using it to predict
data.
That has read every paper, everyarticle, every podcast, every
(05:21):
Slack channel it had access toand so on, that was created and
is available to the public inthe last 20 years.
So it's extremely capable inunderstanding.
Language and concepts andtranslating that into
predictions, what it actuallydoes, which is really incredible
(05:41):
when you think about theoutcome, thinking how simple it
is.
It literally just guesses thenext word in the sentence, so it
doesn't really know stuff.
It just uses a huge amount ofhistorical data on multiple
topics to create a sentence wordby word, and then the second
(06:02):
sentence and then paragraphs,and then really long segments as
long as you want on specifictopics based on historical
information and based on theinformation that you give it and
the requests that you provided.
And those requests are calledprompts.
So you can prompt it to.
Give you answers on specifictopics in specific lengths in
(06:25):
specific formats.
Playing specific roles as anexpert on whatever topic you
want.
And that's why it caught likewildfire because suddenly you,
as a random person in businessor your personal life, has
access to the accumulated datain the internet to ask its
specific questions on specifictopics and getting answers in
(06:47):
specific formats in secondswithout paying anybody.
And that's magical, and that'swhy it started rolling very
quickly.
Once ChatGPT was introduced.
What is a large language model?
What is even AI and how thewhole thing works, and how is it
different than traditionalsoftware programming?
The main difference is softwareprogramming is deterministic You
(07:11):
tell the computer what to do,and the computer does it.
In artificial intelligence ormachine learning, you actually
do not tell the computer.
Anything on what to do.
It learns just like a humanlearns.
So you give it a large amount ofdata on specific topics, and
then you start asking itquestions based on that data,
(07:34):
and then you provide it feedbackon how it's doing with its
answers, And then you follow thecycle again and again, giving it
more and more feedback.
It learns just like a humanwould.
What's correct, what'sincorrect?
What's a good answer, what's abad answer?
And that's how it learns thedifferent aspects.
This process is called traininga model, and these models comes
(07:56):
in differents modals, meaningthey can do different things.
Large language models are likethe name suggests, they're
language modes, but there's alsomodes that can write code that
can create or analyze imagesthat can create or analyze music
that can create or manipulatevideo.
And so on, and there's evenmultimodality, meaning there are
different tools out there thatcan do more than one of those
(08:19):
things.
Then eventually we will get intowhat's called AGI or artificial
general Intelligence, which willbe a AI entity that will be able
to do everything that a humandoes at a human level or above,
rather than doing it at a humanlevel or above on specific
(08:40):
different aspects like ithappens today.
Today we already have all thesestand-alone ones, and there's a
few multimodal tools.
We still don't have agi,probably every company out there
that is working in this field,that's their holy grail, and
that's what they're workingtowards.
And there's a big debate on whenthat's gonna happen.
Is it gonna happen in a year, infive years, or maybe never?
It depends who you ask.
(09:00):
And these answers are comingfrom people who are way smarter
than me.
So the biggest expert in theindustry that has been
delivered, that have beendeveloping this for years, have
different opinions on when wewill reach AGI or Super
Intelligence, which even aboveAGI level.
So what can you do?
With these AI tools that canhelp you in either your personal
(09:23):
or your business life.
So let's take some of them oneby one and give a few examples
and we'll start with largelanguage models because they're
probably the most commonly usedright now.
So what can the model do?
First of all, you need tounderstand that the model
understands the language,meaning things like text to
speech when you speak to yourphone and it writes on your
(09:43):
behalf is, Language model of anAI natural language processing.
So when you speak and itactually understands what you're
saying, like Siri or Google or,Alexa, all of these use natural
language processing in order tounderstand what you're saying.
But because it can do it in somany different languages, it
also can translate from onelanguage to the other and I'm
(10:05):
not talking about justtranslating literally word to
word because that sometimesdoesn't make any sense in the
other language.
But actually understanding thelanguage, understanding the
context, understanding whatyou're trying to say, and saying
the right sentence in the otherlanguage.
So translation is a big benefitof that.
But it also knows how to analyzelarge amount of text data and
(10:26):
produce text data based on thedata that it has consumed and
was trained on.
So what can you do with them inyour work?
you can write anything.
It can help you write detailedemails or short emails.
It can help you write Slackmessages on specific topics that
you need to update the team.
It can help you write PowerPointpresentations, social media,
(10:47):
post full blog post, et cetera.
So basically anything you needto write it can help you write.
And the more data you give itand the more access you give it
to information within yourbusiness, it will be able to do
a better and better job inrepresenting the way you want to
write like your voice, yourtone, the company's brand
guidelines, et cetera.
(11:08):
But in addition to writing, itcan also help you analyze data.
now, for a very long time, wehad multiple tools who are very
good at analyzing quantitativedata.
Anything from Excel, GoogleSheets, to dedicated tools on
specific tasks in accounting orsales, et cetera.
But now for the first time, wehave a tool that can analyze
(11:31):
qualitative data such ascustomer reviews, such as
self-reported attribution sa,such as sentiments on social
media mentions such as theeffectiveness of sales calls.
How is our brand representedacross the communication, all
these kind of things.
It can analyze because itunderstands the language, it's
(11:53):
very good at qualifying text.
And going back to what I saidbefore, recordings can be
translated into text and thenthe text can be analyzed with
the other tools.
So it doesn't matter how thelanguage is available, whether
it's in video, audio, or text,you can convert it to text and
then do the analysis and thenget very detailed analysis that
was almost impossible to dobefore large amounts of data
(12:14):
because you had to have humansin the loop doing it.
It's obviously also very good inanalyzing quantitative data
better than we had before.
So analyzing large quantities ofdata, whether financial
performance.
Sales velocity.
Any other numerical informationwe have, aI can be very helpful
in analyzing and making andgetting into relevant,
(12:34):
actionable conclusions.
What else can you do with AIlanguage related tools?
So creating chat bots.
Creating internal chat bots thatcan help people within the
company understand what's goingon, get answers on anything we
need, instead of going to Sheilafrom accounting or to Dave from
sales because they're theexperts on something.
You can take the organizationalknowledge, train a bot on that,
(12:55):
and make it available to anybodyin the company to get immediate
answers, accurate answers onmultiple topics.
You can use similar chat botsfor external communications,
whether for customer service,for sales, for marketing, for
answering questions that peoplehave on the website without
having to actually communicatewith the employees of the
company using the same conceptsof training, the model on
information you want to give it.
(13:17):
What tools can you use to dothat?
First of all, they're the corelanguage models themselves, so
ChatGPT from OpenAI, Bard fromGoogle, Claude from
Philanthropic are probably thebiggest names you can think of.
But there are a lot of otherlarge language models.
Some are closed, some are opensource that you can use for many
different tasks.
But since those tools all haveAPIs, many companies created
(13:41):
tools that are built on top ofthe underlying large language
model layer.
And so you have the largelanguage model layer provided by
the companies I just mentioned,and then you have an application
layer developed by othercompanies, helping solve very
specific issues on very specificuse cases such as writing
assistants, seo, assistantmeeting, transcription and
(14:04):
summary, et cetera, et cetera.
Like almost anything you canimagine that has to do with
understanding and analyzinglanguage.
There's dedicated tools todaythat are using an underlying
level of a large language modelfrom one provider another, and
in some cases more than one.
So these are a lot of exampleson the language modality.
(14:24):
Now let's talk about image.
AI with images can do one ofthree things.
It can generate images based ona prompt out of thin air,
meaning you can describe to itexactly what you want in the
image, what type of image youwant.
Do you want a Renaissancepainting?
Do you want a photo realisticimage?
Do you want a cartoon, likewhatever style you want it, it
(14:46):
will be able to do it, and youjust tell it what you want to
have in the image and it willcreate it for you in seconds.
What does that mean for yourbusiness?
First of all, it means youshould stop paying for any stock
photo membership that you haveright now, and you can start
generating photos on the fly.
It will be faster than searchingthrough the stock photos.
(15:07):
It will be exactly the image youneed.
It will be unique because it's aone-off.
It was created specifically foryou, and it's either free in
some tools or almost free onsome of the other tools.
So you can create any image forany need, whether it's creating
emails, whether it's PowerPointpresentations for sales or for
(15:28):
your next speaking gig, whetherit is for newsletters or blog
posts, social media, any imageyou need in any format, in any
style, in any resolution you cancreate with these tools in
seconds, and it's absolutelyamazing.
What tools can you use for that?
The one that I use most, becauseI find that right now it
provides the best results is midjourney.
(15:50):
It actually does cost money, andthe only way to use it right now
is through a discord server,which is not the most user
friendly thing to do, but it's,as far as I'm concerned right
now, the best in getting youexactly what you need in the
best resolution, in the mostaccuracy.
Next one on the list isprobably, stable diffusion from
stability AI, and the third oneis Dall-e from OpenAI.
(16:11):
The same company that developedChatGPT.
But again, there are multipleothers that you can use.
The last two that I mentionedare completely free.
So if you're using stablediffusion in its various forms
and shapes through differentwebsites, or if you're using
Dall-e, it's completely free touse.
That's for image generation, butit also knows how to analyze
(16:32):
images.
You can actually upload imagesto an AI platform and it can
tell you what is in the image orwhat's gonna happen next, or
whatever question you want toask it When it comes to image
processing.
Now some of the more advancedtools allow you to edit images
based on understanding what's inthe image, analyzing the
background and the ability tomake really crazy changes to
(16:54):
existing photos.
change the pose of an animal ora person that you actually took
a picture off.
So you have a static image andyou can make the person in the
image turn their head or a lionroar.
While it wasn't when you tookthe original picture, just by
dragging and dropping.
And because it understands whata person is and how that person
looks like, and it can guess howit's gonna look like when that
(17:15):
person is looking to the left orto the right and have the
lighting perfect and so on.
So you can manipulate images,you can also add things to
images without knowingPhotoshop.
So, Adobe, recently introducedAdobe Firefly.
That allows you to write whatyou want to add to an image,
circle the area where you wannaadd it, and it will add it in a
very realistic lighting, etcetera, that it would look as if
(17:36):
it's a part of the photo.
In the same way, you can removethings from images, like people
that stood in the way when youtook a picture of a famous
monument you were visiting, youcan take those people out of the
picture and it will fill out themissing pieces.
The last thing that it knows howto do it knows how to expand
images.
So if you have an image ofsomething and you want the AI to
(17:59):
make stuff around it, that wouldlook realistic as if it was a
part of the original imagetaken.
It knows how to do that as well.
So there's multiple ways tocreate images, analyze images,
and manipulate images.
All that can be used for anykind of business task that you
can imagine from ad creation tosocial post and everything I
mentioned before.
The next big domino that'sprobably gonna happen that's
(18:20):
gonna have a huge impact onbusinesses and the way we
produce different kinds ofcontent is video.
There's already initial teststhat are working that you can
play with that allows you towrite a prompt and create a
short video from it out of thinair without a camera, without
editing, without anything.
It's still not amazing, but it'smoving in that direction.
The other aspect of this isobviously deep fakes, is the
(18:43):
ability to manipulate a video tolook like somebody else, to
sound like somebody else, or.
To have a completely realisticview of a situation that doesn't
really happen.
That's obviously extremely scaryon one hand, but if you think
about the creative possibilitiesfor a business, they are
literally endless becausesomething that used to cost tens
(19:03):
of thousands or hundreds ofthousands of dollars and take
months.
Can now take minutes with onesingle person in a home
computer.
So the ability to create stuffthat is professional grade, just
based on the fact you're reallycreative, becomes really
incredible.
And at the reach of anybody thathas good creative skills.
(19:26):
now that you have a basicunderstanding of what is
artificial intelligence Andlarge language models and
differents and what they can doand how can it work in your
business?
How do you approach this?
How do you really get started?
What are the things you need todo?
A checklist as a business leaderor somebody who aspires to be in
business or just as somebody inthe organization that wants to
(19:47):
step up and saying, I'm gonnahelp my company understand how
to implement ai.
I'm gonna start with someconcepts before we dive into the
checklist.
And the first concept is thatwe're moving from a very long
era, hundreds of years of tryingto improve efficiencies of
processes to an era where we canjust get the outcome without
(20:11):
having to worry about theprocess.
And I'll give you a fewexamples.
Example, number one, customerservice, multiple tools and
capabilities and processes andbest practices.
Are being continuously updatedin order to have better customer
service.
How can we as answer faster?
How can the first answer be theright answer, et cetera, et
(20:33):
cetera.
All these kind of things.
Entire companies are builtaround building these tools that
will help other companiesprovide better customer service.
But the goal is not providingmore efficient customer service.
The goal is having happycustomers that can get a answer
in a very short amount of time,and it's still one of the
biggest frustrations for peopledealing with companies and the
(20:56):
bigger the companies they'redealing with, usually the worst
the customer service is.
But what if we can train an AIon all the knowledge that a
customer service person has toknow and also give you tools to
take all the actions that theyneed, upgrade them, downgrade
them, send them the bill, revoketheir rights, add rights to
(21:18):
them.
Anything that a customer serviceextremely capable agent can do,
give that to the ai.
Then you have the outcome.
You have happy customers thatget immediate results because
they don't have to wait.
And they don't have to talk tothree different people and they
don't have to re-explainthemselves, and they don't have
to go through a phone IVR orthrough 20 different forms.
(21:40):
They just go in, they type orspeak what they actually need,
and they will get an answer,most likely, the right answer
immediately.
So you're going and you'recircumventing the process.
You're not trying to make theprocess more efficient anymore,
you're just going straight tothe outcome.
Another example, you.
Want to rank higher on Google,so you had to do SEO keyword
(22:04):
research.
You had to understand what's thecompetitive landscape looks
like.
You had to start definingdifferent topics you need to do.
You had to then create outlinesthat will be aligned with those
topics that you need to cover.
You have to design the websiteaccordingly.
Then you have to actually writethe blog post, optimize them for
SEO, and then deploy them.
That's a very long process thattakes a lot of people with a lot
(22:26):
of expertise and takes a lot oftime.
And now you can just tell it.
Here's my competition.
Tell me what keywords they'reranking for.
I wanna rank higher than them.
Write as many articles as youneed and set them up in the
right places, in the rightformat, in the right website,
and you get it.
Now what does that mean longterm for SEO?
I don't have a clue.
I think what I just mentioned isan amazing short-term
(22:48):
opportunity that I think will goaway once a big amount of
company will start doing that.
In addition, what's happeningfor SEO is that because chatbots
like Bard and Bing and ChatGPTare going to probably replace
traditional search because ofthe same reason we used to try
to make search better andbetter.
So Google leading this chargeare the best in giving you the
(23:12):
best search results for whatyou're looking for.
But then you still have to gothrough multiple links and try
to figure out the bits andpieces of information from each
link that actually answers yourquestion versus.
The outcome is I just want theanswer, and if the chat bot can
give it to me, I don't need toclick any links.
So the whole concept of SEO willerode either quickly or slowly,
but in the next few years, itwill either be very different
(23:33):
than it is today or disappearedcompletely.
So the first concept is reallyunderstanding that we're moving
from.
Improving process efficienciesto moving to this is the
outcome.
Can I get the outcome withoutgoing through the process?
I'll touch on a few additionalconcepts once I go through the
checklist, but the first thingon checklist, and is the most
(23:54):
important thing is education.
you have to commit to continuouseducation on the topic of ai.
Why am I saying that?
I'm saying that because the AIworld is moving so fast, there's
really amazing news, amazing newtools, new capabilities that are
coming out daily and whateveryou learn right now may be
(24:16):
obsolete and Middle Ages like.
A month from now.
So do you have to do this everysingle day?
No.
But do you have to keep yourselfcontinuously educated on what's
happening in this field so youcan at least not be left far
behind?
Absolutely.
Yes.
So that's the number onerecommendation.
How do you do that?
find the right podcasts likeleveraging AI or any other
(24:37):
podcast.
There's multiple other podcastsout there that will educate you
on that.
Read books, read articles,follow the right people on
social media that share tips andtricks and strategies around ai.
Keep educating yourself on thistopic so you can stay, if not
ahead of the curve, at least onthe curve and not behind it.
The next step is you do not wantto take this path alone.
(25:01):
Why?
Because you have your expertiseand your capabilities, but you
also have a job to perform, andyou're limited with your
understanding, and you'relimited with your area of
expertise.
So the right thing to do withinevery business today is to start
a aI Innovation Committee.
And the people on the committeeneeds to be people who, a, are
quote unquote geeks and willactually enjoy doing this.
(25:21):
B people who come from differentdepartments because they will
represent different needs anddifferent mindsets and different
approaches while looking throughthe AI lens on different
processes within the company.
those people need.
Time on their calendar to do AIrelated things, and we'll talk
in a minute on what they shoulddo.
And they need access to theright resources, whether it's a
(25:43):
sandbox to play with, so theydon't mess up anything that is
currently company infrastructureor budget to have access to
tools they want to test.
So they need resources in orderto do their work correctly.
What does this committee needsto do?
First thing is they're in chargeof educating themselves and the
company all the time.
That means, Learning aboutdifferent tools.
(26:05):
It means experimenting withdifferent tools for different
use cases.
It means developing processeswhen they find a tool that could
actually be used within thecompany.
Developing the processes on howto implement the tool within the
business, and then training thepeople within the company on how
to use the tool.
They also become the go-to groupwhen anybody has an AI related
question and they need todevelop a few very specific
(26:29):
deliverables.
One deliverable is.
AI guidelines, what each personin the company should know that
they should do or shouldn't dowhen it comes to using ai.
Just recently, a lawyer arguinga case for an airline in a court
in the US referenced sixdifferent cases in order to
(26:50):
argue his case.
That do not exist.
So think about the embarrassmentto A, the lawyer, and B, the law
firm that he's representing whenhe presented cases that don't
exist.
And the reason they don't existis because he used ChatGPT in
order to do the research.
And maybe the biggest problemthat these large language models
have is they hallucinate, theymake stuff up, and the bad news
(27:14):
is you have no clue.
When they're factual and whenthey actually make stuff up,
meaning you have to fact checkthem in these kind of scenarios.
But he did not know that.
And hence he used ChatGPT to dothe research, went to court, and
I'm sure again, he was extremelyembarrassed when he found out
that these cases don't actuallyexist.
(27:34):
So creating guidelines for goodand bad is extremely important,
and that's something thecommittee has to define.
Part of the people in thecommittee have to come from the
company's leadership a because.
It will involve making decisionsand B, because that helps lead
by example for the people inleadership to show the rest of
(27:55):
the company that they arehands-on involved in the
process.
But the other reason to haveleadership people within that
process is because one of themost important things you have
to do right now is to reevaluateyour entire business strategy.
Why do you do that?
Because the world is changingextremely fast, faster than it
has ever changed before, and theneeds of clients and the
(28:19):
capability to deliver thoseneeds is changing very fast.
You have to go and after you'veeducated yourself and have a
decent understanding of whatpeople can do today and what
they might be able to do a yearfrom now or two years from now
is go and reevaluate who aregonna be your customers, what
will they be willing to pay forin a year or two?
(28:42):
And the flip side, what otheropportunities within the scales
and the expertise you have inyour company are now open?
That did not exist before.
So let's take two quickexamples.
One for each side.
And I'll take two very extremeexamples.
One example, there are multiplecompanies today who offer SEO
(29:02):
writing services.
They will go extinct veryquickly.
Why?
Because the entire thing thatthey do can now be done by a
machine faster and in manycases, probably better than
these people can write.
And yes, I know some peopledisagree with me and will say
that AI writing is not as goodas human writing.
I think that's not true.
(29:22):
I think it's how you use andleverage the AI tools in order
to write, in order to get theresults you want.
But it's definitely way moreefficient than letting humans
write articles for you rightnow, assuming you are fact
checking it in the end.
As I mentioned before, samething goes by the way for
various types of consulting andtraining.
if you are writing onlinecourses today and that's your
(29:43):
main income, Your income mightbe at serious risk because there
is no doubt in my mind that thebig platforms like Coursera will
develop an AI model that cancreate courses on the fly,
customized to the specific needsof each person.
So instead of going and browsingthrough existing courses, you
will say, I need a beginner'scourse on topic X.
(30:06):
It's for X amount of people.
That's their level ofexperience.
That's what they know right now.
The course has to be withvideos.
It needs tests within it.
And we have six hours to takethe course spread over six weeks
with one hour a week, and youwill create that course for you.
It might even send a preliminaryquestionnaire to each and every
one of the people, so it cancustomize it further to each
(30:27):
individual needs.
So if you're creating coursesonline, as your main source of
income that may go away withinthe next year or two.
The flip side of that, if youhave a lot of data, like if
you're a large consultingcompany, and you can take all
that data and train models withthat, you can now provide much
more efficient, faster, moreaccurate consulting services
(30:50):
than ever before.
You can go after lower tieredclients that couldn't pay you
before because now you canautomate a lot of the processes.
So there's.
Amazing opportunities as well asbig risks.
And that's true for almostanything unless you're cutting
grass or manufacturing somethingvery specific.
And even there, there's probablyefficiencies to be made through
(31:11):
supply chain management usingai, et cetera.
when looking at buildingstrategy, there's a few key
points that are different thantraditional strategy, and I'll
mention them right now.
The first thing is I see threedifferent paths for business
success in the AI era.
Path.
Number one is, as I mentioned,proprietary data.
If you own proprietary data, andthe more of it you own, the
(31:34):
better you can train models.
On that data and provide valuethat nobody else can provide, or
very few companies who haveaccess to this kind of data can
provide.
This way, you can stand out andcreate a blue ocean for
yourself.
That was not possible before.
The second option is exactly theopposite side of the scale is
(31:55):
you're in a highly competitivemarket and you are the best in
implementing these AI tools toget better efficiencies.
And hence you have bettermargins and can be more
competitive with your pricingand win through efficient
implementation and usage of AItools versus your competition
that may not be as good as that.
the third path for success ishuman relationships.
(32:16):
And that becomes amplified inthe B2B world because so many
companies will be able to doalmost everything really well
and vanilla like everybody else,because using those tools will
allow everybody to do thisrelatively easily and without
huge amount of resources.
The human connections are gonnabe a huge differentiator.
So that's another path that Ithink every company should have
(32:36):
done so far, but becomes waymore important moving into the
AI future.
the last concept that I wannamention when it comes to
thinking on the strategy ofusing AI before we dive into the
rest of the checklist is theunderstanding that the existing
concept of diminishing returnsmay not apply anymore to
(32:57):
different aspects of how yourbusiness runs.
An example I used earlier is,let's say you're writing
capability right now to createblog posts.
You can create three a month orthree a week, and now you can
create 300 in a day.
and I can give you multipleexamples like that, but the
ability to take something thatwas a significant bottleneck and
had diminishing returns in orderto scale, it can now scale
(33:20):
infinitely.
With almost zero marginal cost.
And that concept of looking forthings in your business, that AI
allows you to scale dramaticallywith zero or almost zero
marginal cost, leading you tolimitless growth on those
particular topics.
If these things are majorbottlenecks of your growth right
(33:41):
now, they can go away,immediately.
It's not a process.
It's a almost a zero to one gamethat happens almost overnight
once you figure out how to applythese tools to those specific
bottlenecks.
You have to have that thing inmind when you are analyzing your
current and your futurestrategy.
(34:02):
Now I want to go back to thechecklist.
We said continuous education,number one.
We said AI innovation committeeas number two, creating
guidelines as one of the thingsthe committee has to do,
educating the company.
You must continuously educateother people in the business.
You can do this through talkingabout these things in town Hall.
You can share this on Slackchannels.
(34:22):
You can define specific thingsas mandatory reading or
mandatory listening, and thenhave people come in smaller
groups in their departments, inwhatever teams they're in, and
share what they've learned fromthe things that was mandatory to
read or listen.
There are multiple ways on howto educate your business, but
you have to have multiplepeople, preferably everybody in
the business educated on AI andits capabilities and how it can
(34:44):
be leveraged in order to run thebusiness in a more efficient
way.
You need to define KPIs on ai,and yes, there's.
Trailing indicators such as wenow have more sales or our
margins are better, but thatwill take a very long time.
But what you can start with asKPIs is Lin leading indicators.
How many people are using AI inthe company right now?
(35:05):
How many processes we haveimplemented AI into?
How many tools have thecommittee evaluated even if we
haven't implemented them, and soon.
So these are great leadingindicators to see that you're
moving in the right directionfor implementing AI across your
company in a successful way.
Another important aspect isethical guidelines, because AI
allows you to make stuff up.
(35:27):
It can lead into doing thingsthat are beyond the red lines of
your company, yourself, yourcore values, et cetera.
And that line is different forevery individual and for every
different company.
But you have to define what'sacceptable and not acceptable to
do in your business usingartificial intelligence.
Next thing on the checklist Isencourage AI usage.
(35:50):
How do you do that?
share success stories.
This person from that departmenthas done this and was able to
achieve that result.
do this in company town hall.
Invest a few minutes in sharingthose kind of things.
Show that you are doing it as abusiness leader.
Again, lead by example.
People will see you doing this.
They will understand it'simportant to you, hence they're
gonna follow your footsteps.
(36:11):
Reward people that successfullyimplement AI capabilities and so
on.
There's multiple ways to dothat, but the usage of AI has to
be encouraged within.
The guidelines that were definedby the committee, so nobody does
anything stupid that actuallyhurts the business.
Next topic is look for lowhanging fruits for
implementation.
How do you look for low hangingfruits?
(36:31):
AI is really good at veryspecific things.
It's very good at doingrepetitive tasks.
It's very good at data analysis.
It's very good at dataprediction.
So taking the existing data andextrapolating from that, and
it's very good at generatingcontent at scale.
So any task that falls into anyof those buckets will be a
(36:52):
relevant thing that AI can helpyou make more efficient.
If you have tasks that fall intomore than one of these buckets.
Then the benefits you will getfrom implementing AI for those
tasks will be amplified by thefact it gets tick in several of
those different boxes.
So again, repetitive tasks, dataanalysis, data prediction, and
content creation.
So what you need to do, you needto look at all the tasks that
(37:15):
the company's doing, andhopefully you have some kind of
a task management tool likeTrello or Jira or Monday or
ClickUp or whatever other toolthat you're using, or even just
an Excel spreadsheet that haswhat tasks people are doing and.
See which of those buckets theyfall into, and that will help
you prioritize the task, whichwill get you the most amount of
(37:36):
benefit by leveraging differentAI tools to solve or help within
these tasks.
Once you have them mapped andyou have them prioritize, look
for the right tools, give themto the committee, have the
committee evaluate them, andonce you pick up and define the
right process, implement itacross the business and measure
the amount of time or efficiencythat you're actually getting as
an outcome of thatimplementation so you can learn
(37:58):
for the next cycle.
The last component has to dowith what I mentioned before.
How do you nurture humanrelationships within the
business and outside of thebusiness in order to create a
moat that is not technological?
Because technological moats willbe very hard to create in this
new era because it will beextremely easy to catch up to
almost anything you can develop.
(38:20):
So a quick summary of everythingI shared with you today.
The first thing is AI is here tostay.
It's gonna have dramatic impacton society and on businesses and
on individuals, and the best wayto come ahead out of this
revolution is to a continuouslyeducate yourself, experiment
with it.
Find the right ethical ways toimplement it in your personal
(38:42):
life and in your business, andeducate the people around you so
it can have more brainstormingapproach to this.
Revalue your business strategy.
Find fruits and keep runningfaster than your competition.
That's it for this episode and Ihope you found this helpful.
And this is obviously just anintroduction to AI in
businesses.
If you wanna learn more, justcheck out additional episodes of
(39:04):
leveraging AI where me and myguests dive into specific use
cases, specific strategicapproaches that can help you in
a much more practical way thanthis introduction.
I wanna thank you so much fortaking the time and listening to
this.
If you enjoy this episode.
If this was helpful to you.
Please share it with yourfriends or people that you think
(39:25):
can benefit from this.
please give us a five star andwrite a review on the platform
you're on.
That really helps us reach morepeople, which means you can help
more people understand AI andhelp them implement it in their
businesses.
Thank you so much and until nexttime, have an amazing week.
(39:54):
As I mentioned the beginning ofthis episode, this is at least
for now, gonna be the lastepisode of Business Growth
Accelerator, and if you wantmore of me and my approach to
business and how to grow it,just sign up for leveraging ai.
If you've been listening to theshow for a while, and I know
some of you have been aroundwith me from the beginning, so
three and a half years, I reallytruly, I.
(40:16):
Appreciate the time that you'veinvested in listening to the
show.
I know you have other things youcan do in that time and other
podcasts you can listen to.
And I do not take this forgranted and I'm promising you
I'm investing the same energyand the same intellect with a
lot of renewed energy intoleveraging ai So go check it out
(40:37):
and thank you so much
Welcome to the Business Growth Accelerator.
This Isar Meitis, your host, andthis is a a bittersweet episode
for me.
I have been recording and airingthe Business Growth Accelerator
podcast for three and a halfyears now, and I'm going to take
a break.
And today, episode is uniquebecause of that, because it's
gonna be the last show for awhile.
(00:20):
It, I'll probably come back torecording some new shows in the
future, but as of now, I'mtaking a break and I'm taking a
break because I'm focusing andI'm seeing a lot of amazing
results with my new podcast thatis called Leveraging AI.
So first of all, if you've beenlistening to the show and you
like the content and you likethe way I present it, just open
your podcasting app right nowand search for leveraging ai.
(00:43):
You'll see the same face thatyou know from Business Growth
Accelerator, different colors,obviously a different brand
name, and just subscribe to thatso you can keep on listening to
content that I bring over there.
And the content leveraging AI isstill hardcore business
practical content that you canimplement to grow your business,
but it's focused on how to bringAI capabilities into your
(01:06):
business in order to make itmore efficient, grow your
business and advance yourcareer.
But this is going to be the lastepisode of the Business Growth
Accelerator, at least in theimmediate future.
Today's episode is anintroduction to artificial
intelligence, introduction toai.
So if you have been hearing alot about artificial
(01:26):
intelligence, so maybe playingwith it a little bit and you're
trying to understand how you cantake it and implement it in your
business, this is gonna be anamazing episode for you.
If you haven't heard anything,it's time that you hear about it
and you get a betterunderstanding because it's most
likely gonna have a dramaticimpact on your business, your
industry and the business world,and the world as a whole.
(02:48):
So let's begin.
I will start by saying that youhave been using ai, whether you
know it or not for a while now,because many tools and
applications that we useregularly use AI behind the
scenes.
When you talk to Alexa or Siri,that's what actually is running.
It is an AI algorithm thatreplies on their behalf when
(03:08):
you, listen to songs on Spotifyand it suggests the next song.
That's an AI function.
When you get recommendations forproducts on Amazon or for movies
on Netflix, that's an AIfunction.
When.
You have navigation on whatevernavigation platform you're
using, and you get suggestionson where to turn and how to
turn.
There's an AI algorithm definingall of that, and so on and so
(03:29):
forth.
So you've been using AI and weas, the human race, have been
using AI and been using itregularly, daily for a while
now.
And you've also been using toolsin your work probably that has
AI infused to them.
If you've been using tools likescript for editing, videos and
audio, if you've even been usingsimple web apps like Remove bg,
(03:51):
that removes backgrounds forpictures, that's an AI tool.
So you have been using AI toolsboth on your personal life and
your business life for.
A while, so why the hell issuddenly everybody going crazy
about ai this AI that AI gen,what the hell happened?
(04:12):
What happened was on November oflast year, open ai, which is a
company that's been around for awhile.
Has released one of its modelscalled ChatGPT, to the world in
a simple chat interface, andthat started a wildfire of ai.
And the reason for that, theybrought in a mature, very
(04:35):
capable model in an interfacethat was an extremely easy to
use user interface.
It's basically one line and youcan chat back and forth with it.
And it made advanced AI languagecapabilities accessible to
basically anyone very quickly.
And that started the snowballrolling.
But what is ChatGPT?
(04:55):
ChatGPT is what's called a largelanguage model or an L L M, and
what it basically means, thinkabout somebody who's an expert
on everything that doesn'tforget any information it sees.
That is amazing at analyzingdata and using it to predict
data.
That has read every paper, everyarticle, every podcast, every
(05:21):
Slack channel it had access toand so on, that was created and
is available to the public inthe last 20 years.
So it's extremely capable inunderstanding.
Language and concepts andtranslating that into
predictions, what it actuallydoes, which is really incredible
(05:41):
when you think about theoutcome, thinking how simple it
is.
It literally just guesses thenext word in the sentence, so it
doesn't really know stuff.
It just uses a huge amount ofhistorical data on multiple
topics to create a sentence wordby word, and then the second
(06:02):
sentence and then paragraphs,and then really long segments as
long as you want on specifictopics based on historical
information and based on theinformation that you give it and
the requests that you provided.
And those requests are calledprompts.
So you can prompt it to.
Give you answers on specifictopics in specific lengths in
(06:25):
specific formats.
Playing specific roles as anexpert on whatever topic you
want.
And that's why it caught likewildfire because suddenly you,
as a random person in businessor your personal life, has
access to the accumulated datain the internet to ask its
specific questions on specifictopics and getting answers in
(06:47):
specific formats in secondswithout paying anybody.
And that's magical, and that'swhy it started rolling very
quickly.
Once ChatGPT was introduced.
What is a large language model?
What is even AI and how thewhole thing works, and how is it
different than traditionalsoftware programming?
The main difference is softwareprogramming is deterministic You
(07:11):
tell the computer what to do,and the computer does it.
In artificial intelligence ormachine learning, you actually
do not tell the computer.
Anything on what to do.
It learns just like a humanlearns.
So you give it a large amount ofdata on specific topics, and
then you start asking itquestions based on that data,
(07:34):
and then you provide it feedbackon how it's doing with its
answers, And then you follow thecycle again and again, giving it
more and more feedback.
It learns just like a humanwould.
What's correct, what'sincorrect?
What's a good answer, what's abad answer?
And that's how it learns thedifferent aspects.
This process is called traininga model, and these models comes
(07:56):
in differents modals, meaningthey can do different things.
Large language models are likethe name suggests, they're
language modes, but there's alsomodes that can write code that
can create or analyze imagesthat can create or analyze music
that can create or manipulatevideo.
And so on, and there's evenmultimodality, meaning there are
different tools out there thatcan do more than one of those
(08:19):
things.
Then eventually we will get intowhat's called AGI or artificial
general Intelligence, which willbe a AI entity that will be able
to do everything that a humandoes at a human level or above,
rather than doing it at a humanlevel or above on specific
(08:40):
different aspects like ithappens today.
Today we already have all thesestand-alone ones, and there's a
few multimodal tools.
We still don't have agi,probably every company out there
that is working in this field,that's their holy grail, and
that's what they're workingtowards.
And there's a big debate on whenthat's gonna happen.
Is it gonna happen in a year, infive years, or maybe never?
It depends who you ask.
(09:00):
And these answers are comingfrom people who are way smarter
than me.
So the biggest expert in theindustry that has been
delivered, that have beendeveloping this for years, have
different opinions on when wewill reach AGI or Super
Intelligence, which even aboveAGI level.
So what can you do?
With these AI tools that canhelp you in either your personal
(09:23):
or your business life.
So let's take some of them oneby one and give a few examples
and we'll start with largelanguage models because they're
probably the most commonly usedright now.
So what can the model do?
First of all, you need tounderstand that the model
understands the language,meaning things like text to
speech when you speak to yourphone and it writes on your
(09:43):
behalf is, Language model of anAI natural language processing.
So when you speak and itactually understands what you're
saying, like Siri or Google or,Alexa, all of these use natural
language processing in order tounderstand what you're saying.
But because it can do it in somany different languages, it
also can translate from onelanguage to the other and I'm
(10:05):
not talking about justtranslating literally word to
word because that sometimesdoesn't make any sense in the
other language.
But actually understanding thelanguage, understanding the
context, understanding whatyou're trying to say, and saying
the right sentence in the otherlanguage.
So translation is a big benefitof that.
But it also knows how to analyzelarge amount of text data and
(10:26):
produce text data based on thedata that it has consumed and
was trained on.
So what can you do with them inyour work?
you can write anything.
It can help you write detailedemails or short emails.
It can help you write Slackmessages on specific topics that
you need to update the team.
It can help you write PowerPointpresentations, social media,
(10:47):
post full blog post, et cetera.
So basically anything you needto write it can help you write.
And the more data you give itand the more access you give it
to information within yourbusiness, it will be able to do
a better and better job inrepresenting the way you want to
write like your voice, yourtone, the company's brand
guidelines, et cetera.
(11:08):
But in addition to writing, itcan also help you analyze data.
now, for a very long time, wehad multiple tools who are very
good at analyzing quantitativedata.
Anything from Excel, GoogleSheets, to dedicated tools on
specific tasks in accounting orsales, et cetera.
But now for the first time, wehave a tool that can analyze
(11:31):
qualitative data such ascustomer reviews, such as
self-reported attribution sa,such as sentiments on social
media mentions such as theeffectiveness of sales calls.
How is our brand representedacross the communication, all
these kind of things.
It can analyze because itunderstands the language, it's
(11:53):
very good at qualifying text.
And going back to what I saidbefore, recordings can be
translated into text and thenthe text can be analyzed with
the other tools.
So it doesn't matter how thelanguage is available, whether
it's in video, audio, or text,you can convert it to text and
then do the analysis and thenget very detailed analysis that
was almost impossible to dobefore large amounts of data
(12:14):
because you had to have humansin the loop doing it.
It's obviously also very good inanalyzing quantitative data
better than we had before.
So analyzing large quantities ofdata, whether financial
performance.
Sales velocity.
Any other numerical informationwe have, aI can be very helpful
in analyzing and making andgetting into relevant,
(12:34):
actionable conclusions.
What else can you do with AIlanguage related tools?
So creating chat bots.
Creating internal chat bots thatcan help people within the
company understand what's goingon, get answers on anything we
need, instead of going to Sheilafrom accounting or to Dave from
sales because they're theexperts on something.
You can take the organizationalknowledge, train a bot on that,
(12:55):
and make it available to anybodyin the company to get immediate
answers, accurate answers onmultiple topics.
You can use similar chat botsfor external communications,
whether for customer service,for sales, for marketing, for
answering questions that peoplehave on the website without
having to actually communicatewith the employees of the
company using the same conceptsof training, the model on
information you want to give it.
(13:17):
What tools can you use to dothat?
First of all, they're the corelanguage models themselves, so
ChatGPT from OpenAI, Bard fromGoogle, Claude from
Philanthropic are probably thebiggest names you can think of.
But there are a lot of otherlarge language models.
Some are closed, some are opensource that you can use for many
different tasks.
But since those tools all haveAPIs, many companies created
(13:41):
tools that are built on top ofthe underlying large language
model layer.
And so you have the largelanguage model layer provided by
the companies I just mentioned,and then you have an application
layer developed by othercompanies, helping solve very
specific issues on very specificuse cases such as writing
assistants, seo, assistantmeeting, transcription and
(14:04):
summary, et cetera, et cetera.
Like almost anything you canimagine that has to do with
understanding and analyzinglanguage.
There's dedicated tools todaythat are using an underlying
level of a large language modelfrom one provider another, and
in some cases more than one.
So these are a lot of exampleson the language modality.
(14:24):
Now let's talk about image.
AI with images can do one ofthree things.
It can generate images based ona prompt out of thin air,
meaning you can describe to itexactly what you want in the
image, what type of image youwant.
Do you want a Renaissancepainting?
Do you want a photo realisticimage?
Do you want a cartoon, likewhatever style you want it, it
(14:46):
will be able to do it, and youjust tell it what you want to
have in the image and it willcreate it for you in seconds.
What does that mean for yourbusiness?
First of all, it means youshould stop paying for any stock
photo membership that you haveright now, and you can start
generating photos on the fly.
It will be faster than searchingthrough the stock photos.
(15:07):
It will be exactly the image youneed.
It will be unique because it's aone-off.
It was created specifically foryou, and it's either free in
some tools or almost free onsome of the other tools.
So you can create any image forany need, whether it's creating
emails, whether it's PowerPointpresentations for sales or for
(15:28):
your next speaking gig, whetherit is for newsletters or blog
posts, social media, any imageyou need in any format, in any
style, in any resolution you cancreate with these tools in
seconds, and it's absolutelyamazing.
What tools can you use for that?
The one that I use most, becauseI find that right now it
provides the best results is midjourney.
(15:50):
It actually does cost money, andthe only way to use it right now
is through a discord server,which is not the most user
friendly thing to do, but it's,as far as I'm concerned right
now, the best in getting youexactly what you need in the
best resolution, in the mostaccuracy.
Next one on the list isprobably, stable diffusion from
stability AI, and the third oneis Dall-e from OpenAI.
(16:11):
The same company that developedChatGPT.
But again, there are multipleothers that you can use.
The last two that I mentionedare completely free.
So if you're using stablediffusion in its various forms
and shapes through differentwebsites, or if you're using
Dall-e, it's completely free touse.
That's for image generation, butit also knows how to analyze
(16:32):
images.
You can actually upload imagesto an AI platform and it can
tell you what is in the image orwhat's gonna happen next, or
whatever question you want toask it When it comes to image
processing.
Now some of the more advancedtools allow you to edit images
based on understanding what's inthe image, analyzing the
background and the ability tomake really crazy changes to
(16:54):
existing photos.
change the pose of an animal ora person that you actually took
a picture off.
So you have a static image andyou can make the person in the
image turn their head or a lionroar.
While it wasn't when you tookthe original picture, just by
dragging and dropping.
And because it understands whata person is and how that person
looks like, and it can guess howit's gonna look like when that
(17:15):
person is looking to the left orto the right and have the
lighting perfect and so on.
So you can manipulate images,you can also add things to
images without knowingPhotoshop.
So, Adobe, recently introducedAdobe Firefly.
That allows you to write whatyou want to add to an image,
circle the area where you wannaadd it, and it will add it in a
very realistic lighting, etcetera, that it would look as if
(17:36):
it's a part of the photo.
In the same way, you can removethings from images, like people
that stood in the way when youtook a picture of a famous
monument you were visiting, youcan take those people out of the
picture and it will fill out themissing pieces.
The last thing that it knows howto do it knows how to expand
images.
So if you have an image ofsomething and you want the AI to
(17:59):
make stuff around it, that wouldlook realistic as if it was a
part of the original imagetaken.
It knows how to do that as well.
So there's multiple ways tocreate images, analyze images,
and manipulate images.
All that can be used for anykind of business task that you
can imagine from ad creation tosocial post and everything I
mentioned before.
The next big domino that'sprobably gonna happen that's
(18:20):
gonna have a huge impact onbusinesses and the way we
produce different kinds ofcontent is video.
There's already initial teststhat are working that you can
play with that allows you towrite a prompt and create a
short video from it out of thinair without a camera, without
editing, without anything.
It's still not amazing, but it'smoving in that direction.
The other aspect of this isobviously deep fakes, is the
(18:43):
ability to manipulate a video tolook like somebody else, to
sound like somebody else, or.
To have a completely realisticview of a situation that doesn't
really happen.
That's obviously extremely scaryon one hand, but if you think
about the creative possibilitiesfor a business, they are
literally endless becausesomething that used to cost tens
(19:03):
of thousands or hundreds ofthousands of dollars and take
months.
Can now take minutes with onesingle person in a home
computer.
So the ability to create stuffthat is professional grade, just
based on the fact you're reallycreative, becomes really
incredible.
And at the reach of anybody thathas good creative skills.
(19:26):
now that you have a basicunderstanding of what is
artificial intelligence Andlarge language models and
differents and what they can doand how can it work in your
business?
How do you approach this?
How do you really get started?
What are the things you need todo?
A checklist as a business leaderor somebody who aspires to be in
business or just as somebody inthe organization that wants to
(19:47):
step up and saying, I'm gonnahelp my company understand how
to implement ai.
I'm gonna start with someconcepts before we dive into the
checklist.
And the first concept is thatwe're moving from a very long
era, hundreds of years of tryingto improve efficiencies of
processes to an era where we canjust get the outcome without
(20:11):
having to worry about theprocess.
And I'll give you a fewexamples.
Example, number one, customerservice, multiple tools and
capabilities and processes andbest practices.
Are being continuously updatedin order to have better customer
service.
How can we as answer faster?
How can the first answer be theright answer, et cetera, et
(20:33):
cetera.
All these kind of things.
Entire companies are builtaround building these tools that
will help other companiesprovide better customer service.
But the goal is not providingmore efficient customer service.
The goal is having happycustomers that can get a answer
in a very short amount of time,and it's still one of the
biggest frustrations for peopledealing with companies and the
(20:56):
bigger the companies they'redealing with, usually the worst
the customer service is.
But what if we can train an AIon all the knowledge that a
customer service person has toknow and also give you tools to
take all the actions that theyneed, upgrade them, downgrade
them, send them the bill, revoketheir rights, add rights to
(21:18):
them.
Anything that a customer serviceextremely capable agent can do,
give that to the ai.
Then you have the outcome.
You have happy customers thatget immediate results because
they don't have to wait.
And they don't have to talk tothree different people and they
don't have to re-explainthemselves, and they don't have
to go through a phone IVR orthrough 20 different forms.
(21:40):
They just go in, they type orspeak what they actually need,
and they will get an answer,most likely, the right answer
immediately.
So you're going and you'recircumventing the process.
You're not trying to make theprocess more efficient anymore,
you're just going straight tothe outcome.
Another example, you.
Want to rank higher on Google,so you had to do SEO keyword
(22:04):
research.
You had to understand what's thecompetitive landscape looks
like.
You had to start definingdifferent topics you need to do.
You had to then create outlinesthat will be aligned with those
topics that you need to cover.
You have to design the websiteaccordingly.
Then you have to actually writethe blog post, optimize them for
SEO, and then deploy them.
That's a very long process thattakes a lot of people with a lot
(22:26):
of expertise and takes a lot oftime.
And now you can just tell it.
Here's my competition.
Tell me what keywords they'reranking for.
I wanna rank higher than them.
Write as many articles as youneed and set them up in the
right places, in the rightformat, in the right website,
and you get it.
Now what does that mean longterm for SEO?
I don't have a clue.
I think what I just mentioned isan amazing short-term
(22:48):
opportunity that I think will goaway once a big amount of
company will start doing that.
In addition, what's happeningfor SEO is that because chatbots
like Bard and Bing and ChatGPTare going to probably replace
traditional search because ofthe same reason we used to try
to make search better andbetter.
So Google leading this chargeare the best in giving you the
(23:12):
best search results for whatyou're looking for.
But then you still have to gothrough multiple links and try
to figure out the bits andpieces of information from each
link that actually answers yourquestion versus.
The outcome is I just want theanswer, and if the chat bot can
give it to me, I don't need toclick any links.
So the whole concept of SEO willerode either quickly or slowly,
but in the next few years, itwill either be very different
(23:33):
than it is today or disappearedcompletely.
So the first concept is reallyunderstanding that we're moving
from.
Improving process efficienciesto moving to this is the
outcome.
Can I get the outcome withoutgoing through the process?
I'll touch on a few additionalconcepts once I go through the
checklist, but the first thingon checklist, and is the most
(23:54):
important thing is education.
you have to commit to continuouseducation on the topic of ai.
Why am I saying that?
I'm saying that because the AIworld is moving so fast, there's
really amazing news, amazing newtools, new capabilities that are
coming out daily and whateveryou learn right now may be
(24:16):
obsolete and Middle Ages like.
A month from now.
So do you have to do this everysingle day?
No.
But do you have to keep yourselfcontinuously educated on what's
happening in this field so youcan at least not be left far
behind?
Absolutely.
Yes.
So that's the number onerecommendation.
How do you do that?
find the right podcasts likeleveraging AI or any other
(24:37):
podcast.
There's multiple other podcastsout there that will educate you
on that.
Read books, read articles,follow the right people on
social media that share tips andtricks and strategies around ai.
Keep educating yourself on thistopic so you can stay, if not
ahead of the curve, at least onthe curve and not behind it.
The next step is you do not wantto take this path alone.
(25:01):
Why?
Because you have your expertiseand your capabilities, but you
also have a job to perform, andyou're limited with your
understanding, and you'relimited with your area of
expertise.
So the right thing to do withinevery business today is to start
a aI Innovation Committee.
And the people on the committeeneeds to be people who, a, are
quote unquote geeks and willactually enjoy doing this.
(25:21):
B people who come from differentdepartments because they will
represent different needs anddifferent mindsets and different
approaches while looking throughthe AI lens on different
processes within the company.
those people need.
Time on their calendar to do AIrelated things, and we'll talk
in a minute on what they shoulddo.
And they need access to theright resources, whether it's a
(25:43):
sandbox to play with, so theydon't mess up anything that is
currently company infrastructureor budget to have access to
tools they want to test.
So they need resources in orderto do their work correctly.
What does this committee needsto do?
First thing is they're in chargeof educating themselves and the
company all the time.
That means, Learning aboutdifferent tools.
(26:05):
It means experimenting withdifferent tools for different
use cases.
It means developing processeswhen they find a tool that could
actually be used within thecompany.
Developing the processes on howto implement the tool within the
business, and then training thepeople within the company on how
to use the tool.
They also become the go-to groupwhen anybody has an AI related
question and they need todevelop a few very specific
(26:29):
deliverables.
One deliverable is.
AI guidelines, what each personin the company should know that
they should do or shouldn't dowhen it comes to using ai.
Just recently, a lawyer arguinga case for an airline in a court
in the US referenced sixdifferent cases in order to
(26:50):
argue his case.
That do not exist.
So think about the embarrassmentto A, the lawyer, and B, the law
firm that he's representing whenhe presented cases that don't
exist.
And the reason they don't existis because he used ChatGPT in
order to do the research.
And maybe the biggest problemthat these large language models
have is they hallucinate, theymake stuff up, and the bad news
(27:14):
is you have no clue.
When they're factual and whenthey actually make stuff up,
meaning you have to fact checkthem in these kind of scenarios.
But he did not know that.
And hence he used ChatGPT to dothe research, went to court, and
I'm sure again, he was extremelyembarrassed when he found out
that these cases don't actuallyexist.
(27:34):
So creating guidelines for goodand bad is extremely important,
and that's something thecommittee has to define.
Part of the people in thecommittee have to come from the
company's leadership a because.
It will involve making decisionsand B, because that helps lead
by example for the people inleadership to show the rest of
(27:55):
the company that they arehands-on involved in the
process.
But the other reason to haveleadership people within that
process is because one of themost important things you have
to do right now is to reevaluateyour entire business strategy.
Why do you do that?
Because the world is changingextremely fast, faster than it
has ever changed before, and theneeds of clients and the
(28:19):
capability to deliver thoseneeds is changing very fast.
You have to go and after you'veeducated yourself and have a
decent understanding of whatpeople can do today and what
they might be able to do a yearfrom now or two years from now
is go and reevaluate who aregonna be your customers, what
will they be willing to pay forin a year or two?
(28:42):
And the flip side, what otheropportunities within the scales
and the expertise you have inyour company are now open?
That did not exist before.
So let's take two quickexamples.
One for each side.
And I'll take two very extremeexamples.
One example, there are multiplecompanies today who offer SEO
(29:02):
writing services.
They will go extinct veryquickly.
Why?
Because the entire thing thatthey do can now be done by a
machine faster and in manycases, probably better than
these people can write.
And yes, I know some peopledisagree with me and will say
that AI writing is not as goodas human writing.
I think that's not true.
(29:22):
I think it's how you use andleverage the AI tools in order
to write, in order to get theresults you want.
But it's definitely way moreefficient than letting humans
write articles for you rightnow, assuming you are fact
checking it in the end.
As I mentioned before, samething goes by the way for
various types of consulting andtraining.
if you are writing onlinecourses today and that's your
(29:43):
main income, Your income mightbe at serious risk because there
is no doubt in my mind that thebig platforms like Coursera will
develop an AI model that cancreate courses on the fly,
customized to the specific needsof each person.
So instead of going and browsingthrough existing courses, you
will say, I need a beginner'scourse on topic X.
(30:06):
It's for X amount of people.
That's their level ofexperience.
That's what they know right now.
The course has to be withvideos.
It needs tests within it.
And we have six hours to takethe course spread over six weeks
with one hour a week, and youwill create that course for you.
It might even send a preliminaryquestionnaire to each and every
one of the people, so it cancustomize it further to each
(30:27):
individual needs.
So if you're creating coursesonline, as your main source of
income that may go away withinthe next year or two.
The flip side of that, if youhave a lot of data, like if
you're a large consultingcompany, and you can take all
that data and train models withthat, you can now provide much
more efficient, faster, moreaccurate consulting services
(30:50):
than ever before.
You can go after lower tieredclients that couldn't pay you
before because now you canautomate a lot of the processes.
So there's.
Amazing opportunities as well asbig risks.
And that's true for almostanything unless you're cutting
grass or manufacturing somethingvery specific.
And even there, there's probablyefficiencies to be made through
(31:11):
supply chain management usingai, et cetera.
when looking at buildingstrategy, there's a few key
points that are different thantraditional strategy, and I'll
mention them right now.
The first thing is I see threedifferent paths for business
success in the AI era.
Path.
Number one is, as I mentioned,proprietary data.
If you own proprietary data, andthe more of it you own, the
(31:34):
better you can train models.
On that data and provide valuethat nobody else can provide, or
very few companies who haveaccess to this kind of data can
provide.
This way, you can stand out andcreate a blue ocean for
yourself.
That was not possible before.
The second option is exactly theopposite side of the scale is
(31:55):
you're in a highly competitivemarket and you are the best in
implementing these AI tools toget better efficiencies.
And hence you have bettermargins and can be more
competitive with your pricingand win through efficient
implementation and usage of AItools versus your competition
that may not be as good as that.
the third path for success ishuman relationships.
(32:16):
And that becomes amplified inthe B2B world because so many
companies will be able to doalmost everything really well
and vanilla like everybody else,because using those tools will
allow everybody to do thisrelatively easily and without
huge amount of resources.
The human connections are gonnabe a huge differentiator.
So that's another path that Ithink every company should have
(32:36):
done so far, but becomes waymore important moving into the
AI future.
the last concept that I wannamention when it comes to
thinking on the strategy ofusing AI before we dive into the
rest of the checklist is theunderstanding that the existing
concept of diminishing returnsmay not apply anymore to
(32:57):
different aspects of how yourbusiness runs.
An example I used earlier is,let's say you're writing
capability right now to createblog posts.
You can create three a month orthree a week, and now you can
create 300 in a day.
and I can give you multipleexamples like that, but the
ability to take something thatwas a significant bottleneck and
had diminishing returns in orderto scale, it can now scale
(33:20):
infinitely.
With almost zero marginal cost.
And that concept of looking forthings in your business, that AI
allows you to scale dramaticallywith zero or almost zero
marginal cost, leading you tolimitless growth on those
particular topics.
If these things are majorbottlenecks of your growth right
(33:41):
now, they can go away,immediately.
It's not a process.
It's a almost a zero to one gamethat happens almost overnight
once you figure out how to applythese tools to those specific
bottlenecks.
You have to have that thing inmind when you are analyzing your
current and your futurestrategy.
(34:02):
Now I want to go back to thechecklist.
We said continuous education,number one.
We said AI innovation committeeas number two, creating
guidelines as one of the thingsthe committee has to do,
educating the company.
You must continuously educateother people in the business.
You can do this through talkingabout these things in town Hall.
You can share this on Slackchannels.
(34:22):
You can define specific thingsas mandatory reading or
mandatory listening, and thenhave people come in smaller
groups in their departments, inwhatever teams they're in, and
share what they've learned fromthe things that was mandatory to
read or listen.
There are multiple ways on howto educate your business, but
you have to have multiplepeople, preferably everybody in
the business educated on AI andits capabilities and how it can
(34:44):
be leveraged in order to run thebusiness in a more efficient
way.
You need to define KPIs on ai,and yes, there's.
Trailing indicators such as wenow have more sales or our
margins are better, but thatwill take a very long time.
But what you can start with asKPIs is Lin leading indicators.
How many people are using AI inthe company right now?
(35:05):
How many processes we haveimplemented AI into?
How many tools have thecommittee evaluated even if we
haven't implemented them, and soon.
So these are great leadingindicators to see that you're
moving in the right directionfor implementing AI across your
company in a successful way.
Another important aspect isethical guidelines, because AI
allows you to make stuff up.
(35:27):
It can lead into doing thingsthat are beyond the red lines of
your company, yourself, yourcore values, et cetera.
And that line is different forevery individual and for every
different company.
But you have to define what'sacceptable and not acceptable to
do in your business usingartificial intelligence.
Next thing on the checklist Isencourage AI usage.
(35:50):
How do you do that?
share success stories.
This person from that departmenthas done this and was able to
achieve that result.
do this in company town hall.
Invest a few minutes in sharingthose kind of things.
Show that you are doing it as abusiness leader.
Again, lead by example.
People will see you doing this.
They will understand it'simportant to you, hence they're
gonna follow your footsteps.
(36:11):
Reward people that successfullyimplement AI capabilities and so
on.
There's multiple ways to dothat, but the usage of AI has to
be encouraged within.
The guidelines that were definedby the committee, so nobody does
anything stupid that actuallyhurts the business.
Next topic is look for lowhanging fruits for
implementation.
How do you look for low hangingfruits?
(36:31):
AI is really good at veryspecific things.
It's very good at doingrepetitive tasks.
It's very good at data analysis.
It's very good at dataprediction.
So taking the existing data andextrapolating from that, and
it's very good at generatingcontent at scale.
So any task that falls into anyof those buckets will be a
(36:52):
relevant thing that AI can helpyou make more efficient.
If you have tasks that fall intomore than one of these buckets.
Then the benefits you will getfrom implementing AI for those
tasks will be amplified by thefact it gets tick in several of
those different boxes.
So again, repetitive tasks, dataanalysis, data prediction, and
content creation.
So what you need to do, you needto look at all the tasks that
(37:15):
the company's doing, andhopefully you have some kind of
a task management tool likeTrello or Jira or Monday or
ClickUp or whatever other toolthat you're using, or even just
an Excel spreadsheet that haswhat tasks people are doing and.
See which of those buckets theyfall into, and that will help
you prioritize the task, whichwill get you the most amount of
(37:36):
benefit by leveraging differentAI tools to solve or help within
these tasks.
Once you have them mapped andyou have them prioritize, look
for the right tools, give themto the committee, have the
committee evaluate them, andonce you pick up and define the
right process, implement itacross the business and measure
the amount of time or efficiencythat you're actually getting as
an outcome of thatimplementation so you can learn
(37:58):
for the next cycle.
The last component has to dowith what I mentioned before.
How do you nurture humanrelationships within the
business and outside of thebusiness in order to create a
moat that is not technological?
Because technological moats willbe very hard to create in this
new era because it will beextremely easy to catch up to
almost anything you can develop.
(38:20):
So a quick summary of everythingI shared with you today.
The first thing is AI is here tostay.
It's gonna have dramatic impacton society and on businesses and
on individuals, and the best wayto come ahead out of this
revolution is to a continuouslyeducate yourself, experiment
with it.
Find the right ethical ways toimplement it in your personal
(38:42):
life and in your business, andeducate the people around you so
it can have more brainstormingapproach to this.
Revalue your business strategy.
Find fruits and keep runningfaster than your competition.
That's it for this episode and Ihope you found this helpful.
And this is obviously just anintroduction to AI in
businesses.
If you wanna learn more, justcheck out additional episodes of
(39:04):
leveraging AI where me and myguests dive into specific use
cases, specific strategicapproaches that can help you in
a much more practical way thanthis introduction.
I wanna thank you so much fortaking the time and listening to
this.
If you enjoy this episode.
If this was helpful to you.
Please share it with yourfriends or people that you think
(39:25):
can benefit from this.
please give us a five star andwrite a review on the platform
you're on.
That really helps us reach morepeople, which means you can help
more people understand AI andhelp them implement it in their
businesses.
Thank you so much and until nexttime, have an amazing week.
(39:54):
As I mentioned the beginning ofthis episode, this is at least
for now, gonna be the lastepisode of Business Growth
Accelerator, and if you wantmore of me and my approach to
business and how to grow it,just sign up for leveraging ai.
If you've been listening to theshow for a while, and I know
some of you have been aroundwith me from the beginning, so
three and a half years, I reallytruly, I.
(40:16):
Appreciate the time that you'veinvested in listening to the
show.
I know you have other things youcan do in that time and other
podcasts you can listen to.
And I do not take this forgranted and I'm promising you
I'm investing the same energyand the same intellect with a
lot of renewed energy intoleveraging ai So go check it out
(40:37):
and thank you so much
Popular Podcasts
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.