The AI Roadmap Every Business Owner Needs to Win in 2025

whether you're a business that has barely started using g a or you've been on the fence or you have started but you want to make sure that you're on the right track this video is for you I'm going to walk you through how to create the perfect 2025 road map for how you can use gen of AI in your organization tailored to specifically your needs your budget and exactly what will take you to the next level with the thousands of new tools and Frameworks coming out every single month it's unbelievably daunting to know are you

stepping the wrong direction are you being Nimble enough and most importantly are you able to look three six months in the future and understand if what you're working on today will still have Roi in 6 12 months from now these are just a few of the many questions I'm going to help you answer in the step-by-step guide if we haven't met before my name is Mark and I run my own AI automation agency called prompt advisors for the past 2 years as of today's recording we've helped almost 500 businesses and entrepreneurs best leverage AI in

their workflows and we've over 1,200 Discovery calls to fully understand what the market is looking for in every single industry I personally have over 10 years in the data science and machine learning space so that gives me an extra advantage to better understand where we might be in 3 6 9 12 months based on the technology that's out today if you're ready to demystify this whole generative AI space stick with me to the end of the video and you won't regret it let's Dive Right In if you've been a longtime subscriber of the channel you

know that I love going through PowerPoints typically L where I kind of go through Slide by slide and break down my ideas in this video I'm going to do something a bit different and more interactive where instead of thinking of things through the lens of a mind map we're going to think of it through a mind City so we're going to go through different jurisdictions of this city to break down all the different things you need to consider when thinking about how to use gen in your business so let's zoom into our first neighborhood here

and go to the foundational constraints that you have to think about before you even worry about what AI tools what AI Frameworks Etc so the very first thing you want to consider are security considerations and this is usually the very first place you should start is just understanding where are you on that security Spectrum are you dealing with data that's hyper sensitive that is moderately sensitive or if it went out into the world it wouldn't be ideal but nothing would actually happen answering this question is actually very helpful because it helps shape what kinds of

solutions you can consider right off the bat so if we take the stream case if you have data that's hyper sensitive and so restricted that it can't even live on a global cloud provider like Microsoft Azure gcp or AWS then maybe you might want to consider a local large language model infrastructure which is a completely different skill set in a different scope but at least you know from the get-go that this is the area or the Box you have to play in if you're dealing with something that's moderately sensitive meaning you don't have person specific

information that you can't anonymize and you you can use it but ideally you don't want anything leaked from the way you do things day-to-day maybe using something like chat teams which is one tier of security above normal Chach might be sufficient if you need a little bit more a few more magnitudes of security you might want to go to Enterprise and I'm just taking Chach as an example this applies to all products whether it be Claude from anthropic or Gemini from Google if we move on to compliance requirements if we think about Hippa sock 2

ISO gpdr or all all kinds of different Frameworks which ones do you need to adhere to again if you know this you know that you have a checklist of items that certain tools have to match in order for you to even consider them so in a way this is sometimes a silver lining because if you know you can only use so to compliant tools then when you go filtering for what your infrastructure should be which language models you should use which providers you should use you have an easy way to filter what you even can

use the harder part is when you can use anything for everything and that's where you need to dial in by use case and By Priority now the magic BW word and the Very contentious b word of budget budget limitations it's important to understand how much you want to spend on generative AI in general there will be a cost whether you hire internal team whether you subcontract an agency such as mine whether you work and you have an operating budget per month of paying llm costs API costs Etc as well as tools how much you're willing

to spend not only on big tier projects but month to month what are you willing to deal with in terms of subscriptions and recurring fees now if you're starting at scratch or zero implementing from the get-go is actually super cheap and even if you racked up 10 different tools you'll only be paying $1 to $500 per month where the cost really starts to hurt and scale is understanding if you need something to be available to hundreds of thousands of employees or you need to run some large language modelbased operations on millions of rows where the

same thing is happening over and over again but slightly different tastes and nuances you need to understand what are your constraints in terms of budget because again that will help you engineer what is the path of leas resistance now this might seem biased but from my perspective I don't see a world in the next 3 to 5 years where the average business cannot use Zen of AI meaning I think it will be a prerequisite to survive SL Thrive versus just being an option to actually opt into that being said if you want to start small

the most important thing is that you start in general you can leverage Ai and get the benefits of AI not just using generative AI even though that's the whole scope of this video and the scope of the word AI pretty much in the world right now there's a whole other world that I came from which use the old way of AI which is the machine learning and statistics and that stuff although it takes quite a while to develop and test and produce properly once you actually build it it's actually almost free to use over and

over again so there's this whole area of AI that majority of the world is ignoring right now just because it's not as hype and fast moving but just know that even from a budgetary standpoint you can start small if you know exactly the sandbox you're playing in in terms of cost as well as in terms of the use cases you have in terms of how many individuals or stakeholders will be the end user of this Tooling in jbi so this slide completes our visit in the most boring neighborhood of the city and we can now

dive into slightly more engaging at least more thought-provoking which is identifying business business pain points and understanding where geni might fit best based on your priorities so if we zoom in just a tad and you go to identifying core constraints you have to make a core decision and this decision can be double-sided but typically it's one-sided do you want to use gen for Revenue generating stuff or cost-saving stuff if it's Revenue generating then that's easier to even justify with a board of directors your stakeholders or Finance because as long as you're making more revenue from

doing the thing than it costs you to do the thing you should be in the clear where it becomes a bit more tricky is when you spend thousands of dollars to save hundreds of dollars and that's where it doesn't make sense to actually use this technology at least using the most Cutting Edge latest model now there are times where you can get the benefits of Both Worlds where the generative AI activity you do both can make you more money think it's following up with leads or nurturing leads and it saves you time because someone would

have had to take time out of their day and use their salary time to go talk those leads themselves so there are those activities where you have the double leverage where should you pursue it and should it be successful you get benefits from both sides of the actual coin when we engage with businesses at prompt advisers the most common use cases or questions that arise are what should I do first I have all of these things all of these pain points all of these constraints what should I prioritize ahead of everything else in our typical

answer assuming that there's no crazy Nuance is that if you start with the lowest level task that actually takes a lot of time comma it also deres a lot of value and generates a lot of Revenue potentially those are your golden gems that you want to start on and the main idea is you just start from the highest leverage and the lowest input activity that can be easily replicated meaning there's not hyper Nuance from Scenario to scenario that's where you start moving up that actual funnel and getting down to maybe much harder tasks that you

don't maybe automate entirely but maybe you automate 80% and that 80% over 6 months over 2 years 5 years will really justify itself in terms of your initial investment and my last Point here is momentum so the name of the game in gener of AI is starting like I said before once you start racking up some small wins one you take away the skepticism of yourself or other stakeholders that this stuff is actually usable this stuff is production ready and that you can derive a lot of value from building automated workflows with G of AI

sprinkled in or endtoend G of AI processes this is why when an average organization reaches out and says hey here's our security requirements it's medium security here's our budget it's around 500 to 1,000 a month what should we start doing to dip our toes and understand how to leverage Ai and my response is typically to leverage something like chat GPT custom gpts and now you have something brand new called chat projects very similar to Cloud projects which which allow you to build pretty much a cheat code for building a full web application that has a

text interface you can connect tons of services to and you can start building tons of use cases for a fixed price so if you imagine you're paying $25 per person per month and each person can build theoretically tens if not hundreds of custom gbts that all do micro tasks that they can start automating you'll learn so much through that process you'll learn how to do prompt engineering you'll learn how to create a solid know based and what files are helpful and what files aren't you learn how to take knowledge based files and lean them down

so they're comprehensive enough and succinct enough that they're actually useful and you're not sending chat gbt on a wild goose chase to find an answer that should be way easier to find and the cool thing is once you have a few custom gbts or Cloud projects whichever platform you choose that are heavily used and are instrumental parts of your workflow and you want to remove the human out of that Loop that's where you have an amazing use case and a prototype to justify investment to say this process end to end is worthy of investing and

putting our dollars in to make sure that we have this process fully RN out in a way it's kind of like putting training wheels on and the training wheels to me is how can you remove full stack development meaning creating some form of app setting up a server setting up a domain setting up a database how can you take all of that stuff out and just focus on trying to use the AI trying to learn prompt engineering deciding who your organization or who you're going to subcontract is going to take care of those tasks once

you have that confidence and justification that things are working that you can save time and money with this stuff that's where you can start to venture out and try more experimental use cases and really start to ask the question what is and tying into that how do you understand what constraints have the biggest payoffs typically you want to focus on two to three main use cases across the whole organization so let's say you were a standard organization that had the sales the marketing the legal Finance Etc ideally you want to look for two to three

very human intensive tasks comma that are very high value and very high leverage meaning if those things were to not be done you would stop making money or making as much money and one helpful rule of thumb is looking at the actual workflow itself so let's put aside the fact that it's high leverage and it's human intensive if you're dealing with something on pen and paper or you're deal dealing with some form of special CRM designed for only your company that has no backend access to the internet or the world in general that's where it

becomes really difficult to execute a gen project so if you had a perfect triangle of the best case scenario is you have a very high leverage task you have it's very valuable when it's done and the last one is there is a way to actually speak to the systems needed to complete those tasks so if we take a very simplistic one like social media automation every social social media right now Facebook Instagram Etc they all have a backend or an API you can speak to they all need copy they all need written copy that's tailored

to those platforms you have probably a social team that spends tons of time crafting your short form your long form a bit of both or either over making solid social media content could be a very high leverage task it is very human intensive to do properly and once you really automate it you get a ton of Revenue possibly if your messaging is good and assuming you're using gen to its full SP potential for the social media use case you have the perfect triangle which is you have a very high leverage activity it's very human intensive

typically and if you're able to automate it you have all the necessary systems in place to just hook them all together think of something like zapier or make or n8n and you'll have an end to-end process where basically you can start with an input that you obsess over how do you standardize the input that's usually the thing that stumps most companies is what is the trigger for everything to start is it some form of email we upload so it could be a slack message that you send where you say hey can you create a post

about the latest and greatest AI news that just came out and that one very vague input can become a very sophistic output if you have the necessary stuff orchestrate and the last component here is defining a specific mindset for doing generative AI unlike normal AI or the old AI or all kinds of other projects there is no clear beginning and end technically to a lot of these initiatives just because you built an amazing resource system where your employees can now talk to your data that's stored in Azure or stored in AWS and they have some

chat interface using XYZ model doesn't mean that in 3 months from now we get an amazing model from openi that's so good at leveraging those same documents and doing way more detailed analysis and you might then have to consider switching so when you pursue geni projects I like to think of it as it's done for now or this iteration has been completed new that in 3 to 5 months you might want to reassess and ask the question is this still relevant are we paying way too much compared to where the costs are now are we

using the right infrastructure are we using the right models so this entire process while it's more daunting than the typical project where you can do a system migration and then sit on it for 5 years you don't have the luxury of sitting for 5 years with one type of gender VI solution built 6 months ago you want to retroactively and be proactive and see seeing how can you make it better and how can you do more with less and one last point I want to make on this slide is that I've never seen a geni

project whether it's voice whether it's in the education space whether it's a custom development project using make or zapier where we have requirement and we get the perfect result that is perfectly according to our stakeholders needs on the first try a lot of times you don't even know what success looks like until you see the first draft because you can have that perfect workflow where you have a voice note you send on slack and that voice note then gets turned into text then that text gets turned into social media posts and after all that's done

you might look at the post and be like this does not have our brand voice this does not have the Nuance of human speech or personality we're going for so now a process that was just trying to do a toz now is accounting for different letters outside the alphabet that weren't in scope because once you see it you understand where you want to go and what's missing that final mile to take it from okay to great now naturally the more we have up front the more you understand what you're trying to accomplish how you're trying

to accomplish it what your budget is what your security constraints are and you m those all together the more the likelihood is we can maybe finish it by iteration number two iteration number three and making sure that we can actually hit the goals we set at the beginning accounting for all the possible paths we might need to Pivot down the line and and this will Mark the end of our visit in this part of the city and now we'll go to a more new and hip part of the city where we're going to go through

a framework that we use a lot at prompt advisors that's derived from an existing framework on the market so we like to call this the bite scoring framework so this is our bite-sized approach to finding a way to best prioritize the projects even directionally to understand what should come first what is truly highest leverage versus what is this would be cool if we could automate this thing so we actually derived this bite framework from an existing one called rice where it stands for reach impact confidence and effort and in our case we just tried to

make the math a bit easier and a lot simpler so it's very quick to do this list on the go so pretty much it stands for business criticality so how important is this thing to our business to actually experiment with and potentially fail on doing assuming that gen ofi can't do the thing after we try the next one is impact potential somewhat related if we were to fix said problem How would this solve things for us how much time would it save how much cost would it save how much revenue would it create would they

all be happening in tandem the next one is time to Value so if we start this project even if it's super business critical even if the impact is super high if it takes 3 to 6 months and it's your first AI project you're doing there might be this sense of not necessarily just skepticism but worry what if we do this thing and it messes up and we wasted 6 months of resources and cost to take a bet on something that didn't end up being fully feasible so this is why again if you're starting out or

you've already started and you want to kick off 2025 with more momentum to get into the harder and medeor projects having that shorter activity or that shorter project where you can get that win or at least some direction on whether that win is accomplished in the first month month and a half two months that will set you up more for Success than praying that certain technology comes out or certain costs go down within the timeline that you're looking for the last one is obviously ease of implementation if it's something like I said before where you

have this random system from the 2000s that was built by developers that are no longer there and there's no way to connect that thing to everything else but that's your source of Truth probably not the best candidate for now to start with there are different things that came out that can probably help with this in 3 6 months but as of today I probably wouldn't start with that specific project and for Simplicity well use scoring from 1 to five five being super high and one being low so if we go into an example here just

to break it down an example scenario of a small consulting firm that's overwhelmed by client inquiries and let's say the solution they were thinking about is a basic question and answer chat bot and this is a very vanilla use case so if we look at the table here we have business criticality B at four and that's primarily because if you're overwhelmed with all these inquiries it's taking time from something else that's probably way more more High leverage for the long-term survival business we looked at impact very similar probably four out of five because faster responses

happier customers less calls less inquiries if we go to time to Value this one is like a five out of five and the reason why it's a five is because if you go and search for chatbot Solutions and I'll go through all kinds of different tools that are out there right now in general in a few slides but if you look at how many are out there and how easy they are to use for basic use cases not convoluted ones not very convoluted booking flows and nurture flows you can find ones like chatbase Dante AI

upload PDFs have it answer basic questions and then you're good to go and that also leads into the ease of actually implementing it very easy to implement it's made for non-technical folks in this case to set up so for this specific use case this is hypers solvable and probably one of those first quick wins that I'd prioritize if that was a line item on the road map now for more complex scenarios and more complex use cases what I've done is I've put together a huge uh cheat sheet guide on everything that I'm going to be

going through in this video just more in depth where you'll find even some scenarios going through some model examples of how you might do some scoring so these are some hypothetical individuals running hypothetical companies where I go through what the possible context is what the appropriate bite scores would be for every single metric and then why my rationale would point to that as well as showing you the actual table so I have three examples in the cheat sheet and this will be available in the gumroad link in description below so you can check that out

after this video If we go back and keep rolling if we go to the next slide here it's about planning is more important than building so it might seem like I'm contradicting myself do something fast get a quick win at the same time though it is way more important to plan that quick win plan what happens after that quick win because you just don't want to create a bunch of as Alex horoi would say half-built Bridges where you start some initiative you kind of fix it you start another initiative you kind of build something for

it but nothing actually gets pushed through till the end where it's fully useful it's an amazing justification that at the end of 2025 you can look back on and say wow this was a very well-invested well worth project that we pursued so like I said before you want to look at all the Departments and even if you're a department of one and you are everything which I used to be not too long ago you want to take a look at all of your activities in the different segments and see which are the ones that are

the most suffocating to the business moving to the next level and the goal is not to overthink on the individual use cases you shouldn't spend a year planning uh should I bet on custom gbts or Cloud projects or look Google Now is really stepping up their game should we look at Google as a provider just choose one of them because the learnings from one will be transferable to the other especially if you plan some form of system migrations down the line the most important thing is that you have Tribal know know of what worked and

how you approached it and what your internal Sops are and you'll never have those standard operating procedures if you don't start so the most important thing is start where you want to overthink is what happens next so if we take that example of that chatbot so this is a basic chat bot and now this fixes all the easy to answer questions is it valuable enough to build a more bespoke chatbot that's custom we have to get away from the easy providers and we get custom code we build on something like voice or bot press and

now we want to integrate things like connection to our CRM we want to be able to do appointment setting we want to be able to even collect payment potentially for appointments for a doctor a pediatrician a plastic surgeon these are the follow-through questions that are worth overthinking about the last thing here is that the first few initiatives you might start might take a while because the technolog is new you're setting everything up if we take the example of a workflow automation tool let's say make.com even if you're not familiar just bear with me so it's

a platform where you have a bunch of bubbles that represent different automations you can execute at the beginning you have to basically log into every platform set up your API keys go to open or cloud or Gemini and pay to put a credit card get a key and start to set everything up all of this stuff takes longer comma at the beginning but as you get more and more well-versed in having everything set up for you and now you can just focus on building you'll find that projects 3 to 5 5 to 7 and 7

plus become easier to start maybe harder to actually execute and with this we'll take a little fery to a little island off of our Mini City here and look at actual use cases in passing by Department none of this will be groundbreaking for you just more of a reminder of the different types of activities you might want to consider if you want to integrate gen AI in 2025 so if we go into sales and marketing the most common use cases that we get and again uh we have a presence on Fiverr where we are a

top rated agency a Fiverr Pro so every day we get anywhere from 8 to 10 inquiries for any domain in every industry so the most common ones in sales and marketing are leite qualification automated Outreach ad content generation SEO and a brand new thing that's on the block that's called Geo it's generative engine optimization and this has come about because of things like perplexity and more recently uh chat gpt's search GPT where now you can basically say browse the web and it gets you answers to questions in real time Geo will be the new SEO

in the coming years and it'll be that much more important that you're ranked in a way or you position your blogs or website in a way to take advantage of this new opportunity that eventually people will be using voice right using Siri using chat T voice to ask for who's the best plumber who's the best roofer so taking advantage and being at least cognizant of this new thing on the horizon will help you be that much better position moving forward next one if you go to Ops and back office so typically you have the standard

use cases of we have set documents we want to be able to convert said documents in formats that our employees can quote unquote talk to Via our knowledge base and there's a million ways you can actually solve this you can use things like pine cone uh things that are natively on AWS there's even an agent infrastructure on AWS there is cognitive search on Azure there's all these different options it's more so this are the most common use cases next thing is customer support now you have voice which is really showing up in a big way

after 2023 ended and that's why we specialize a lot in voice Solutions and building not only inbound agents but also outbound agents as well as voic driven infrastructures we are betting that the year of 2025 is one of way more advanced agents but more specifically where voice becomes an expectation versus a luxury so just thinking about that from a high level customer support moving forward should have a voice option a chat option that are AI automated and ideally there are still humans in the loop because there are cases where you do need that special touch

and that Nuance so it's not good to think of automating 100% it's you want that Trifecta of human for high value questions and high value scenarios and you have voice and you have chat now there's this little video that I always think of from the office if you're familiar it's with like a little show it's very dark humor kind of funny um um that basically there was a product being developed called wolf where you could send an email a text a Facebook message to a contact and that was revolutionary back then so this is a

little video montage of kind of what is possible today with AI Facebook what's up Facebook I sent you a Facebook message yesterday I still haven't heard anything back you should have sent me a wolf a what you send a wolf it goes to your pH phone cell phone email Facebook Twitter and on screen all set up Ro what. all right a bit of comedic humor but overall the funny thing is this was fantasy and now it's totally possible and this is also an amazing thing to think of when you think of uh lead nurturing lead

drip lead generation Etc now if we go into next portion is legal and Consulting so this one common use cases with attorneys we work with are client intake automations contract summaries drafting assistance and answering mid to level difficulty questions from the legal side so these ones are pretty out of the box the hardest part with these is just making sure that you have the documents standardized in a certain way that are easy to speak to especially if you're dealing with scans or statutes that are really dug into certain web pages you want to be able

to take care of that pre-processing which is not really a generative AI thing it is more so a data engineering thing all right and the next one is Education and Training where pretty much the most important two elements here is one the more you upscale your employees to be able to use G of a versus just having one specialist the more everyone will be able to accomplish and in a very difficult economy which we're seeing right now you'll be able to do a lot more without shaving off any staff you can just make those staff

super human all of them leveraging AI for those different skill sets That's essential and I think it's almost a right in the future to be literate in all these Technologies so if you are Forward Thinking and you start now you can make those employees one feel like they're actually growing which is awesome and two as a business owner or a senior executive in a business you can see a lot more High leverage opportunities you can take advantage of as a business owner you'll now have a staff that can do all kinds of things well outside

their original Comfort zones and skill sets so that's a very high leverage thing to do and one more thing here is designating some form of Team tinkerer where at least one person should be keeping a pulse or a tab on what's out there because d by day literally the day of this recording you have the 11th day of open AI 12 Days of Christmas and there's brand new things coming out literally by the hour by the minute so in an environment where brand new things come out that could change the entire trajectory of what you

do day-to- day it's important and helpful to have someone that could actually take on that role and the last one here is more of a fringe case where you have media and entertainment and this is where you can do image and video generation as of this recording we just got Sora last week we got ve from Google both very potent models with text to video you have very Advanced image models and you have all these different things that can allow you to provide a better customer experience or at least a more experential one where you

can experiment with things that might seem gimmicky but they might be novel Now While most businesses are not doing it or not adopting it you could be using it in your ad copy your Facebook ad copy instead of worrying about hiring an animator or hiring a designer you can start to really experiment and see where do these different things fit that most people even myself even our agency hasn't really figured out where is the perfect use case for AI generated images what are use cases we've never even thought of before so with that we're going

to go to our file neighborhood and in my opinion the most important neighborhood which is future proofing and continuous optimization what's out there how can you take advantage of it and how can you make sure that 2025 is not a year that passes by and you feel like you wasted time by not starting or over indexing on the wrong direction sorry to interrupt the video I just have something super special for business owners that's coming out in 2025 which is our early AI adopters Community which is going to be a paid apply only Community where

only vetted business owners are going to be selected to create our ecosystem where we're going to offer exclusive content have a group of coaches in the automation space in the LM space and the voice space to help you take advantage of everything that generative a has to offer in a very bite-sized and tailored way that has a community feel it will have events and master classes where myself I'll do twice a month uh one hour to go through all the questions to address as many as possible you'll have tons of help resources and exclusive automations

along the way so if that seems interesting to you feel free to go on early adopters doc Community the link will be in the description below and you'll be able to apply for Early Access on the site and once you do that we're going to do some consideration of applications at the beginning of the year and then email those that are selected and we're going to be handpicking five specific business owners or Business Leaders to actually be our Alpha testers where they'll get the first month for free in the community in exchange for their candid

feedback about what they're looking for what are their most common pain points and how we can help address those concerns and make sure that we pave the way for the resources needed the knowledge needed and the expertise needed make sure you can navigate to those very tricky problems if this sounds interesting to you feel free to go on the website it'll be in the link down description below apply and I'll hopefully see you early in 2025 all right back to the video all right so the major thing to be cognizant of is there's pretty much

a tool for almost everything now now there is no mega tool that does every single thing but there are hundreds if not thousands of disperate tools that can all accomplish these microtasks if you go on a website like there's an AI for that or if you go on future pedia you're going to find literally 26,000 AIS again the majority of them are not useful but the idea is there are all kinds of tools out there to help you accomplish exactly what you're looking for for so with that in mind I'm going to give a quick

overview of the different domains or modalities of gen of AI that exist and what you can consider and think of in terms of the major players out there so when it comes to text based agents again like I said before if you want very lowlevel chat bot you can set up immediately you have things like chat base ninja chat Dante Ai and if you want to start making them more custom and bespoke that's where you want to start considering things like voice flow bot press and if you want to build your own rag system which

stands for retrieval augmented generation you can use things like pine cone in combination with make.com n8n to set up very long workflows and there's hundreds of tutorials on how to do this on YouTube on how you can securely speak to your data in a very affordable way and of course I would be remiss to not talk about open AI you have the assistance API from open aai that lets you build lowlevel agents and integrate them in things like mini chat in voice flow in bot press and take advantage of this a opportunity from that standpoint

of course you'll find that the majority of AI tools and now tools that you can start building yourself are pretty much wrapped on top of the open AI infrastructure using their models which are continually getting better and more importantly are continually getting cheaper now if we move to our bread and butter prompt advisors which is voice-based Solutions you have all kinds of providers and I imagine voice is going to become more and more commoditized at the high level in 2025 so you have vendors like Bland AI which are very useful for very convoluted cases that

have a conversational pathway necessity meaning it's a you have like five different ways where one conversation can go down it can depending on the situation that arises can go in six different directions that's where it's super useful VAP Is Awesome from going from zero to voice agent in literally 3 to 5 days even with no technical expertise you can watch one to two YouTube tutorials write a prompt and you can test out a voice agent almost immediately then you have things like retail which are really good from a testing standpoint where you can build these

tree nodes similar to Bland but more simple and you can actually test how that agent is going to actually function without constantly calling it you can just write in a chat and it will emulate what it would have said on the voice so retail is another great option and if you don't want to even see or touch code and you want minimal PRT engineering things like thly and sylow are more out of the box if you have more cooled calling use cases using something like sylow is probably better for that but one thing to keep

in mind is that you have things like the realtime API from open aai which is what powers the chat gbt voice that has literally in the last 24 hours of recording this video reduced their price two to three times the original one thing you're going to find very similar to llm cost in general everything is going to Trend towards zero over time the most intelligent models we have now will be almost free 1 to two years from now so with that in mind you shouldn't let let cost be the delimiter of which platform you choose

you should bet on the platform that is most likely to one quickly adapt with new infrastructures quickly adopt new Frameworks so for example vapy was able to take in the realtime API as a part of their product and let users actually use the realtime API which if you haven't used it is actually very code intensive to set up you have to set it up in something like repet hook it up to twio a phone provider it's very convoluted you want to be able to look at platforms that are very quick and have their pulse on

The Voice space so you can just rely on them to integrate it on their infrastructure and you just worry about nimbly switching settings and features on the go as things inevitably get more and more advanced and even on the most common objection we get when it comes to voice projects which is ah it's not fully human or it's slightly slower than I expected in terms of responding I wish it was instant but perfect that also will resolve itself most likely in the next 3 to 5 months we literally this morning I woke up and 11

Labs which used to just do voice cloning now has jumped into the voice space and they've released an agent that has 75 millisecond latency meaning it's literally almost instant it sounds pretty human and it's pretty sophisticated so imagine you extrapolate from here and the last thing to Keep On Your Horizon are things like high go high levels AI employee where now you have your CRM you have your contact instead of building something let's say on vapy and then hooking that up to your CRM you can have a voice agent now out of the box Ino

high level now mind you it's not as good as these other providers but again as time progresses you're now going to start having super crms that have chat out of the box that have voice out of the box where integration now becomes the moat because just having the CRM itself will not be a Moote and you'll see why on one of my next slides having theof software software in general will not be a moe anymore what you can do with that software and how quickly you can do it will become that Moe this one I'll

Zoom through pretty quickly automation based software make zapier n8n any one of these should be able to help you start automating different tasks especially if you're using a lot of brand name software crms email providers they should integrate into everything zapier has an advantage that it has way more native Integrations than make.com make.com has in my opinion the advantage of being a lot more user friendly and a lot easier to work with in general but it has less Integrations but it's still you pay per cost meaning in zapier or make you pay per operations so

every single time one of these little modules runs you're technically paying what's called an operation or in zap here it's called zaps whereas with n8n you can host it on your own server so security wise you can make more secure and you don't actually pay when an operation runs so theoretically you could have 10 different nodes or 50 different different nodes and if you run it 100 times you're not paying per that run cost it really comes down to what your preference is on the tool you like again for me I'm partially biased to liking

me.com versus the two others but you might try and say I love N8 end which tons of people on YouTube send to actually really like so check all them out and see what fits best for you both from a cost standpoint as well as an operation standpoint an image in vid based probably by this time next year there'll be a 100 different vendors you have mid Journey for image generation you have Runway Sora pabs and Google VO2 which as of right now is the biggest and hottest thing and the best performing video model probably because

they trade on YouTube um all of these are very Advanced compared to where they were two years ago or one year ago where you could see the meme of a AI Will Smith eating spaghetti that is no more you have much higher resolution and much higher quality outputs and finally we get to my favorite here which is the wild cards there are tons of softwares a few have gone over on this channel one being repet repet agent as well as Bolt um and I'll be making a few more videos on some other tools as well

coming up but these help you go from idea to building a fully functioning app so when I was saying before on the crms that software will not be aote anymore I think that in 1 to 3 years the average company will be able to actually create their own CRM should they wish to create their own softwares that they were typically paying assassin V4 and the entire software as a service Market will start to really change and flip on its head when you can own the technology and you can build it just using text or maybe

your voice in the future that changes a lot of business models so these are the wild cards that I strongly encourage you at least Tinker with or at least put on your horizon because I can imagine in 6 months it won't be a very painful process like it is now sometimes to create an app where you have to go back and forth prompt in a certain way spoon feed in a certain way that takes a bit of a skill set you will be able to go from instruction vague instruction at that to a fully functioning

app that you can deploy on a website almost instantaneously and have yourself or other people start using it so these are really important to keep in mind especially if you're starting to think about should I overinvestment is it really worth the risk to build that SAS where someone else can build that same thing for $20 subscriptions that's one question you have to assess so where do we go from here after everything I've gone through so pretty much if you look at the pros and cons I won't read it step by step the most important thing

for future proving yourself is understanding to not look at as they say in Canada where the puck is in hockey but where the puck is going to be in the same vein you can see what's out there today the most important thing is to see what's out there so I'll give you an example anthropic came up with something called computer use where it can take control of your computer and click around and do some things right now it's super inefficient because it takes screenshots of your actual screen sends each screenshot as an input and then

changes the position of the cursor accordingly super inefficient 3 months from now that will become real time stream we're now streaming your lapped up in real time so that cursor movement will be a lot more meaningful and maybe they'll actually make it have a much more complex model that is fine-tune on navigating tabs computers logins Etc so this will change all kinds of processes where let's say you did have that use case where you had that CRM that didn't talk to other systems maybe now that's not a problem you don't need an API you don't

need an endpoint because you can programmatically deal with it from a front end where the AI is taking control of your computer or col these computers and accomplishing the task so if we were to make some bets right I think that by this time next year voice platforms the cost to run per minute will be close to zero I think that they'll be very humanlike I think that latency will be very much mitigated if not instantaneous I think that text to video will be exceptional where it's very comparable to what you would get from a

video editor I think that when it comes to text based Solutions you won't have as formulaic and robotic sounding llms you'll have something that's more bespoke that's more tailored and adapts very quickly to feedback in changing of style and if you can take away one major thing from this video is that gen of AI is a process it is an infinite game it's not a finite game so you don't win at gen of AI you invest in gen of AI over time and it never really ends your investment starts but it will change over time

you'll have to Pivot in different directions but as you're able to be nimble as you pivot in those directions you'll be able to realize a lot of business benefit and you'll realize what is the highest value thing for your yourself as a business owner as a business leader as a stakeholder as an employee what's the most valuable thing to your users your customers and when you bring all that together every one of these avatars can be addressed by some form of generative AI so if you take that in the rearview mirror and you know that

over time cost of everything will Trend towards zero knowing that this is something you have to focus on moving forward is very important designating a team even a single person to start if you're a small organization is very important to make sure that you're ready and Nimble to take advantage of this huge opportunity that has this small window of Arbitrage where enough businesses and people have not used or harness the power of geni but if you can the next 1 to two years you do have an unfair advantage over many now with that like I

said at the beginning of the video I'm going to provide you this entire cheat sheet as well as a bite buddy custom gbt that will help you create a bite framework and better understand and prioritize your tasks and projects just by either speaking to GPT typing GPT that'll be super helpful for you and absolutely Shameless plug if your organization would rather Outsource the strategizing the identification of these projects how you can approach these projects doing the bite scoring for you we offer a special package called The White Glove AI strategy audit that will basically have

us sit down with your team for 1 to two hours go through all your processes we'll send you over a checklist of things to prepare so we can make full advantage of the time we spend together we'll generate a full road map document on how you could use G VII in all the different spheres of your business which ones seem to be the highest leverage opportunities and what are the costs associated as well as the expected ETA or expected time to complete said projects so you have a full blueprint of exactly where you can head

as an organization moving forward with our team's guidance to help you better understand everything that might pop up along the way to save you time effort and most importantly money if that sounds interesting to you just go on book your white glove AI strategy audit today book yourself in and we'll be super excited to help you set off your AI Journey for 2025 if you enjoyed this video if you found it valuable and helpful if you've been a longtime subscriber of the channel and appreciate this slightly higher production value than I usually do let me

know in the comments below it'll tell me that I'll probably have to save up to do this more often and equally as important if we haven't met before let me know in the comments down below whether or not this was helpful and what other areas of genbi you might want want me to dive into to make sure that 2025 is the year where you derive the most Potential from LMS and gen of AI in general leave a like sub the channel and I'll see you all next time