How To Use AI Like A 7-Figure Agency Owner

Рет қаралды 13,639

Күн бұрын

Пікірлер: 41

@nicksaraev 7 ай бұрын

My new step-by-step program to get you your first automation customer is now live! Results guaranteed. Apply fast: skool.com/makerschool/about. Price increases every 10 members 🙏😤

@kamikaze-ub9kz 8 ай бұрын

This man is a genius. Zero gatekeeping in this little corner of the internet 💜

@warezit 9 ай бұрын

🎯 Key Takeaways for quick navigation: 00:00 *🚀 Introduction to AI usage in business* - Introduction to how AI has transformed a content writing company, leveraging AI for significant business growth. - Highlighting the common misuse of AI and the potential for achieving better results with proper knowledge. 01:28 *🛠️ AI workflows and technical insights* - Overview of AI model systems and technical aspects for optimizing performance. - Discussion on the importance of AI technology in future job markets and business operations. 02:11 *📷 Image analysis for personalized communication* - Utilizing image analysis to personalize email campaigns, enhancing engagement and conversion rates. - Explanation of the process and tools used for scraping and analyzing image data to tailor email content. 11:21 *🔍 AI as a filter for decision-making* - Using AI to filter through data for relevance, particularly in sorting job proposals on platforms like Upwork. - Detailed breakdown of creating an AI-powered workflow for identifying high-value opportunities and crafting initial contact messages. 15:47 *🤖 Enhancing AI's Precision with Examples* - Demonstrating how to train AI models with user and assistant pairs for more accurate job filtering. - Discussing the importance of providing clear examples to AI for better understanding and output quality. 22:04 *📝 Content Generation and Proposal Crafting* - Exploring AI's role in generating detailed proposals and content from brief human inputs. - Highlighting AI's ability to expand small pieces of information into comprehensive documents or proposals. 27:50 *🌐 AI-Powered Web Scraping* - Utilizing AI for efficient web scraping and information extraction from complex HTML structures. - Emphasizing the cost-effectiveness and speed of AI in parsing and extracting specific data from web pages. 32:07 *🔄 Advanced AI Integration Techniques* - Discussing the potential for AI to revolutionize workflows through extraction and data parsing. - Emphasizing the cost-saving and revenue-generating potential of AI when properly leveraged. 32:36 *🧠 Technical Insights on AI Utilization* - Sharing expertise on AI's capabilities and limitations from years of practical experience. - Offering insights on optimizing AI model performance and the importance of understanding the technology's fundamentals. 41:04 *📝 Content Generation and Tone Adjustment* - Highlighting changes in AI model outputs over time and strategies to adjust the tone of AI-generated content. - Suggesting practical tips for improving the relevance and quality of AI-generated texts, including markdown usage and prompt optimization. Made with HARPA AI

@rodneyg2200 6 ай бұрын

Every video I consume provides breadcrumbs that if you actually take notice and pick them up are essentially golden nuggets that are multifaced in nature. 1. They educate in the simplest of terms even if you have minimal knowledge. 2. They provide a path for intentional application for your respective use cases. 3. They provide multiple paths for scalability depending on your purposes post-viewing. #noknowledgegatekeepinghere

@irishmanericlive7549 5 ай бұрын

I love the work flow dude

@PeterJohnBarnett Ай бұрын

Hey Nick, did you end up releasing the Notion doc for this video?

@stinna5758 9 ай бұрын

Awesome Nick thank you so much for these valuable videos. If you get the question more often, could you consider creating a video on how you structure the pricing for your deliveries? Or if it's a short answer could you please answer below :) How does one justify charging X amount for certain automations? Cheers!

@nicksaraev 9 ай бұрын

I liked this comment so much I recorded a whole video on it! Link: kzbin.info/www/bejne/hKS0iIqkqpypsKs&lc=UgzAL14lJzC-0VmRzj14AaABAg Thanks for the suggestion.

@stinna5758 9 ай бұрын

@@nicksaraev Checked it out already, helped a ton! Really appreciate it man, thanks for the video! :)

@karamjittech 9 ай бұрын

Awesome but Nick, openai module really consumes so much tokens for each operation, can you try with some other LLM module like huggingface etc. or any other option.

@nicksaraev 9 ай бұрын

Appreciate the support my bro Yes, you can absolutely do this with any other LLM provider (Anthropic, HuggingFace, Bloom AI, etc). But imo quality varies so much in practice it's usually not worth it. The applications in which I use AI deliver such disproportionate ROIs that OpenAI tokens are minimal-usually

@travisross8704 7 ай бұрын

I would like to learn from you how to leverage make automation for outreach. From research (fetching company-level data points) for personalization down to message crafting - all programmatically. This is key to actual building a business. Would love to see your process ❤

@handfacts 9 ай бұрын

For a topic idea, how would someone generate content like a list with open AI that has multiple pieces of data such as someone’s name and also their email address and then ultimately input it into a Google sheet. The issue with splitting it is that it doesn’t stay linkedwith the name and the email address I have attempted this

@nicksaraev 9 ай бұрын

I'd recommend sidestepping that entirely by forcing the OpenAI module to output in JSON. Something like: Generate a list of profiles using the following JSON format. {"name":"","email":"","location":""} Then give it 2 examples of what you're looking for, and pass the output into a Parse JSON module (which will let you grab the specific elements) in later modules. If you're stuck with an array for whatever reason, you can also just convert it into an object. There are many ways to do this but I'd just create one in JSON. What you'd do is add a "Set Multiple Variables" module. Then add a variable name like "profileObject". For the value, do something like {"name":"get(split(input;separator);index)", "email":"get(split(input;separator);index)"}, where separator is the character you're using to turn the string into an array, and index is the place in the array that your name or email address will consistently be. Afterwards you'd feed that into a "Parse JSON" module and now you'd have access to the variables in Make like usual. Hope this helps 🙏

@handfacts 9 ай бұрын

This is very helpful (as usual), thank you! Ok I have one more for you if you have any thoughts on it. I am trying to get chat GPT to access a template to use/reference. I made each template into a variable and in the OpenAI prompt am telling it to reference the variable as a template but it is still outputting whatever it wants and not using the templates. Any thoughts on how to "store" these template and force OpenAI to actually use it? You are truly the best! @@nicksaraev

@sinbun 8 ай бұрын

Serious value here! quick question -- I'm trying to figure out how to scrape multiple websites in one scenario so I can build out a competitor price sheet to observe. Is it possible to scrape multiple websites in one scenarior i.e. load a google sheet with URLs, scrape it daily, and then populat a spreadsheet with the data?

@ProperEng 9 ай бұрын

Thank you Nick! So awesome. Seriously the value you are giving in these videos is unreal. Can you talk about your process for delivery of the systems, account creation process, client off-boarding after delivery, etc?

@nicksaraev 9 ай бұрын

Absolutely-will defo record this soon.

@johnringo6155 8 ай бұрын

@@nicksaraev how can one get to your mentorship/course please ?

@livingworkingoutsidebox 6 ай бұрын

Very helpful. But I still feel like a 8th grader taking a college course 😊 I do remember using markdown text when I used to write Steam articles.

@raycarrasco5997 9 ай бұрын

Why don''t you go for more replicability by hard coding a seed & continually customising your examples until you fine tune the prompts until you get the response you're after?

@nicksaraev 9 ай бұрын

Simply put: it's more work! I imagine my approach yields 80% of that result for maybe 30% of the time, and I prioritize moving quickly. But there are certainly other ways to do it and they can deliver objectively superior results 🙏

@angeladipietro3428 8 ай бұрын

Hey nick love this video you really inspired me to code on make. Although where do you get your lead lists from? Did you make your own scraper or do you get them from a different website? Just having a hard time with that

@vaidasvaidas393 9 ай бұрын

Great video 👍🎉 thanks for sharing! Would be also interesting to see how to make automated Istagram posting with AI and how to use specific GPT Assistant to answer email based on predefined few document database🧐

@nicksaraev 9 ай бұрын

Glad I could help Vaidas 🙏 sure that suggestion sounds interesting. Re: answering emails based on documents, we use Missive in our agency. Video I recorded a few weeks ago if you're interested: kzbin.info/www/bejne/p6bGqYNjfJmGn8U Note: still some manual work-in our case doesn't make sense to automate completely, since most emails are inbound inquiries and we want to maximize conversion rate.