More of this! And more details, please. It’s great to know the use case, but more details on how they built it. This is invaluable information. Great video.
I recently built a RAG for our old project management software that uses GPT4 alongside vector based retrieval of posts for context, mixed in with elasticfull text search with a bit of a RRF and it's been a life saver because many old topics weren't properly categorized back in the day and a lot of related information is scattered across different topics/categories
🎯 Key Takeaways for quick navigation: 00:00 💼 *Introduction and the purpose of the video* - The video host, Greg, introduces the topic of the video: how businesses are actually using AI. - He explains that he sent out a tweet asking for real-life examples of AI use in businesses and received many responses. - He emphasizes that he's interested in hearing about AI tools or workflows that are bringing real value to businesses. 01:28 🛌 *Use of AI in Restorative Sleep by Bright* - Bright uses AI to analyze survey results and chat messages from their users, saving the team at least 20 hours per survey. - They also use AI to categorize and extract data from customer chat sessions, such as feature requests and bugs. 02:25 📑 *AI for Automated Template Generation at TASC* - TASC uses AI for automated template generation, increasing their output from 2-3 templates per week to over 100. - This not only improves customer experience but also increases organic traffic through SEO. 02:54 📧 *AI in Customer Support at TASC* - TASC uses AI to draft responses to customer support emails based on previous email answers, reducing response time from 3-30 minutes to 30 seconds. - The process involves a human in the loop to review the AI-drafted responses. 03:37 📊 *Work Recap Tool at Pulley* - Pulley uses a tool that searches for instances of employees mentioning their own work on Slack and recaps it for them. - This tool not only helps employees track their career progress but also saves them at least 1 hour of work per quarter per person. 04:03 📞 *AI in Sales at Zapier* - Zapier uses AI to extract information from sales call transcripts and transfer it to HubSpot, earning them an extra 100K of ARR per month. - The AI extracts details from the call summary and recommended next steps, which are then added back to the CRM. 05:00 🤖 *AI in Documentation Maintenance at Mendable* - Mendable uses a bot to create a Slack thread with user questions about their documentation, draft documentation, search GitHub for relevant files, and submit a new documentation PR. - This process saves them at least 4 hours per week of writing, revisiting, and maintaining their own documentation. 05:41 📧 *AI in User Onboarding at Zapier* - Zapier uses AI to send email recommendations to new users based on their title, role, and company, increasing their new user booking rate from 8% to 12%. 06:09 🗂️ *Knowledge Bot at Pulley* - Pulley uses a knowledge bot that answers questions with previously locked away tribal knowledge from Slack history. - The bot is used between 30 and 40 times a day by the team of 30 people, saving them an estimated 40 hours of leadership time. 06:53 📝 *Conclusion and Call to Action* - Greg concludes the video by expressing his interest in hearing more about AI tools or products that are making a real impact within companies. - He invites viewers to share their experiences and join his email list for updates. Made with HARPA AI
Greg, I am a corporate guy. I'm not focused on money; in my case, LLMs grant freedom by boosting my productivity. I now complete my tasks in 3 hours, whereas previously it took 8-10 hours. This leaves me with plenty of free time to invest in activities that I love. Freedom is worth more than money.
I work for a pharmaceutical company in the medical department. Usually, I use GPT-4 with RAGs for generating insights from various sources (medical publications, books, conferences,podcast, social media, market research data, sales data), creating interesting educational content for doctors (videos, audio, short texts), and self-education. In my case, full automation does not work well - I achieve better results when I closely cooperate with LLMs, providing curated data, prompts, and feedback.
@@micbab-vg2mu that's pretty cool. It seems that your work is less deterministic than mine. Yours require more creativity and I feel like LLM can really be good at stringing together ideas. My business is very deterministic and heavily regulated so simpler automation can work. I have been testing autogen to figure out what I can do to exponentially scale my work output. Love that there is this huge opportunity with AI and LLM in general!
@@DataIndependent for automation, we have a GUI that takes in the human input and automates through the existing excel files. I spend about 1/4 less time than manually doing it myself. Now, I am trying to use autogen to initiate certain tasks and langchain to query sql database (client, project, etc). I am hoping to uncover more nuggets so it can do more and more work for me.
@@DataIndependent I'm uncertain on the effectiveness of using AI do actual engineering task since a lot of them require multiple inputs to make decisions. I don't doubt that this will happen at some point but I'm not knowledgeable enough to get this to work. Maybe your insight would be very helpful. I think it would be very applicable to many service businesses!
Basically, for personal usage / super small business ( run by 2 person max ) scale, the AI can't be or not found to be useful? I also wonder about this question yesterday, I try to learn so much, but I didn't really know how to apply, and apply to what?? @_@. Especially, I'm only a freelancer. It makes me wonder so much...... Hope you can find more data that impact to a single person work ( that doesn't need to use huge sum of money for train or use other peoples AI ). Thanks for sharing anyway :)
Hey Greg, You said you got a Bunch of replies, may i have that data where you have saved these replies, because I wanna know how any people are using AI. I am Starting my AI Automation Agency, so I wanna know how businesses uses these AI so I can learn from It and Help Other Business who don't use AI.
More of this! And more details, please. It’s great to know the use case, but more details on how they built it. This is invaluable information. Great video.
nice that's fun - which one was your favorite? I might be able to get them on a video to see how they did it
@@DataIndependent The TASC Automated template generation, and both Pully examples, the knowledge bot and the work recap tool all piqued my interest.
I recently built a RAG for our old project management software that uses GPT4 alongside vector based retrieval of posts for context, mixed in with elasticfull text search with a bit of a RRF and it's been a life saver because many old topics weren't properly categorized back in the day and a lot of related information is scattered across different topics/categories
This seems like an intro video to something with more detail. Where's the elaboration of these use cases?
🎯 Key Takeaways for quick navigation:
00:00 💼 *Introduction and the purpose of the video*
- The video host, Greg, introduces the topic of the video: how businesses are actually using AI.
- He explains that he sent out a tweet asking for real-life examples of AI use in businesses and received many responses.
- He emphasizes that he's interested in hearing about AI tools or workflows that are bringing real value to businesses.
01:28 🛌 *Use of AI in Restorative Sleep by Bright*
- Bright uses AI to analyze survey results and chat messages from their users, saving the team at least 20 hours per survey.
- They also use AI to categorize and extract data from customer chat sessions, such as feature requests and bugs.
02:25 📑 *AI for Automated Template Generation at TASC*
- TASC uses AI for automated template generation, increasing their output from 2-3 templates per week to over 100.
- This not only improves customer experience but also increases organic traffic through SEO.
02:54 📧 *AI in Customer Support at TASC*
- TASC uses AI to draft responses to customer support emails based on previous email answers, reducing response time from 3-30 minutes to 30 seconds.
- The process involves a human in the loop to review the AI-drafted responses.
03:37 📊 *Work Recap Tool at Pulley*
- Pulley uses a tool that searches for instances of employees mentioning their own work on Slack and recaps it for them.
- This tool not only helps employees track their career progress but also saves them at least 1 hour of work per quarter per person.
04:03 📞 *AI in Sales at Zapier*
- Zapier uses AI to extract information from sales call transcripts and transfer it to HubSpot, earning them an extra 100K of ARR per month.
- The AI extracts details from the call summary and recommended next steps, which are then added back to the CRM.
05:00 🤖 *AI in Documentation Maintenance at Mendable*
- Mendable uses a bot to create a Slack thread with user questions about their documentation, draft documentation, search GitHub for relevant files, and submit a new documentation PR.
- This process saves them at least 4 hours per week of writing, revisiting, and maintaining their own documentation.
05:41 📧 *AI in User Onboarding at Zapier*
- Zapier uses AI to send email recommendations to new users based on their title, role, and company, increasing their new user booking rate from 8% to 12%.
06:09 🗂️ *Knowledge Bot at Pulley*
- Pulley uses a knowledge bot that answers questions with previously locked away tribal knowledge from Slack history.
- The bot is used between 30 and 40 times a day by the team of 30 people, saving them an estimated 40 hours of leadership time.
06:53 📝 *Conclusion and Call to Action*
- Greg concludes the video by expressing his interest in hearing more about AI tools or products that are making a real impact within companies.
- He invites viewers to share their experiences and join his email list for updates.
Made with HARPA AI
Quality Gregg. About to dive deeper down this rabbit hole!
Greg, I am a corporate guy. I'm not focused on money; in my case, LLMs grant freedom by boosting my productivity. I now complete my tasks in 3 hours, whereas previously it took 8-10 hours. This leaves me with plenty of free time to invest in activities that I love. Freedom is worth more than money.
Your freedom is closely tied to financial security, as you've highlighted by earning more per hour now.
In corpo - usually increase productivity is not equal increase $ - it is more politic and PR. @@titoine79
what type of tasks are they if you don't mind me asking?
I work for a pharmaceutical company in the medical department. Usually, I use GPT-4 with RAGs for generating insights from various sources (medical publications, books, conferences,podcast, social media, market research data, sales data), creating interesting educational content for doctors (videos, audio, short texts), and self-education. In my case, full automation does not work well - I achieve better results when I closely cooperate with LLMs, providing curated data, prompts, and feedback.
@@micbab-vg2mu that's pretty cool. It seems that your work is less deterministic than mine. Yours require more creativity and I feel like LLM can really be good at stringing together ideas. My business is very deterministic and heavily regulated so simpler automation can work. I have been testing autogen to figure out what I can do to exponentially scale my work output. Love that there is this huge opportunity with AI and LLM in general!
Automation and AI helped me reduce my work hour by half but grow revenue by 50 percent. That is around 3x. Im in service business.
That's awesome - what is the use case?
@@DataIndependent for automation, we have a GUI that takes in the human input and automates through the existing excel files. I spend about 1/4 less time than manually doing it myself. Now, I am trying to use autogen to initiate certain tasks and langchain to query sql database (client, project, etc). I am hoping to uncover more nuggets so it can do more and more work for me.
@@DataIndependent I'm uncertain on the effectiveness of using AI do actual engineering task since a lot of them require multiple inputs to make decisions. I don't doubt that this will happen at some point but I'm not knowledgeable enough to get this to work. Maybe your insight would be very helpful. I think it would be very applicable to many service businesses!
Basically, for personal usage / super small business ( run by 2 person max ) scale, the AI can't be or not found to be useful?
I also wonder about this question yesterday, I try to learn so much, but I didn't really know how to apply, and apply to what?? @_@. Especially, I'm only a freelancer. It makes me wonder so much...... Hope you can find more data that impact to a single person work ( that doesn't need to use huge sum of money for train or use other peoples AI ).
Thanks for sharing anyway :)
Insightful vid, thanks for sharing! Did you come across any who were using LLMs over their company workspaces e.g. a company Google Drive?
Like a chat bot for google drive?
Sort of - but trained/with knowledge of the work that you or your company do
hm, for retrieval, yes I've seen it, but not for fine tuning on what a company does. That's a ways away it seems like@@avishvj
Great video! Would love more like it.
Nice! Thanks Adam - what part was particularly cool? So I know what to double down on
Hey Greg, You said you got a Bunch of replies, may i have that data where you have saved these replies, because I wanna know how any people are using AI. I am Starting my AI Automation Agency, so I wanna know how businesses uses these AI so I can learn from It and Help Other Business who don't use AI.