재피어의 재파톤: 자동화 효율성을 위한 AI 워크플로

Zapier의 Zapathon은 작업을 자동화하여 사용자의 생산성과 효율성을 향상시키는 혁신적인 AI 워크플로우를 선보였습니다.

### AI로 업무 자동화: Zapier의 자파톤 하이라이트
* 🤖 Zapier는 최근 Zapathon을 개최하여 리소스에 액세스하기 위해 이메일에 가입할 필요 없이 **AI를 활용하여 업무 프로세스를 간소화하는 혁신적인 워크플로우**를 선보였습니다.
* 📊 뛰어난 워크플로 중 하나는 **Asana에서 프로젝트 관리를 자동화**하는 것으로, AI를 활용하여 상세한 프로젝트 개요를 생성하고 프로젝트 관리자와 후원자가 쉽게 액세스할 수 있도록 관련 정보를 **Slack 메시지로 컴파일**합니다.
* 📝 이 웨비나에서는 **블로그 게시물에 대한 SEO 친화적인 요약**을 생성하는 메타 설명 생성기와 콘텐츠 작성자가 잠재고객의 참여를 효과적으로 유도하는 **소개 생성기**와 같은 맞춤형 AI 챗봇도 소개했습니다.
* 💬 또한 하나의 블로그 URL에서 여러 개의 게시물을 생성할 수 있는 소셜 미디어 콘텐츠 봇이 강조되어 **여러 플랫폼에서 콘텐츠 제작의 효율성**을 높일 수 있는 가능성을 보여주었습니다.
* 📚 작업 관리의 경우 **Zapier와 Zendesk의 통합**을 통해 지원 티켓을 자동으로 분류하여 엔지니어가 다양한 기능에 소비하는 시간에 대한 인사이트를 제공합니다.
* 💡 프레젠테이션에서는 **매일 스페인어 문구 생성기, 더 나은 정리를 위해 Slack 메시지를 실행 가능한 작업으로 요약하는 시스템 등 AI를 통한 개인 생산성 향상**에 대해 강조했습니다.
* 💵 주목할 만한 통합 기능은 **신용카드 알림을 분석하고 비용을 분류**하여 지출을 추적함으로써 재무 관리에서 AI의 실질적인 적용을 보여줍니다.
* 📑 마지막으로 이 세션은 **Zapier를 사용하여 맞춤형 보도 자료 생성기**를 구축하는 데모로 마무리되며, AI가 어떻게 콘텐츠 제작 프로세스를 간소화하면서 브랜드 일관성을 보장할 수 있는지 설명했습니다.

Zapier just had their Zapathon, and they shared a bunch of workflows that you can use to automate your work with AI. Now, as the guy that does the most AI automations and does the best AI automations, I was dying to know what was inside this webinar. So, I'm just going to show you the coolest things, and by the end of this video, I will compile all of them into a document that you can download at the end. You don't even need to sign up an email or anything like that. Let's here's an AI-powered automated personal message workflow. Never sleeps, never makes mistakes, and is always ready to get to work. So, let's kick things off in the Asana roadmap. Picture this: a project manager gets assigned to a project with a detailed description, and then voila! We tag it in our AI-driven zap swings into action on screen. You'll also see our safe filters and sharing. This zap sticks to its lane: our projects and not anything that's unrelated. Next up, meet our secret weapon, Chat GPT, the AI wiz who dresses up our

project briefs based on our prompts. It dishes out an articulate brief, and all we have to do is ask because cat GPT has an attention span. Yes, even AIS do. We've set up two separate prompts to ensure we gather all of the project details. The grand finale, our set, turns out a Google doc in our PM AI folder and pulls a new project from my project template. Plus, it uses Zac APIs to find the usernames of the PM and the sponsor. It's like magic. Lastly, it compiles all of that information into a lovely Slack message. Now for the moment of truth. We'll switch gears to Slack right here. You can see a message tagging the project sponsor and the project manager, serving up the name of the project along with a link to the brief, the schedule, and some other info. And it's all happening automatically and with just a click of this link. Boom! You're looking at the generated project brief. Take a little bit of data from this and from this end, from this and from this, and then packages all into a

slack message. It's like formatting plus a little bit of messaging, plus a little bit of summarization, all into one massive time saver. Here's how to create your own custom AI chatbot and custom AI workflows to make content a lot more efficient. My meta description bot takes care of writing the boring SEO description that has to accompany every blog post you write. All you have to do for my bot is copy-paste the text of the article that you've already written and give it a keyword that you're going after for SEO, and it will generate a meta description that's the perfect length for Google. The results are not revolutionary, but meta descriptions honestly shouldn't be. My blog introduction AI chatbot lets you input a quick summary of an article that you're about to write and then spits out inspiration for an introduction to that article in the style and voice of a Zapier blog post. My goal was to inspire Zapier content writers to push the envelope and make sure they're connecting with their audience, right?

from the jump, my social media content bot is designed simply to speed up the content creation process. All you need to do is take the blog URL that you would like to use, and the bot pretty much takes it from there, creating as many LinkedIn posts as you like, or you can specify how many posts you would like the bot to create for you. Wow, I would love to try to do this exactly with a YouTube video and create it into multiple tweets, LinkedIn posts, TikToks, etc. My article summarizer is designed to read any articles that are on a website and take the key takeaways and feed them back to me. It uses the power of Zapier in a Chrome extension, so whenever I go to any article that I find online and I just don't have the time to actually read through it, but I know it's interesting to me, I just click a button. It feeds all of that information into OpenAI, takes the main takeaways, and then it sends it back to me. It's given me so much more time as a leader to be able to

make sure I'm taking key takeaways that I think can help me in my job. This one I kind of use already, some great examples there. Here are four AI workflows for task management. We use Zendesk at Zapier for extreme on a lot of our support functions, and while it's great, it can be difficult to figure out exactly what kind of the topics of those conversations are. So, this app, you can see the trigger at the top here, that's coming directly from Zendesk. So, whenever a ticket moves to closed, it kicks off a webhook, which kicks off this app. From there, I have a code step where I format the conversation. Finally, I send it to OpenAI. The prompt here, nothing too magical, okay? It's categorizing it under all these four parameters, all the contacts that OpenAI has about the features that allows my team to see the number of conversations around a specific feature, the number of engineers who are involved in those conversations, and that lets me track how much time our engineers are spending supporting each individual.

feature over time. My zip is designed to make life easier for two user groups. The first user group is managers on the data team. So, this is a data org to do add items to the team that can create to-do tasks for our team just by simply putting a reaction on a slight message, and all the information gets automatically recorded in a structured way. So, we extract the folks that were Tech to extract the deadline. We extract the core of the message. This enables the people to directly get a structured list of what needs to be done and by when and who is it relevant for, so they can get directly to executing rather than having to keep track of this information themselves. This is a huge one, right? I think a lot of people have sent email else, and the communication is like one page, and then you're like, So, what do you want me to do? I also use the same type of workflow in chat gbt where I kind of ask for the action items that I have to do, but here you can do it automatically.

what my job does is it basically takes all the Twitter mentions from both my personal and my business account and it puts it all into one slot channel so that I can read all the responses all in one spot. And it basically does this by triggering the Twitter search mention for both those accounts and then takes an action by just setting it to the slot. It saves me a ton of time because I can basically see all the replies across all my accounts in one spot. Saved slack messages can often contain a lot of information and be relatively verbose. This app allows you to pare down the text to a single actionable to-do item when you save this level with the prompt, Provide a brief title or a to-do list item based on this slack message. It then provides a one-line task in a designated folder in Todoist. I found this particularly useful working in a busy global work environment. Once I've identified the conversations that need my attention, it allows me to easily triage and organize these items as tasks to tackle in my day-to-day. If you're bad

at productivity, and you're wasting a lot of time. It's not your fault; maybe there's just too many things going on. So, here are some AI workflows that you can use to boost your personal productivity. My Spanish face generator provides me with a daily dose of Spanish via Slack message. It prevents me from receiving duplicate phrases by extracting the phrase each day and storing it at a table that includes all of those previously used phrases in the OpenAI prompt with instructions. Now, repeat itself. It's helped me maintain a consistent daily learning habit with minimal effort. So, I manage all the things I need to do across my teams using a kanban board in Trello, and with my day being so filled with meetings. So, this girl was tracking her accomplishments. I often don't have time to fully read a long Slack message and parse what there is for me to do there. But if I can scan it and realize quickly that I need to come back to it, I can at least like add a quick Trello react G, and then ChatGPT will read it in more.

detail summarize what I need to do into a short to-do title and then add that to my backlog in Trello so this we see a lot summarization action step and then actually doing that, and I think that's such a good way to use communication, right? Like just breaking it down to exactly what you need. That power by OpenAI was designed to quickly add short summaries of my work accomplishments to a dedicated space in Notion, where I can keep track of my wins for one-on-ones, growth calls, and quarterly check-ins with my manager. By reacting with the specific emoji, mine is Incognito swag cat, to a Slack message in a public channel, OpenAI then summarizes that into a concise message for me and adds it into my database in Notion. By using this bot, I've been able to say goodbye to my old, very time-consuming process of digging through three months of Slack messages, handwritten notes, and reporting documents before writing my check-in. I use Zapier interfaces to host a chatbot that's powered underneath by

chatgpt and I instructed that chatbot to explain the concept that we provided and why it's important to someone with a college-level education but with no computer science background. I also tell the chatbot that it should write concisely enough that the answer fits on a flashcard to help it simplify the answer. Once it provides an output, I add a button that saves both the prompt that we provided it and the flashcard explanation to our dedicated page of Notion where we keep all of our key terms. This lets us review technical concepts as we're reading papers, reading blog posts, and also gives us a starting point for how we should explain these ideas to our more business-focused audience. And that's one of the best parts: like take this hard-to-understand thing and turn it into this easy-to-understand thing is what chatbots are really, really good at. Nice app helps me keep track of my expenditure and where my money goes. It watches for email alerts anytime there is a spend on my credit card.

GBT it extracts the amount in vendor. It then decides what category the expansion falls under using the vendor information and sends these details over to my Google sheet where I create my charts and get inputs. Amazing way to keep track of your budget and if you could train it on your data. You can maybe even say how much money do I have left for marketing this month or profit expenses. This one is really cool. Here are four zapping integrations back to back quite fast that all have chat GPT inside of them. The analyze free text sentiments app analyzes the sentiment of each row in a spreadsheet and adds its assessment to the row. This app allows me to easily chart what percentage of the survey responses fall into the sentiment categories of positive, neutral, or negative. And it's a quick way to summarize a large data set and communicate its meaning to my team. I built a custom button on the customer record in HubSpot. The sales and success teams can click that button to look up information from Looker, such as recent

support tickets and titles of live zaps of the customer within use chat GPT to summarize the support tickets and guess that the customer's main use case based on their zap titles. We take all of that information and create a note in HubSpot on that same record that the sales and success team click the button on so they have all that information in one place without even leaving HubSpot. My AI teammate's zap is powered by chat GPT and AI. It can be invoked by writing the AI names between square brackets anywhere in our team channel. It then keeps track of what happens in threads where they are invoked. The zap is designed to help the team leverage AI in a familiar way and to provide an additional source of information when necessary. So this one is so cool because you just have it in Slack and you can ask it, Hey, what do you think about this? and then put their name in brackets, and they will actually write in Slack. I call summarizer robot is designed to summarize sales and success calls and bucket those calls into use.

cases. This app initiates from a web hook push from Gone From There. We have another API call to gong to get the transcripts from the call. Next, we do a formatter step and we remove the beginning parts of the call that are typically small talk. Using OpenAI, a prompt request for a summary to be provided in a bulleted format and bucketed into a use case. We have a lookup step that then looks to find the contact that attended the gong call from our CRM and provide the link to the contact object into the zap. Last, we create a slack message with this summary, the use case, a link to the call, and the CRM object with the ability to summarize calls. People from all over the organization can have better insight into our customers without spending hours watching lengthy call recordings. The status page message generator bot is designed to increase the speed and efficiency of creating public-facing messaging about outages. That's happier. This automation starts with a type form, which asks our incident lead a few questions about the outage behavior that

information is then sent to chat gbt, where the AI creates a message around a specific set of language rules. Once the message is drafted, it's posted to Slack for approval, and with the click of a button, it's automatically posted to our status page. We want to provide an adhesive and streamlined experience for our users. With the status page message generator, we no longer have to spend extra time pondering over the wording. Our messages are clear, concise, and consistent. Like all of these automations, save anywhere from, you know, one minute all the way to 10 minutes, but it's just the ability to not even have to think about it anymore that I'm striving for in most of my business. Don't use chat gbt; use your own custom chat bot instead. Today, I'm going to show you how to build a press release generator that is customized to your company. Building an AI bot sounds intimidating, but it'll take about 30 seconds to do. We'll spend more time tweaking and perfecting your directive, which we'll get into. So, the first step is to log into your

zapier account and select interfaces on the left-hand side of your screen. From here, you'll click create. There are tons of different types of interfaces you can create, but we're going to scroll down and select zap chat. You'll see the screen as the interface is created. It may take a couple of minutes, but just be patient and don't refresh. When your interface is created, it'll look like this. You'll be able to connect tables and zaps to your interfaces as well. Before we do anything, we're going to customize this page. We're going to customize these fields to our project, so let's call this zapathon. This will update the URL for our press release generator and the name of the interface in zapier, so it's easy to find and organize. Save those changes, and it's going to take you back to the main screen. We're going to click on our new zapdot page. Then we'll hover over the text and click the edit box that appears on the right-hand side. You'll see all the fields that you can customize for

your Bot. Since we're going to be building a bot that will write press releases, I'm going to name this press release generator. You can also customize the gradient and prompt placeholder. Next, we're going to write the directive, which is what the bot will reference when generating results. This is a lot of text, but here's a template that you can use to give your chatbot very specific instructions. Usually, these things have a lot of good information that you should write or tailor to your company. I like to start by giving it a job and telling it what the main task to complete is, and then tell it how to generate press releases based on some industry standards. Then, I get into customization. I tell it who to attribute quotes to, what our brand attributes are, and that way it will weave those throughout the press release. I also include key messages. These things just save so much that the company wants associated with itself. Let's take an example of a real company so you can see these in action. I'm gonna do one that everybody knows, Disney. We're

gonna paste this into our directive in the AEI bot. We're almost ready to give this a try, but first, let's customize our URL and Page name so it's easy to locate later. Just click on the settings, and here you can adjust those settings, including the name and URL. If you own a custom domain, you can also use it here. We'll click save changes, and we're ready to try to give it a whirl. We're going to click on the link shown on the screen. This is going to open up the page that users will see and interact with. I'm going to use a real announcement from Disney so we can compare the results. Here's our prompt. I'm giving it a few basic facts, and let's see how it does. We actually have a quote from Bob Iger. It gave us a quote from the CEO of Make-A-Wish, and it scattered our brand attributes throughout. It also linked out to Make-A-Wish's correct website without me telling it to do that. It's pretty cool. Now, you'll also notice that I

hallucinated a contact name and information at the bottom since I did not provide it with that. This means it made that information up completely. That's one reason why it's important to still have a person review and edit these releases. Now, yeah, the jane.smith@disney.com. Compare it to the actual release, not too shabby. Disney's press releases are much longer than 400 words, so this is just a sample of the actual release that you can find on their website. But you can see how this will take you from a blank page to your first draft in about 10 seconds. I want to build a zap chat based around all of my previous YouTube videos so I can eliminate dumb comments forever. Hey, what's that AI tool that you talked about 10 videos ago that you probably remember that you only talked about for one minute? Well, if you have that question, ask my chat bot. This app is a game changer for any salesperson or anybody doing deals. Hey everyone, my name is Nicola and I work on the sales team here at Zapier, and I

built a bot that allows me to prep for my deals in an easy way. So, one of the problems that I was trying to solve was we get a ton of different customers here on sales, and in order for me to better prepare for these calls, I need to go to research each and every customer to learn more about them, to make sure that the solutions that I'm prescribing fit the challenges that they might have. So, I built this app that connects HubSpot, ChatGPT, and Slack. The way this app works is that for every time there's a new deal that gets created within HubSpot, we have a delay step that allows the HubSpot automation to happen through data enrichment. I then continue if this deal is assigned to me, and I have some steps here to split the domain so I can use that within a link later that I'll show, as well as grab the custom user ID within HubSpot, so it makes it easier for me to masquerade it. Now, for the ChatGPT steps, my prompt says, Can you tell

more information about the organization DraftKings, and I gave it some instructions as well. So I say, Hey, you're a helpful assistant helping me research my leads. Your goal is to provide me information that will help me in a sales rep. I want to make sure that the information is clear and concise, and that if there's no information about this customer since it was founded up for 2021, just let me know. The second thing that I asked the bot is, Hey, can you please give me challenges, or what are some of the top challenges a company like DraftKings might face, and then how my automation solves those challenges? What's unique about this prompt is that I ask it for every challenge presented, Please provide a discovery question that a sales rep could ask after that. What I do is I take all the information from this app and I send it to me within Slack in one single message. So here I can say, I have the contact information, I have the company size from steps five and six from above, I can enter in DraftKings to fill out the domain.

inspector link, as well as the account masquerade link with the custom user ID. This was something I added just a few days ago, where I also want to see how many zaps have been created within the account. And then finally, I have information from the chat GPT responses. So to give an example of what this looks like, I had a customer recently that I worked with from Global Silicon Valley. Here I can see just some high-level information about the company, such as their Venture Capital firm, and they focus on edtech. And then down here below, I have information around what are some top challenges that we see for my face right: communication and collaboration, operation lead management, data entry. And I have some solutions such as integrating collaboration tools and Discovery questions about asking them how they communicate today. All in all, like this app allows me to prep for all of my calls in one single place versus spending a bunch of time looking through various locations for this information. If you have any questions about this app, please let me know. Happy to help. I have

to say that this opened my mind. Like having all the information in one place allows you to just sit there with the data and meditate on different ways to find problems your customer might have and then solve them with Creative Solutions. So, this one is a game changer. So, if you want to try out any of these apps, I recommend checking out the link right here to get all of these summarized into a nice little document. Also, if you want to know my AI automations that I have on a weekly and daily basis, check out Patreon in the description down below. But I hope you got some value from this. I hope you use some of this to save you some time. And that's it. Alright, see you in the next one. Peace.

Top 19 Zapier ChatGPT Automations [Automate With AI]

Documenting The AI Singularity Daily With a Touch of Humor


AI Andy

Zapier의 Zapathon에서 제공하는 AI 기반 워크플로

Zapier는 개인 메시지 워크플로와 프로젝트 관리 자동화 등 업무 자동화를 위한 혁신적인 AI 기반 워크플로를 공유했습니다.

ChatGPT는 프로젝트 요약과 Slack 메시지를 생성하는 데 사용되어 커뮤니케이션을 간소화합니다.

이 비디오에서는 메타 설명, 블로그 소개, 소셜 미디어 게시물을 작성하기 위한 맞춤형 AI 챗봇과 핵심 내용을 요약하는 기사 요약 기능도 소개합니다.

Zapier의 Chrome 확장 프로그램으로 생산성 향상하기

Zapier의 Chrome 확장 프로그램을 사용하면 사용자가 흥미로운 기사의 핵심 정보를 OpenAI로 전송하여 빠르게 핵심 정보를 수집할 수 있습니다.

또한 이 확장 프로그램은 Slack 메시지와 트위터 멘션에서 할 일을 자동으로 생성하여 작업 관리를 간소화합니다.

또한 사용자가 작업을 정리하고 분류하는 데 도움을 주어 궁극적으로 시간을 절약하고 개인 생산성을 높일 수 있습니다.

사용자는 OpenAI의 도움으로 매일 사용하는 문구를 저장하고 요약하여 일관된 학습 습관을 유지합니다.

Trello 칸반 보드를 사용하여 작업을 관리하고 ChatGPT가 작업을 요약하고 백로그에 추가하여 Slack 메시지에 빠르게 반응합니다.

또한 복잡한 개념을 설명하고, 비용을 추적하고, 설문조사 응답 감정을 분석하는 데 챗봇을 활용하며, 이 모든 작업은 효율성을 위해 ChatGPT로 구동됩니다.

팀 효율성을 위한 AI 및 자동화 활용하기

영업 및 성공 팀은 HubSpot 버튼을 사용하여 고객 정보 및 지원 티켓에 액세스할 수 있으며, AI 기반 요약 및 사용 사례 제안이 제공됩니다.

Slack 기반 AI 팀원은 추가 정보와 지원을 제공하며, 통화 요약 로봇은 영업 및 성공 통화에서 얻은 인사이트를 제공합니다.

상태 페이지 메시지 생성 봇은 정전에 대한 대중 대상 메시지를 간소화하고, 사용자 지정 챗봇을 신속하게 설정하여 회사의 보도 자료를 생성할 수 있습니다.

보도 자료 생성 봇 사용자 지정

이 페이지와 필드는 보도 자료 생성기에 초점을 맞춘 'Zapathon'이라는 프로젝트에 맞게 사용자 지정되었습니다.

봇의 지침은 보도 자료 생성을 위한 실제 기업 사례로 디즈니를 사용하여 구체적인 지침으로 맞춤화되어 있습니다.

봇의 출력물에는 인용문, 브랜드 속성 및 링크가 포함되지만 연락처 정보도 포함될 수 있어 사람의 검토가 필요하다는 점을 강조합니다.

영업 통화를 위한 고객 조사 간소화

이 앱은 HubSpot, Chai GPT, Slack을 통합하여 영업 거래를 위한 고객 조사를 자동화합니다.

데이터 강화 및 도메인 분할을 사용하여 정보를 수집하고, Chai GPT는 조직과 그 과제에 대한 인사이트를 제공합니다.

이 앱은 모든 정보를 하나의 Slack 메시지로 통합하여 사용자가 효율적으로 통화를 준비할 수 있도록 도와줍니다.