My AI Workflow: Conquering My May 2026 Inbox & To-Do List

Hey everyone, Ryan here from agntwork.com. Hope you’re all having a productive week. It’s May 13, 2026, and if you’re anything like me, your inbox is probably a warzone, your to-do list is a mythical beast, and your brain is constantly trying to decide which fire to put out first. We live in a world that demands more and more from us, and honestly, sometimes it feels like we’re just treading water.

I’ve been obsessed with AI workflows for years now – not because I love jargon, but because I genuinely believe they can give us back something invaluable: time. And mental bandwidth. Today, I want to dive into something incredibly specific, yet universally applicable: how to stop drowning in your email and Slack messages using AI to triage and draft initial responses.

We’re not talking about a full AI takeover of your communications here. That’s a recipe for disaster and impersonal interactions. Instead, we’re focusing on the grunt work, the repetitive stuff that eats up hours every single day, preventing you from doing the deep work that actually moves the needle. Think of it as having a ridiculously efficient, always-on personal assistant for your digital comms.

The Inbox Apocalypse: My Personal Struggle

Let me paint a picture. A few months ago, before I really buckled down on this particular automation, my mornings went something like this:

• Wake up, grab coffee.
• Open laptop.
• See 70+ new emails and 100+ unread Slack messages.
• Panic.
• Spend the next two hours just reading and sorting, not actually doing anything.
• Feel exhausted and defeated before noon.

Sound familiar? For a tech blogger, someone who’s supposed to be on top of this stuff, it was embarrassing. My “inbox zero” philosophy was more like “inbox one thousand.” Client requests were getting delayed, pitches were being missed, and important internal messages were buried under a mountain of newsletters and notifications.

I knew there had to be a better way than sheer willpower. Willpower is a finite resource, and mine was clearly running on fumes. That’s when I decided to get serious about applying AI to this specific problem: intelligent inbox triage and first-pass drafting.

Why This Matters Right Now

Why is this topic timely? Because AI capabilities for natural language understanding and generation have gotten genuinely good. Not just “toy project” good, but “actually useful for real-world business” good. The tools are more accessible, the APIs are more robust, and the costs are lower than ever. If you tried this a year or two ago and gave up, it’s time to revisit. Seriously.

Also, the sheer volume of digital communication isn’t going down. It’s only going up. Relying on manual processing is no longer sustainable for anyone aiming for high productivity and focus.

Setting Up Your AI Comms Assistant: The Core Idea

The goal is simple: identify common communication patterns, extract key information, and either flag for immediate attention or draft a sensible initial response. We want to reduce the cognitive load of decision-making and typing for routine items.

Step 1: Identify Your Communication Patterns

Before you build anything, you need to understand what kinds of messages you get most often. Spend a week or two observing your own inbox and Slack. What are the recurring themes?

• Client inquiries: “How much for X service?” “Can you explain Y feature?” “When will Z be ready?”
• Internal requests: “Can you review this draft?” “Meeting reminder.” “Project update.”
• Support requests: “My login isn’t working.” “I can’t find X.”
• Sales pitches: “We offer a great tool for Y.” “Want to partner?”
• Newsletters/Promotions: The endless stream.

For me, as a blogger, a huge chunk was interview requests, guest post pitches, and inquiries about specific AI tools I’d reviewed. Another big one was internal team updates and requests for feedback on drafts.

Step 2: Choose Your Tools (No-Code Friendly)

You don’t need to be a developer for this, though knowing a bit of scripting helps. Here are my go-to tools:

• Email: Gmail with Google Apps Script, or Zapier/Make.com integrations.
• Slack: Slack Workflow Builder, Zapier/Make.com, or custom Slack apps (if you’re feeling adventurous).
• AI Engine: OpenAI API (GPT-4 or GPT-3.5 Turbo), or Claude API. I lean towards OpenAI for its broad capability.
• Automation Platform: Zapier or Make.com are fantastic for connecting everything without code.

Practical Example 1: Triage and Summarize New Client Inquiries (Email)

This is a lifesaver. Instead of reading every long-winded email from a potential client, I get a concise summary and a recommendation.

The Workflow:

1. Trigger: New email arrives in my “Client Inquiries” folder (set up a Gmail filter for this).
2. Action (Zapier/Make.com): Extract email subject, sender, and body.
3. Action (OpenAI API): Send the email content to GPT-4 with a specific prompt.
4. Action (Slack/Email): Post the summary and recommendation to a dedicated Slack channel or send a new, prioritized email to myself.

The Prompt (OpenAI):

You are an AI assistant designed to triage client inquiries for a tech blogger.
Read the following email and perform two tasks:
1. Summarize the core request in 2-3 concise sentences.
2. Based on the request, suggest a priority level (High, Medium, Low) and a recommended next action (e.g., "Schedule a call," "Send pricing info," "Decline respectfully," "Research further").

Email Subject: {email_subject}
Email Body:
{email_body}

Output format:
Summary: [Your 2-3 sentence summary]
Priority: [High/Medium/Low]
Next Action: [Recommended action]

This simple prompt, when fed through Zapier (Email -> Webhook -> OpenAI -> Slack), transforms my inbox from a minefield into a prioritized list. I get a Slack notification like:

*New Client Inquiry!*
Summary: [Client X] is interested in a sponsored post on AI workflow tools, specifically reviewing their new platform launching next quarter. They want to discuss rates and timeline.
Priority: High
Next Action: Schedule a 30-min discovery call to understand their platform and discuss pricing.

Boom. I know exactly what it’s about and what I need to do without even opening the email. This saves me easily an hour a day.

Practical Example 2: Draft Initial Responses for Common Requests (Email/Slack)

This is where the real time savings come in. How many times do you type “Thanks for reaching out! I’m reviewing this and will get back to you shortly” or “Here’s a link to our pricing page”? Too many.

The Workflow (Email – using Google Apps Script):

For recurring requests, like guest post pitches that don’t quite fit, or requests for my media kit, I use Google Apps Script to automate a first-pass draft.

1. Trigger: New email in a specific folder (e.g., “Guest Post Pitches – Review”).
2. Action (Google Apps Script): Read the email content.
3. Action (OpenAI API – via UrlFetchApp): Send email content to GPT-3.5 Turbo with a specific prompt.
4. Action (Google Apps Script): Create a draft reply in Gmail with the AI-generated text.

Google Apps Script Snippet (simplified):

This assumes you have your OpenAI API key stored securely and the `callOpenAI` function set up.

function autoDraftReply() {
 const label = GmailApp.getUserLabelByName("Guest Post Pitches - Review");
 const threads = label.getThreads();

 threads.forEach(thread => {
 const message = thread.getMessages()[0]; // Get the first message in the thread
 const emailBody = message.getPlainBody();
 const sender = message.getFrom();

 // Define the prompt for the AI
 const prompt = `
 You are an AI assistant for a tech blogger.
 The following is a guest post pitch. Analyze it and draft a polite, professional reply.
 If the pitch is generic or doesn't fit the blog's focus on AI workflows, draft a polite decline.
 If it's promising, draft a reply asking for more details or a specific topic proposal.
 Keep the tone friendly but professional. Do not sign off with a name.

 Guest Post Pitch from ${sender}:
 ${emailBody}
 `;

 const aiResponse = callOpenAI(prompt); // Custom function to call OpenAI API
 
 // Create a draft reply
 message.createDraftReply(aiResponse);
 
 // Move the email to a "Drafted" label to avoid re-processing
 thread.addLabel(GmailApp.getUserLabelByName("Guest Post Pitches - Drafted"));
 thread.removeLabel(label);
 });
}

// Placeholder for callOpenAI function - you'd implement this to hit the OpenAI API
// using UrlFetchApp.fetch and parse the JSON response.
function callOpenAI(promptText) {
 // ... (Your code to call OpenAI API with promptText) ...
 // For example:
 // const API_KEY = "sk-YOUR_OPENAI_API_KEY";
 // const options = {
 // "method" : "post",
 // "headers": {
 // "Authorization": "Bearer " + API_KEY,
 // "Content-Type": "application/json"
 // },
 // "payload" : JSON.stringify({
 // "model": "gpt-3.5-turbo",
 // "messages": [{"role": "user", "content": promptText}],
 // "max_tokens": 300
 // })
 // };
 // const response = UrlFetchApp.fetch("https://api.openai.com/v1/chat/completions", options);
 // const json = JSON.parse(response.getContentText());
 // return json.choices[0].message.content.trim();
 return "This is a drafted response based on your pitch. I'll review it shortly!"; // Mock response
}

Now, when I open my “Guest Post Pitches – Drafted” folder, I see fully formed replies, ready for a quick review and send. Instead of writing “Thanks but no thanks” twenty times a week, I just skim, maybe tweak a sentence, and hit send. It’s like having a writing assistant who already knows my voice and my business rules.

The Workflow (Slack – using Slack Workflow Builder + Webhook):

For common Slack questions, you can use Slack’s Workflow Builder to trigger an AI response. For instance, if someone asks “Where is the documentation for X?”, you can have an AI generate a reply with the correct link and a brief explanation.

1. Trigger: A specific keyword or emoji reaction in a channel.
2. Action (Slack Workflow Builder): Send the message content to a webhook.
3. Action (Make.com/Zapier): Receive webhook, send message to OpenAI API with a prompt to find relevant info from a knowledge base (if integrated) or provide a stock answer.
4. Action (Make.com/Zapier): Post the AI-generated response back to the Slack channel.

Example Prompt for Slack (OpenAI):

You are an AI assistant for a tech blog team.
A team member asked: "{slack_message_text}"
If this message asks for our team's internal media kit, provide the following link and a brief description:
Link: https://agntwork.com/media-kit
Description: Our media kit includes audience demographics, past partnerships, and rates for sponsored content.
If the message is about something else, respond politely that you can only answer questions about the media kit for now.

This can be expanded significantly. Imagine if your AI could pull information from your Notion docs or Google Drive. That’s a whole other level of assistant, but this simple example shows the power of automating common Q&A.

Important Considerations and My Learnings

This isn’t a “set it and forget it” solution. Here’s what I’ve learned:

1. Review is Key: ALWAYS review AI-generated drafts. They are first passes, not final. Sometimes the tone is off, or it misunderstands context. A human touch is non-negotiable.
2. Be Specific with Prompts: The clearer your instructions to the AI, the better the output. Experiment with different phrasings. I iterate on my prompts constantly.
3. Start Small: Don’t try to automate everything at once. Pick one or two high-volume, low-complexity communication types first. Get those working well, then expand.
4. Privacy & Security: Be mindful of sensitive information. Don’t feed proprietary client data or confidential internal comms into public AI APIs without careful consideration of your provider’s data policies. For highly sensitive info, self-hosted or private LLMs might be necessary, but for most general comms, commercial APIs are fine.
5. Ethical Implications: Be transparent if you’re using AI for initial drafts, especially with clients. I often add a small note like “Drafted with AI assistance” or just make sure my final review adds enough personal touch that it feels like me.
6. Cost: OpenAI API usage is generally inexpensive for text generation, but monitor your usage. A few cents per call adds up if you’re processing thousands of messages.

Actionable Takeaways for Your Week

Ready to reclaim your inbox and sanity? Here’s your homework:

1. Audit Your Comms: Spend 30 minutes tomorrow looking at your last 50 emails and 50 Slack messages. What are the top 3 recurring types of messages you receive?
2. Pick One Target: Choose ONE of those recurring message types that you dread dealing with manually. This is your first candidate for AI automation.
3. Map the Workflow: On a piece of paper or in a simple doc, sketch out the steps for handling that message type manually. Then, think about which of those steps could be handled by an AI (triage, summarize, draft).
4. Experiment with a Simple AI Prompt: Go to ChatGPT or Claude.ai (the free versions are fine for initial testing). Copy-paste an example of your chosen message type and ask the AI it or draft a reply based on your desired outcome. Play with the prompt until you get something useful.
5. Consider a No-Code Tool: If your experiment goes well, look into Zapier or Make.com. Most offer free tiers to get started. Try connecting your email or Slack to the AI for a simple triage step.

This isn’t about replacing human connection; it’s about eliminating drudgery so you can focus your human energy where it truly matters. Imagine having an extra hour or two every day, not just to catch up, but to get ahead. To think. To create. That’s the real promise of AI workflows, and it’s a promise I’m seeing fulfilled every single day.

Let me know in the comments below what communication challenges you’re facing and if you’ve tried any AI solutions!

Until next time,

Ryan Cooper

agntwork.com

🕒 Published: May 13, 2026