Module 1 — full and free. No payment, no email required. If it's good, you'll know what to do.
Most people who take this course have already tried automating something. Maybe you asked ChatGPT to rewrite a few emails and it worked okay. Maybe you followed a YouTube tutorial for Zapier and got a zap running for about a week before it broke. Maybe you read a thread about "AI workflows" and came away feeling like you were missing something everyone else could see.
Here's what you were missing: you started with the tools instead of the work.
The people actually saving 15–20 hours a week didn't start by asking "what can I automate?" They started by figuring out where their time actually goes. Then the automation choices got obvious.
That's this lesson. Before you touch a single tool, you're going to map your week. Not a journaling exercise — just a tally. Where does the time actually go?
Most people are genuinely surprised by what they find.
Grab a Google Sheet or a piece of paper. Create these five columns:
Task | How Often | Minutes Per Session | Weekly Total | Notes
Now go through your last two weeks and write down every recurring task you did. Not projects — tasks. The stuff that happens regularly.
Some prompts to help you remember:
You don't need to be exact on the minutes. Rough estimates are fine. The point is to build a map, not an audit report.
When I ran this exercise before my first automation project, I found that about 40% of my working hours were going to things I could describe as "information shuffling." Moving data from where it landed to where it needed to be. Turning raw inputs into formatted outputs. Following up on things I'd already followed up on.
None of it was hard. None of it required any judgment. It was just there, every week, eating time I could have spent on work that actually needed me.
The exercise tends to produce two kinds of surprises.
The first kind: tasks that feel quick in the moment but add up over a week. Fifteen minutes of inbox sorting, twice a day, five days a week is two and a half hours. It never feels like that because you do it in small chunks.
The second kind: tasks you didn't even consciously register as tasks. The weekly copy-paste from your project management tool into a client email. The manual formatting of reports before you send them. The checking and re-checking of a spreadsheet to see if a number changed.
Write them all down.
Once you have your list, sort each task into one of three groups:
Category A: Pure Repetition
Same task, same inputs, same outputs. No judgment required. Someone (or something) just needs to follow the steps reliably. Examples: formatting reports, moving data between tools, sending scheduled updates, archiving files.
Category B: Template + Input
The structure is always the same, but the content changes based on something. Examples: client emails that follow a pattern but need context, proposals that reuse sections but require customization, status reports that need real numbers dropped in.
Category C: Judgment Required
This stuff can't be automated away. Strategy decisions, client relationship management, creative work that requires your specific perspective, anything where the answer genuinely changes based on context you can't easily write down.
Most people find their category A and B tasks take up more time than they expected. Those are your targets.
Before the next lesson, complete the worksheet and categorize everything. By the end, you should have at least 5 category A tasks and at least 3 category B tasks.
If you're struggling to get to that number, you're probably not being granular enough. Break big things down. "Email" isn't a task — "drafting weekly status updates to clients" is a task.
One more thing: once you have your list, look at the top 3 tasks by weekly total time. Those are your automation shortlist candidates. We'll come back to those in Lesson 3.
The most important shift in this lesson isn't the worksheet itself — it's the habit of asking where your time actually goes before deciding what to do about it. Every good automation starts with a real problem. The ones that stick are the ones solving something you'd notice immediately if it stopped working.
Your category A tasks are the ones to automate first. Fast to build, immediate ROI, and they teach you the core mechanics you'll use everywhere else.
Proof of Work
When I started working with a client on her reporting workflow, I asked her to do this exercise first. She estimated she was spending maybe 4 hours a week on reporting. The worksheet said 11. The difference was tasks she did in small chunks throughout the week that she'd stopped consciously registering as time.
By month two, she'd reclaimed 9 of those 11 hours. But we wouldn't have known where to start without the audit.
The worksheet takes about 30 minutes. Those 30 minutes will determine whether the rest of this course is worth your time.
There's a lot of content out there about what AI can do. Most of it is either breathlessly optimistic or weirdly dismissive. Neither is useful when you're trying to figure out whether to spend 3 hours building an automation.
So here's the practical breakdown. Not what AI will do someday. What it does reliably right now, in workflows like yours, without a developer on speed dial.
Drafting from context. Give AI a brief, some background, and a format — it produces a solid first draft. Works for emails, reports, proposals, summaries, social posts, documentation, anything with a predictable structure. The output usually needs editing. But "needs editing" is much faster than "writing from scratch."
Extracting and structuring information. Take a messy input — a PDF, a long email thread, a block of notes — and AI can pull out the relevant pieces and put them in a format you can use. This is one of the highest-ROI use cases. A 30-page contract becomes a table of key terms. An email thread becomes a one-paragraph summary. A meeting transcript becomes a list of action items.
Categorizing and sorting. Tickets, emails, feedback, leads — anything where you're making the same judgment about many similar items. AI can apply a consistent ruleset much faster than a person can, and it doesn't get tired or distracted by email 200 the way you might.
Following multi-step instructions. This is the one people underestimate. If you can write a clear sequence of steps for a task, AI can execute that sequence reliably on new inputs. The key word is "clear." Vague instructions produce vague outputs. Specific instructions with examples produce usable outputs.
Translating between formats. Data in a spreadsheet that needs to go into a report. Meeting notes that need to become a formal document. A verbal brief that needs to become a structured spec. All of this is AI at its most reliable.
Judgment calls where the right answer isn't in the input. If the correct response to a client depends on context that's not written down anywhere — the history of the relationship, things you know about their business, your read on how they're feeling — AI can't reliably replicate that judgment. It'll try. The output will look confident. But it's guessing.
Anything that changes every day without a clear pattern. Markets, trends, news, current events — AI can process these, but automating a response to them is hard because the right response changes. You can build systems that help you process them faster, but you can't usually take yourself fully out of the loop.
Creative work where "your voice" matters. AI can write. But if a piece of writing needs to sound like you specifically, with your specific opinions and style, you'll spend a lot of time fixing AI drafts. It's still faster than starting from scratch, but it's not a set-it-and-forget-it automation.
Customer-facing communication for sensitive situations. AI handles routine communication well. But anything involving conflict, disappointment, or a relationship that could break — keep yourself in the loop. Review before it sends.
Make consequential decisions you haven't pre-defined. You can automate the information gathering and the first-pass analysis. You cannot automate the decision itself in high-stakes situations. What you can do is automate everything around the decision so that by the time you're deciding, you already have everything you need.
Invent solutions to novel problems. If the problem is new and there's no pattern in your existing data to draw from, AI isn't going to figure it out. This is actually fine — these are the problems worth spending your time on.
Maintain quality without feedback loops. AI outputs drift over time if you're not auditing them. A prompt that produces great output in week one can produce noticeably worse output by month three as your workflow evolves. Build in checkpoints.
Here's the test I use before deciding whether to automate something:
If you answer yes to the first three and the stakes are manageable, it's a good automation candidate.
The biggest mistake people make is automating the wrong tasks — usually because they thought about what AI can do before thinking about what they actually need. The second biggest mistake is expecting AI to be fully hands-off when the task requires real judgment.
The sweet spot is: repetitive, structured, low-stakes-for-mistakes, high-volume. That's where you get the 15–20 hours back. And that's where we're building everything in this course.
You've mapped your week. You know where the hours are going. You know which tasks are pure repetition (Category A), which are template-plus-input (Category B), and which actually require your brain (Category C).
Now you need three specific tasks to commit to automating during this course. Not a wish list. A working list — tasks you'll have running by the time you finish Module 7.
This lesson shows you how to pick them.
Not every Category A or B task is worth automating first. Some are too infrequent to justify the setup time. Some have inputs that are too unpredictable. Some are so fast they don't matter.
The ones to prioritize have most of these:
High frequency. At least once a week, ideally multiple times a week. Daily tasks are the best candidates. The more often a task runs, the faster the automation pays for itself.
Consistent inputs. The task starts from something you already have: a spreadsheet, an email, a form submission, a calendar event. Messy, variable inputs slow everything down. Structured inputs make automation fast and reliable.
Clear definition of "done." You know what the output looks like. Not vague ("a summary") but specific ("a 3–5 bullet summary of the email thread, saved to the client folder in Google Drive"). If you can write down what good looks like, the automation can produce it.
Low stakes for a mistake. The first automation you build will probably fail at least once. Pick something where a failure doesn't matter — an internal doc, a draft that you review before sending, a data move that you can re-run if it goes wrong. Save the client-facing stuff for when you've got more confidence in the system.
Currently annoying. Honestly, this matters. Automating something that genuinely bothers you every week is more motivating than automating something theoretically optimal. Pick at least one task that you'll be relieved to never do manually again.
Weekly status reports. You have data somewhere — a project tool, a spreadsheet, a CRM. Every week you format it into an update you send to clients or management. The data is already there. The format is always the same. The only thing you're doing is the translation. This is a Category A automation that usually takes an afternoon to build and saves 2–4 hours a week permanently.
Email drafting for repeat scenarios. Think about the emails you write that follow a pattern. Inquiry responses, follow-ups after a meeting, project kickoff emails, client updates when something changes. You're always pulling from the same mental template. That template can be written down, given to AI, and used to generate drafts you review and send. Not fully automated — but much faster.
Follow-up sequences. A lead goes quiet. A client hasn't responded. A proposal is sitting unanswered. If you track these in any tool at all, you can build a system that notifies you when something needs attention and drafts the follow-up. The drafts need your approval before they send — but right now you're probably forgetting to follow up half the time, so even a draft-and-remind system is a significant improvement.
Data moving and formatting. Something lands in one place and needs to be in another place in a different format. Form submission to spreadsheet to formatted report. Invoice to expense tracker to monthly summary. Social post engagement metrics to your weekly review doc. These are often the easiest automations to build because they require almost no AI — just plumbing between tools.
Meeting prep briefs. Before client calls, you probably look at the last few emails, check what was promised, maybe review a status doc. A 20-minute prep ritual that happens 3–4 times a week is an hour a week doing research you already have the tools to automate. Brief pulls automatically from email thread + last update + open items and lands in your notes 30 minutes before the meeting.
Now pick your three. Use this format:
AUTOMATION SHORTLIST Task 1 (build in Module 2): Name: ________________ Trigger: what starts it? (e.g., new email from client) Input: what data goes in? (e.g., email thread contents) Output: what comes out? (e.g., draft reply saved to drafts folder) Current time: ___ minutes, ___ times per week = ___ min/week Task 2 (build in Module 3 or 4): [same format] Task 3 (build in Module 4 or 5): [same format]
The trigger/input/output breakdown matters more than you think. When you can state these three things clearly, the automation is 80% designed. What follows is just execution.
Module 2 takes your first shortlist task and builds it. By the end of the module, you'll have a working automation running against something from your actual workflow.
Not a demo. Not a toy example. Your stuff.
Proof of Work
When I started working with a client on her reporting workflow, I asked her to do this exercise first. She estimated she was spending maybe 4 hours a week on reporting. The worksheet said 11. By month two, she'd reclaimed 9 of those 11 hours — but we wouldn't have known where to start without the audit.
Module 2 through 7 build the actual automations — email pipelines, reporting systems, client work flows, and the connected stack. One working automation by the end of Module 2.
14-day refund policy. Questions: t.me/sage_executable