Copy-paste automation builds for freelancers, consultants and solo operators. No code. Free-tier route documented for every single one.
Four things that will save you a wasted weekend. Read these — they matter more than any individual workflow in here.
The failure mode for this Blueprint is reading all twelve, feeling energised, planning a full rebuild, and finishing nothing. Pick the task you dread most. Build only that one. Come back next week.
The rule that keeps this safe: anything that leaves your name on it and reaches another human gets a human glance first. Collecting, sorting, drafting, formatting — let it run unattended. Sending — approve it. That glance costs you eight seconds and prevents the one embarrassing failure that would make you abandon all of this.
Nine of these twelve run entirely free. Upgrade when a limit genuinely annoys you, not because paying feels like commitment. Every workflow below lists its free route.
An API changes, a format shifts, something silently stops. Check your automations once a month. Build a five-minute monthly review into your calendar now, before you need it.
You open your laptop and immediately spend forty minutes inside other people's priorities before touching your own work.
DECISION, REPLY, or NEITHER. Constraining the output to one word is what makes the routing reliable — leave it open-ended and you'll get paragraphs you can't branch on.NEITHER gets archived. DECISION gets a label and a three-line summary written into the thread. REPLY goes to step 4.Every proposal is 80% identical to the last one, and you rebuild it from scratch anyway — usually at 11pm.
{{CLIENT}}, {{SCOPE}}, {{TIMELINE}}, {{PRICE_MID}}.{"scope": "...", "timeline": "...", "stakeholders": [...]}. Structured output is what makes the next step reliable.You post consistently for three weeks, client work gets heavy, you go quiet for two, and start from zero.
Chasing money is the worst part of working for yourself — so you delay it, and then you're both owed money and annoyed about it.
The recap you promised at the end of the call gets written two days later, if at all — and by then you've lost half the detail.
Someone enquires at 11pm. You reply at 9am. By then they've messaged two competitors and one of them answered first.
Month-end reporting is four hours of copying numbers between tools and writing the same commentary in slightly different words.
You think clearly while walking and badly while staring at a blank document. So the document doesn't get written.
Most work you lose isn't lost to a competitor. It goes cold because nobody followed up and it felt awkward to chase.
Every new client sector means hours of reading before you can speak credibly about it.
The first week with a new client is your best chance to look organised, and it's the week you're most scattered.
Sunday evening anxiety is almost always a lack of visibility, not a lack of capacity. You don't know what's outstanding, so you assume everything is.
Nothing in this Blueprint requires payment to work. Here's the zero-cost route for each paid recommendation — slower and clunkier in places, but functional.
| Paid recommendation | Free route | The trade-off |
|---|---|---|
| ElevenLabs (transcription, voice) | Whisper via free tier, or built-in phone dictation | Lower accuracy on accents and crosstalk; no speaker separation |
| Jasper (drafting, repurposing) | Any free AI chat tier, run manually | Loses automation — you paste in and out by hand |
| Make.com paid tier | Make.com free — 1,000 operations/month | Enough for 4-5 workflows; you'll hit the ceiling around workflow six |
| Paid schedulers | Native platform scheduling | One platform at a time; no cross-posting |
| Symptom | Usual cause | Fix |
|---|---|---|
| Workflow stopped silently | Expired authentication | Reconnect the account; set a monthly reminder to check |
| AI step returns unusable output | Prompt too open-ended | Force a structure — JSON keys or a single word. Constraint fixes most of these. |
| Duplicate actions firing | Trigger firing on updates as well as creation | Add a filter for first-run only |
| Hit the operations limit | Trigger too broad | Narrow the filter before the trigger, not after |
| Wrong information sent out | Approval step skipped | Re-add it. This is why the rule exists. |