Hello Reader
& happy Friday! 🌿🕊

​
This month inside the Quiet AI Accelerator, we're moving up a level with AI (level 3, see below). Here's what's inside this week:

→ The upkeep AI tutorials tend to skip: traps hiding in AI automations
→ A 3-minute Testimonial Miner walkthrough (short video, use 1.5 speed if my accent is too much 😆)
→ Automations you can test easily
→ A small holiday check-in ☺️

​
​
The upkeep AI tutorials tend to skip

On a recent Zoom call, someone told me they'd switched off their automations because they were draining their weekly Claude allowance. The tutorials had shown them the setup. They hadn't shown the upkeep.

I wasn't surprised. That's what can happen when "automations" come first.

Even a simple scheduled task can hide traps:

→ Instructions that contradict each other or point to missing files
→ Folders the AI can't fully work through in a single run
→ Permissions that stop the task halfway, or allow changes you didn't intend
→ Runs that happen too often, or at the wrong time (yes, set them up ~5 hours before you wake up so you don't lose your morning session)
→ A model mismatch: paying for heavy reasoning on a light task, or the reverse

And so on.
Not to mention automating skills (AI assistants) before they perform at the level you need.

My rule is simple: make level 2 work well first. Your skills should already be saving you real time. Then decide whether each one would benefit from an automation on top. If it would, add them one at a time and watch how each performs before adding the next.
​

​
A small skill that now saves me real time

Last week I showed you how my Pitch Writer automation increased my monthly pitches by 500% (I didn't expect that either).

This week it's another AI tool that has made a big difference for me: my Testimonial Miner.

Collecting client feedback used to be something I put off. Kind words sat in Google Drive, emails and DMs, and plenty slipped through because I hadn't set aside time to gather them.

Now it takes a few minutes per call transcript, and Claude organises the quotes inside Google Drive for me. (If you use Senja- the testimonial app, the Miner works well as the step before it: it digs deeper into transcripts and surfaces quotes you'd otherwise miss)
​

It took a few rounds of changes. Now I'm happy enough with it to add an automation on top: one that checks for new transcripts, runs the whole process, and updates my library on its own.

Watch the Testimonial Miner walkthrough →
​

​

​
Simple Cowork automations you can test

Depending on the files and access you give it, you could start with:

→ Receipts to expenses: organise receipt details into a spreadsheet, with totals and anything unreadable flagged for review
→ Spreadsheet cleanup: identify duplicates, standardise formatting, mark incomplete rows and save a clean copy
→ Folder organisation: group files and suggest what could be archived. It reorganised my Downloads folder of 8,000 files 😅 (I would have deleted them all otherwise)
→ Morning overview: bring together your calendar, tasks and unread emails
→ Competitor snapshot: compare pricing, homepage messaging and lead magnets in one table

And plenty more.

Some are useful whenever you need them. Others may deserve a recurring schedule once you're happy with the results.

For a solopreneur expertise-based business, pitching and content repurposing are where I see the most value. That's what my clients will implement next.
​

Where are you with this? Still improving your assistants, or already happy with what you've automated?

Reply and tell me which of the examples above you'd like to try. I'll send you a prompt to get started, along with the instructions you'll need to adapt it.
​

And a small holiday check-in 🌿

The water was wonderful, as always.

While I was away, I did something I now do whenever I step back from work: I held my vision up against the life I'm actually living.

How far have I come? What has shifted? Is there enough joy and peace in the ordinary days? 🌝

The earlier version of me skipped this entirely, my own why remained unexplored for a long time. Now I regularly think about the life I want, then build the business that fits inside it.

That order also helps me decide where AI belongs, how much complexity I'm willing to manage, and what I can leave out.
​

And it matters more each year. With AI reshaping jobs, I expect more professionals to start something of their own. Daniel Priestley makes the same point in one of his recent videos, arguing for building assets around your expertise: IP, content, data, code, and books too.
​

If you're considering an expertise-based business, give yourself room to think about both: first the work that fits your expertise, values and desired lifestyle, and the specific problem someone would pay you to help solve (basically an aligned niche that combines identity and topic).
​

The same applies to the AI you bring into it. There's enough to do already. It needs to earn its place.
​

Warmly,
​
Diana 🩵
​
​

P.S. Quiet AI Business Accelerator reopens in October for its final founding-member round at $79/month, slightly above the first round. These founding rates won't be available again. Inside, we work through all eight business systems, beginning with VITO (Vision), and build your AI infrastructure one layer at a time.
​
If you'd prefer to begin with the vision work alone, my VITO workshop covers it in detail.