how I automate 90% of my bookkeeping (and why it's wrong)

The dark side of automation that nobody talks about

I feel like everyone’s completely burnt out right now.

The teams I coach
The network I speak with
The other Founders I know

Maybe it’s just the time of year.

But, they’re all totally frazzled.

I don’t know about you?

Yesterday, I posted on LinkedIn about ‘The Dark of Steve Jobs’.

One of my best performing posts in a while.

(the video wasn’t mine, but the caption was).

As the whole thing got me thinking about this concept of:

“Signal vs Noise”

People don’t burn out because they’ve got too much to do.
People burn out because they’re doing too much of the wrong things.

High-performers have a ruthless focus on activities that support their goals.
And a ruthless intolerance for ‘noisy’ activities that take them away from them.

Everybody sells the dream of ‘automating away’ your problems.

But you’ve got to be careful that solving problems now.
Does NOT create more problems for you in the future.

Here’s the trap to avoid.

Technical Debt

This is the dark side of automation that nobody talks about

Technical debt.

It works in the same way as financial debt.

You get a short term benefit (time or cashflow)
You get a long term drawback (lost time or increased spending)

The simplest example:

I create a spreadsheet for a process that saves me 1 hour a week right now.
I evolve that spreadsheet into 13 that then lose me and my team 3 hours updating it.

This then falls into the “This is how we’ve always done it” mentality.

And for years, you end up wasting time on noise, losing opportunities to create signal.

  • 30 mins updating a spreadsheet is 30 mins lost networking

  • 60 mins manually downloading files is 60 mins lost speaking with clients

  • 90 mins writing meeting notes and generating reports is 90 mins time lost delivering more value to clients

And if that unscalable process compounds?

You quickly see how people end up time poor, burnt out and grumpy.

System that Scale

The general rule is.

Don’t set up anything now that’s going to cause you headache in the future.

The simple trick is just a ‘play it forward’ exercise.

Ask yourself:

“If I create this spreadsheet, process, automation, documentation now.
How will this change when I’m double the size, or have double the work?”

Here’s me going through this exercise with one of my own internal processes.

Current Workflow (plus future plan)

Here’s my current semi-automated bookkeeping process.

If you copy it, will it save you a tonne of time?

Yes…

Just be mindful of creating technical debt that might lose your time in the future (as you’ll see below)

The below step by step includes:

  1. My current process

  2. A commentary of what I plan to do to make this more scalable.

(you’ll have to stay tuned for a future newsletter to see whether I actually achieve this or not..)

Tools

  1. Outlook Folder

  2. Make.com Automation

  3. SharePoint File Storage

  4. Notion Database Audit trail

  5. FreeAgent (and Mettle) Matching

1. Move invoice to ‘Pending Invoices’ Folder

Every time I receive a billing e-mail from my supplier, I drag it into a folder.

I don’t have a mailbox rule for this (it probably only takes 30 seconds as it is)

Works fine atm, but easy to miss the e-mail when grow to paying 50-100 invoices each month.

Future Plan

  1. Have AI monitor my inbox

  2. Any e-mails from known suppliers that contain an invoice

  3. Move automatically (or not at all if the AI can trigger the workflow without needing a folder)

2. Make.com Attachment Download

Make E-Mail Attachment Automation

I then have a Make.com automation that triggers once a week.

It looks at the e-mail folder, downloads the attachments.

Then moves them to a specified supplier folder in SharePoint

The problem with this, is every time I onboard a new supplier, I have to create another chain of logic. And if the supplier e-mail changes, then the automation breaks.

Future Plan

  1. Have AI intelligently recognise what invoice is from what supplier

  2. Have AI automatically store it in the correct folder

Dedicated tool?

Right now, I store invoices outside of FreeAgent, because I literally only bring bank statements into it. I don’t use any of the ‘Bills’ capability, as - right now - I don’t have any payment terms with suppliers.

If I did use this facility then using something like Sage AutoEntry, or Dext would be an easier off the shelf option.

3. Notion Database for Quick Matching

Notion Invoice Audit Trail, VAT and Matching

At the end of the Make.com automation, it creates a new Notion database item, with the invoice name, supplier, and original e-mail.

The database also includes a column for whether that supplier has VAT or not (so I can make sure it’s correct in FreeAgent)

Future Plan

I think I’d keep this as it’s useful to have an audit trail, and see progress on what has / hasn’t been matched in FreeAgent.

4. Notion Database for Manual Downloads

Notion Manual Invoice Download Locations plus VAT treatment

On the same Notion page, I then have an additional database, that shows the billing link I need to use to download invoices manually if I don’t get them sent by e-mail.

This is the most manual part of the process.

Future Plan

  • Option 1 - Use something like ChatGPTs Atlas browser to autonomously open the links I give it to download the invoices

  • Option 2 - Use Claude for computer use to achieve something similar

(this will all come down to how useable these tools are)

The biggest issue with both of these will be 2 Factor Authentication.

It might be able to try and login if it’s got my password pre-saved in the browser.

But will they be able to access my phone / e-mails for the authorisation codes? Probably not.

To be continued…

5. FreeAgent Matching

Then, all that’s left to do is to go into FreeAgent, pull up my bank transactions and match / explain them off once I’ve got the documents filed.

The Bottom Line

Even with a simple automation like the one I’ve explained above.

I’ve still created:

  1. An element of technical debt that doesn’t scale

  2. An element of manual work (albeit, not a massive amount)

So, the next time you think about automating something.

Pause.

Play it out.

Ask yourself.

“Is the short term benefit going to create long term pain”.

Sometimes it pays to automate
Sometimes it pays to autowait

(see what I did there?)

And don’t forget.

Breaking through the noise for you might not be automation.

Taking words out of Eisenhower’s mouth… You can also delegate and delete.

Don’t do something that can be delegated
Don’t delegate something that can be automated
Don’t automate something that can be deleted

Simple is often better.

Until next time

Keep on building

Adam

P.S. - What did you think of this week’s edition? Hit reply and let me know (I read all replies)

P.P.S - In exchange for referring 5 people to my newsletter, I will give you 30 mins FREE consulting (normally $99) on anything you're struggling with that can be solved with AI & Automation. From speeding up your delivery, to working less hours, or winning more clients, we can tackle anything you like. Share using the link below.