Stop Copying Data Between Tools
Why I Do This
When I run an AI audit for a small or medium business, one of the first questions I ask is simple.
Where are your people copying information from one system to another?
The answers are usually the same:
- A website lead gets typed into the CRM by hand.
- Invoice numbers get copied into a spreadsheet.
- An order comes in, someone sends the confirmation, then messages the team.
- A receipt arrives as a PDF, and someone reads it and retypes the details.
None of these tasks is difficult.
But when they happen every day, they get expensive.
This is where many businesses go wrong with AI. They jump straight to one question:
“Which AI tool should we buy?”
Sometimes AI is the answer.
Often the answer is simpler: connect the systems you already have.
You don’t need someone copying the information. You may not need an AI agent either.
You may just need an automation workflow.
The hidden cost of copy-paste, in four numbers
30
enquiries arriving every day
5 min
to copy each one into the CRM
150 min
of data entry, every single day
~55 hrs
a month, for just one process
Illustrative example: 30 enquiries × 5 minutes × 22 working days.
When People Become the Glue
And that’s only one process.
The same business is often also copying:
- Orders into a spreadsheet
- Invoices into a management report
- Email attachments into Excel
- Bookings into a calendar
- Customer details into the CRM
Without noticing, the business has turned people into human connectors between pieces of software.
That isn’t a people problem. It’s a workflow problem.
Each tool is good at its own job. The trouble starts when information has to move between them.
Automation platforms such as n8n, Zapier and Make fix exactly that. The rule behind them is simple:
When something happens in one system, automatically do something in another.
A website form is submitted. A CRM lead is created, a salesperson is assigned, a follow-up task appears and a confirmation email goes out.
Nobody copies anything. The workflow does it.
How I Actually Work
I design every workflow in three layers.
Automation moves the information. AI understands it, but only where it’s messy. People make the decisions.
Move the information
Fixed rules. Same input, same result, every time. Runs around the clock and logs every run.
Examples: form → CRM, invoice → sheet, order → email, booking → team alert.
The plumbing
Understand the information
Used only where the input is messy: emails, PDFs and free-text requests.
Examples: read an email, classify a lead, extract from a PDF, draft a reply.
The thinking
Make the decision
People approve anything that touches money, contracts, staff or key customers.
Examples: review the draft, approve the quote, handle the exceptions.
The judgment
The safe order for any customer-facing workflow
That last link matters.
Automation → AI → Human → Customer is usually much safer than letting AI reply to customers on its own, especially when money, contracts, staff or important clients are involved.
And because the AI is just one step in the chain, it can be ChatGPT, Claude or Gemini. If a better or cheaper model comes along, you swap that one step. You’re not locked into a single AI provider.
Here’s how I work through a business, step by step:
Map Where People Act as Connectors
We start with the business, not the tools.
What do you sell? Where does the time go?
Then I look for the tell-tale verbs: copying, pasting, downloading, uploading, renaming, reformatting, and building the same report every week.
List the Tools You Already Pay For
CRM, accounting, email, Google Workspace or Microsoft 365, your website, WhatsApp, spreadsheets and project tools.
Most of what you need to connect is already paid for.
Pick the Smallest Fix for Each Task
This is where judgment matters. The answer might be:
- No new technology, just a better process
- An existing plugin or integration
- A simple automation
- Automation with one AI step
Recommending five AI subscriptions is easy. Choosing the smallest fix that works is the job.
Build the Plumbing First
Fixed steps go into a workflow tool: n8n, Zapier or Make, or a WordPress plugin such as OttoKit when your website sits at the centre.
They run on triggers around the clock and keep a log of every run, so you can see what worked and what failed.
Add AI Only Where Interpretation Is Needed
AI earns its place on messy input: an email in someone’s own words, a PDF invoice, a free-text booking note.
It reads, classifies and extracts. The workflow does the rest.
Start With One Workflow, Then Measure
Choose one process and time it today.
Build the simplest version and run it for a few weeks.
Then measure the result and decide whether to expand.
See It on One Messy Enquiry
Here’s where the three layers meet. Imagine this email lands in your inbox.
Nothing about it is neatly structured. The system has to work out who this is, what they want, how urgent it is and whether it’s a sales lead.
Email received
The workflow picks it up from the inbox.
AI reads and classifies
Is this a sales lead, and for which service?
AI extracts the details
Company size, requirement and urgency.
CRM record created
No retyping, no typos.
Lead assigned
Routed to the right salesperson.
AI drafts a reply
In your tone, ready for review.
A person reviews
Edit, approve or reject the draft.
Sent and scheduled
Reply goes out, follow-up is booked.
■ Automation fixed rules · ■ AI interpretation · ■ Human judgment
AI isn’t replacing the automation here.
It sits inside the automation and handles only the part that needs interpretation. Everything else is fixed steps, and a person still approves what goes out.
Three Everyday Examples
Let’s make it concrete with three situations I see all the time.
1. A website enquiry
A small property business gets leads through its website form.
Before
- Someone checks the inbox
- Copies the name, email and phone
- Opens the CRM
- Creates the lead
- Sends a reply email
- Sets a reminder, if they remember
Slow first replies, typing errors and forgotten follow-ups.
After
- CRM record created automatically
- Lead assigned to a salesperson
- Follow-up task created
- Confirmation email sent
Every enquiry logged the same way. No AI needed, just connected systems.
The person is still involved, but now where their judgment matters, not where data entry is needed.
2. Invoice reporting
Before: someone downloads each PDF invoice, reads it, and copies the supplier, invoice number and amount into a spreadsheet and then the management report.
After: the workflow catches the attachment, AI reads the invoice and pulls out the fields, and the spreadsheet and dashboard update on their own.
AI is useful here because an invoice is an unstructured document. Everything around it is ordinary automation.
3. A new booking
For a hotel, clinic or consultancy, a new booking can update the calendar, send the confirmation, alert the team and create the customer record automatically.
No AI needed.
If the customer writes a complicated request in the booking notes, AI can read it and flag it for the right person.
Don’t Add AI Where It Isn’t Needed
There’s a habit right now of putting AI into everything.
A business has a simple rule: when a form is submitted, create a record.
Then someone says, “Let’s add AI.”
Why? The workflow already knows exactly what to do.
Adding AI brings another moving part. You may pay for model usage. The output can vary. And a simple automation has quietly become an AI project.
AI should solve problems that need intelligence, not be added because it sounds more modern.
The same goes for choosing tools. I never start a project with “Let’s buy n8n” or “Let’s buy Zapier.”
The first question is always: what are we trying to automate?
Zapier makes it easy to connect lots of popular apps. Make gives you a visual way to build more complex flows. n8n offers more control, including the option to host it yourself. The right one depends on your team’s skills, how many workflows you run, your security needs and your budget.
I compare two of them side by side in n8n vs Zapier: Which Automation Tool Fits Your Business.
What I Bring
We don’t stop at a list of ideas.
You leave with one workflow running and a clear plan for the next one.
Your Copy-Paste Map
Every place your team moves the same information by hand.
Ranked by time lost, so you know where to start.
One Workflow, Running
Your most valuable process, connected end to end.
With a log, so you can see every run succeed or fail.
A Simple Decision Rule
Which tasks get automation, which get AI and which stay with people.
So the next workflow is quicker to build.
Frequently Asked Questions
Do I need AI to stop copying data between tools?
Usually not. Most copy-paste work is predictable, so a plain automation handles it. AI is only needed for messy input, such as emails written in someone’s own words or PDF invoices.
What’s the difference between automation and AI?
Automation follows fixed rules: the same trigger always produces the same result. AI interprets information that isn’t neatly structured. Good workflows use automation for the plumbing and AI for the reading and thinking.
Should I use n8n, Zapier or Make?
It depends on your team’s technical skills, how many workflows you run, which apps you use, your security needs and your budget. Decide what you’re automating first, then pick the tool.
My business runs on WordPress. Can that be connected too?
Yes. Plugins such as OttoKit connect WordPress forms, WooCommerce orders and other plugins to outside apps, and n8n, Zapier and Make can connect to WordPress as well.
Will automation replace my staff?
That isn’t the goal. Automation takes the data entry off their plate, so they get hours back for customers, sales and the work that actually needs a person.
Want your tools talking to each other?
Book a 20-minute call and we’ll find the one copy-paste task costing your team the most time, and what it would take to automate it.