I read a piece recently about why nobody at work wants your AI workflow, and it landed with me because I have seen this play out in very different rooms.
Big bank. Small business. Charity. Internal team. Same problem, different furniture.
Everyone says they want to be more efficient. Then you look at the actual work and it lives in Word documents, Excel sheets, shared drives, inboxes, Teams chats, old CRMs, finance systems, and that one spreadsheet nobody wants to touch because it somehow runs the whole process.
That is the bit people miss. You cannot just drop an AI tool into the middle of that and expect people to change how they work because the demo looked clever.
The Work Already Has a Home
Most businesses are not short of tools. They are usually drowning in them. There is a tool for accounts, a tool for projects, a tool for sales, a tool for documents, a tool for signatures, and then a pile of spreadsheets holding the bits that never quite fitted anywhere.
People know those tools. They know the awkward bits. They know which report takes too long, which export needs cleaning up, and which customer update always gets copied into three places.
So when someone says, "we are going to introduce an AI workflow", the first reaction is rarely excitement. It is usually another password, another system, another thing to remember, and another meeting about process.
I understand that reaction. A lot of technology projects ask people to change before the technology has earned any trust.
People Do Not Change Because a Tool Exists
This is true in large organisations and it is true in small ones. The bank might have governance boards, procurement rules, data teams, and change managers. The small business might have one person who knows how everything works because they have been doing it for years.
The scale is different. The human problem is the same.
People change when the new thing clearly makes the work easier, when it fits around the way the job is already done, and when it does not make them feel silly for knowing the old way.
That last bit matters. A lot of office work looks messy from the outside because it has grown around real constraints. A spreadsheet that looks ugly might be carrying ten years of small decisions. A shared inbox might be the only place where three teams can see the same thing. A document template might exist because the official system cannot handle one important exception.
If you ignore all of that, your AI project is already wobbling.
Start With the Busywork
The better starting point is much more boring. Ask where the busywork is.
Where is someone copying the same information twice? Where are staff waiting for a reply before they can move something on? Where does a customer email become a spreadsheet row, then a calendar entry, then an invoice note? Where does a manager lose half an hour every Friday pulling numbers out of three systems for a report nobody enjoys writing?
Those are the places AI and automation can help first.
Not by replacing the whole process. By taking a chunk of friction out of it. Summarise the long email. Draft the reply. Pull the numbers together. Check the form for missing details. Turn the call notes into actions. Move the data to the right place. Flag the odd case for a human instead of pretending the system knows everything.
That is not as exciting as a keynote demo. It is much more useful.
Keep the Tools People Trust
A lot of the best automation work does not start with a new platform. It starts with the tools people already use.
If your team lives in Outlook, Teams, Excel, Google Sheets, Xero, a CRM, or a shared folder, the first question should be how to make those tools work harder for you.
Sometimes that means a proper integration between systems. Sometimes it is a small internal web app that sits in the middle and gives everyone one clean view. Sometimes it is a document workflow that stops people hunting through folders. Sometimes it is a careful bit of AI that reads, drafts, checks, or sorts information before a person makes the decision.
The important thing is that people should not have to abandon everything they know on day one.
We have written before about business process automation, and this is the same principle with AI added carefully. Find the repeated task. Understand the judgement involved. Automate the safe bits. Keep a person in charge of the bits that need context.
Where Community Tech Helps
This is the kind of work we like at Community Tech because it is practical. We are not there to sell a shiny AI product and hope the team bends around it. We sit down with the people doing the work and ask what actually happens between the customer asking for something and the job being finished.
Then we look for the drag.
Forms that could fill themselves from existing data. Emails that could be drafted from a template and checked before sending. Reports that could build themselves from the systems you already pay for. Internal tools that stop information living in five different places. AI checks that help catch missing details before they become tomorrow's problem.
For some organisations that becomes a small automation project. For others it becomes a custom web app or internal system. Sometimes the honest answer is to leave the software alone and tidy the process first.
We are fine saying that. Bad automation just makes a bad process faster.
Use AI Where It Earns Its Place
I use AI every day. Writing, coding, checking ideas, reading through messy information, getting through admin that used to eat half a morning. Used well, it saves a serious amount of time.
But it only works when you are honest about the job. AI is good at handling drafts, summaries, patterns, classification, and repetitive checks. It is not magic. It still needs boundaries, review, and someone who understands the consequences when it gets something wrong.
That is why the best AI work often looks small at first. It removes one irritating step. Then another. Then another. People trust it because it has helped them with the work they already had, not because someone told them they were now part of a transformation programme.
The Point
Nobody wants an AI workflow.
They want the quote done faster. They want the report to stop taking Friday afternoon. They want the inbox under control. They want customer details to be in the right place without typing them twice. They want fewer mistakes, fewer chasers, fewer "where did we put that" moments.
That is the useful work.
If your team is already using good tools but still losing time to copying, chasing, checking, reporting, or moving information around, we can help you find the bits worth fixing. Get in touch and we will talk through what is actually happening now before suggesting what to automate.