A one-page technology plan for the year ahead
How to write a simple one-page technology plan: where you are, what matters this year, a few priorities, a rough budget and a named owner.
· 5 min readSomeone in the office needs the current price list. They search the shared drive and find four candidates: "Price list 2025", "Price list 2025 FINAL", "Price list 2025 FINAL v2 (with edits)" and one sitting in an email attachment from March. They pick the most likely one and ask a colleague to double-check.
Most businesses have a version of this. It's a small, daily bit of friction that nobody quite gets round to fixing. And when you start reading about AI, you'll soon come across the advice that AI "needs good data", which can sound like a large, technical project standing between you and any real benefit.
It's much more manageable than that. For a small business, getting ready for AI mostly means ordinary housekeeping: knowing where things live, keeping one version of important documents, and making sure the right people can see the right things. It's work that pays off whether or not you ever switch on an AI tool. This article explains what it involves and how to start small.
What "good data" means for a small business
When large organisations talk about data readiness, they often mean databases, reporting systems and dedicated teams. For a business of 5 to 100 people, the picture is simpler. Your "data" is mostly documents, spreadsheets, emails and the information inside your everyday systems, such as your accounts package or CRM.
AI assistants that work inside your business tools, such as Microsoft's Copilot or Gemini in Google Workspace, answer questions by searching the files, emails and chats a person can already reach. They are only as useful as what they find. If there are four price lists, the assistant has the same problem your colleague had, except it may not think to ask.
So "good data" here means three practical things:
- Findable: information lives in known, shared places rather than on desktops and personal drives
- Current: there's one clear version of important documents, and old ones are archived or removed
- Appropriately shared: people can see what they need for their work, and sensitive information is limited to those who should have it
Know where your information lives
Start with a rough map. You don't need a formal audit, just a single page listing where the main types of information are kept. For example:
| Information | Where it lives today | Who relies on it |
|---|---|---|
| Client contracts and proposals | Shared drive, plus some in email | Directors, sales |
| Pricing and quotes | Several spreadsheets | Sales, office manager |
| HR and payroll | A folder on the finance PC and the payroll system | Finance lead |
| Job records and drawings | Sector software, plus a shared folder | Operations |
The gaps usually show themselves quickly. Perhaps important documents sit on one person's laptop, or a key spreadsheet is emailed around rather than shared. Those are the places to focus on, because they're invisible to colleagues and to any AI tool you might use.
If you've written a one-page technology plan, this map fits neatly alongside the inventory of systems.
Reduce duplicates and "final_v7" files
Multiple versions of the same document are one of the most common sources of confusion, and one of the easiest to improve.
Modern file storage helps here. SharePoint and OneDrive keep a version history of each file, and Google Drive does something similar, so people can edit one shared copy and go back to an earlier version if they need to. Once a team trusts that, the habit of saving new copies with new names tends to fade.
A few simple steps make a difference:
- Agree a single home for each important document, and link to it rather than attaching copies to emails
- Move older versions into an "Archive" folder, or delete them if they're no longer needed
- Agree a plain naming style, such as the client name, then the document type, then a date where it matters
There's a data protection benefit too. The ICO's guidance on storage limitation under UK GDPR says you must not keep personal data for longer than you need it, and should review and delete what you no longer need. A tidy-up of old files is a good moment to do exactly that.
A sensible folder and permission structure
Folder structures don't need to be elaborate. Organising by how the business works, such as by client, by department or by project, usually serves better than organising by person. Many businesses find a small number of shared areas (one for each team, plus one for the whole company) is easier to live with than dozens of ad hoc shares.
Permissions deserve a little more care, because AI tools make them more visible. Microsoft states that its Copilot only surfaces organisational data that each user has at least view permission to access, and Google says that Gemini has the same access to Workspace data as you do. That's reassuring: the assistant doesn't give anyone new access.
It does mean that old sharing decisions become easier to notice. A folder shared with "everyone" years ago, for convenience, can now be found with a simple question rather than a lucky click. Microsoft publishes a deployment blueprint on addressing oversharing for exactly this reason. For a small business, the practical version is simple: check who can see your most sensitive folders, and tighten anything that's wider than it needs to be. We covered this in more detail in our article on Microsoft 365 Copilot for a small team.
Keep sensitive information in the right places
Most businesses hold some information that needs extra care: salaries and HR records, health information, bank details, and client information covered by confidentiality agreements.
This doesn't mean locking everything down. A good aim is for sensitive information to sit in a few clearly defined places, with access limited to the people who need it, instead of being scattered through general folders and inboxes. That protects the business and the people whose information you hold, and it means you can introduce AI tools to the rest of the business with more confidence.
It's also worth being clear with the team about what shouldn't go into AI tools at all, particularly free consumer tools used with personal accounts. A short, friendly AI use policy helps here, and it's the subject of a future article.
Start with one well-organised area
You don't have to tidy everything before you begin. A good approach is to pick one area where good organisation would clearly help, get it into shape, and use it as the pattern for the rest.
Good candidates are areas that many people use and that change often, such as sales proposals and pricing, standard operating procedures, or project documents for current work. Once one area is tidy, people notice the difference quickly: fewer "which version is right?" questions, and faster answers when they search. If you later trial an AI assistant, that area also gives you a fair test, because you'll know the information it's drawing on is sound.
This is also a natural way to start with AI at all, as we described in AI for small businesses: a sensible place to start: begin with everyday friction and the tools you already have.
A sensible first step
This week, choose one shared area your team uses every day and spend an hour on it:
- List what's there, and move old versions and duplicates into an archive folder
- Check who has access, and remove anyone or any group who doesn't need it
- Ask two colleagues where they'd expect to find the key documents, and adjust the structure to match
That hour will make the area easier to use straight away, and it gives you a clear, low-risk place to try an AI assistant when you're ready.
If you'd like help planning a tidy-up or getting your information ready for AI, our AI Adoption and Cloud & Microsoft 365 services are a good place to start.
