Your team uses AI. Does your business?
A new Irish government study finds most firms use AI for individual tasks, not for the work the business runs on. How to take the step from one to the other.
Abhijit Chatterjee Chattopadhyay

Your team probably uses AI already. Someone drafts emails with it, someone summarises reports, someone asks it questions all day. That is useful. It is also not the same as your business using AI.
A new study for the Irish government puts a name on the gap.
What the study found
On 13 October, the Department of Enterprise, Tourism and Employment published A Sectoral Study on Artificial Intelligence in Enterprise in Ireland (opens in a new tab), prepared independently by Accenture. It looked at manufacturing, construction and tourism, with a separate part on the ICT sector's role in building and supplying AI, and drew on 272 responses to a consultation, four sector workshops and more than 30 interviews.
Its finding on how firms use AI: "Use also remains concentrated more heavily in individual tasks than in integrated workflows." In the consultation, about half of respondents described AI used for tasks, while far fewer were piloting it or had it embedded in a workflow, and about one in ten had scaled it.
Firms also described "fragmented data, continued reliance on spreadsheets and adoption centred on general-purpose assistants." And the study names half of the challenge for Ireland as helping firms "move from isolated tools and pilots into repeatable operational use and measurable business value."
Task versus workflow
A task is one person, one prompt, one result. A workflow is a piece of work the business runs again and again: every quote, every booking, every supplier invoice.
AI in a task saves one person some minutes. AI in a workflow changes how the work is done, for everyone, every time. That is where the value the study talks about sits, and also where the risks are, because a workflow runs whether or not anyone is watching.
Where to start with AI in your business
- Pick one workflow that repeats. Something that happens every day or every week, with a clear start and end, such as answering quote requests.
- Find where the data lives. If the information is spread across spreadsheets and inboxes, that is the first job, before any AI.
- Name an owner and a measure. One person responsible, and one number you will compare before and after, such as time to send a quote.
- Run the checklist before it goes live. Owner, budget, fallback, tests and an off switch. More in The one-page AI launch checklist.
The study also lists public supports in Ireland for firms at this stage, from Local Enterprise Offices to Enterprise Ireland and the European Digital Innovation Hubs. Worth a look before you spend your own money. If you want an outside view as well, see AI consulting for Irish teams, from Dubai.
Do this today
- List the AI tasks your team already does. Each one is a hint about which workflow to try first.
- Pick the one workflow that repeats most often and write down where its data lives.
- Decide who owns it and what number would show it worked.
Your next step
If you want help choosing that first workflow and getting it ready, book a free 30-minute 1:1 assessment. I look at your setup with you and tell you what the work would take.
Drafted with AI help, edited and approved by me.