AI agents for business: choosing your first task to automate

Choose a recurring task with accessible source data and an output someone can reliably check. Keep the first AI agent pilot narrow: gather information or prepare a draft, then measure the work left for a person. A fast model response does not, by itself, make the process worthwhile.

Consider a manager handling workshop enquiries. One message includes a date and a group size; another says only “we’d like to bring the team”. Copying fields is easy. Working out what is missing, finding the relevant terms and drafting a useful question takes judgment. That small stretch of work is a plausible pilot candidate.

Do you need an agent, a workflow or a simple integration?

An AI agent uses a model to choose its next steps and call tools in pursuit of a goal. Definitions vary, so agree on the intended freedom: will the system follow a fixed route or decide what to investigate next? Anthropic distinguishes predefined workflows from agents that dynamically direct their own work.

Compare the pilot with the simplest useful alternative, such as a better form and a response template. Comparing only with fully manual work can hide a cheaper solution.

• Anthropic: workflows and agents

Find the task by watching someone do it

Walk through recent enquiries with the person handling them. Ask where they copy information, make a decision or discover that a fact is missing. A real record gives you a firmer brief than “automate sales”.

• Frequency: the task recurs often enough to collect examples and observe a meaningful change.
• Data: there is an authoritative, current source with access limited to what the task requires.
• Verification: someone can distinguish a correct answer from a plausible one without repeating the whole job.
• Consequences: a wrong draft can be corrected before a message, payment or record change takes effect.
• Ownership: a named person reviews output and handles exceptions.

If five documents disagree about the price, resolve that first. A stronger model cannot know which price your manager approved unless the approval is recorded. Cleaning up the source material may turn out to be the most useful part of preparation.

Example: drafting a reply for a workshop

We use Forma, the fictional pottery studio from our website guide. This is a proposed process, not a deployed agent or a report of client results.

• The previous step: building and testing the enquiry form

The system extracts the request, checks approved terms and prepares a draft with source references. Missing dates trigger a clarification question. A class timetable is not evidence of available places: that requires a separate current source.

Start with read access and a separate draft destination. If you later allow sending, approval should cover the exact recipient and text. A retry after an interruption must not quietly send a duplicate.

Test the pilot before involving customers

1. Measure the existing process, separating active work from waiting for the customer. Include corrections.
2. Collect routine cases and exceptions. Write expected outcomes and forbidden actions in advance.
3. Use some cases to improve instructions and hold others back for evaluation.
4. Run without sending customer messages. Record system version, output and review time.
5. Count every attempt, including failures completed manually.
6. Decide whether to expand, narrow the task, fix the data or stop.

A single accuracy percentage conceals very different mistakes. A clumsy greeting and an invented price need different responses. Track unsupported terms, missed clarifications and actions beyond permission separately. Add targeted tests for rare, serious failures; not seeing one in a small sample does not establish that it is unlikely.

Before testing, agree what would justify continuing. Every draft must be reviewed, unconfirmed prices must not reach customers, average human work should fall and the team must be able to handle the exceptions. Set thresholds for your own process; there is no success rate or sample size that fits every pilot.

• Download a blank pilot evaluation log (CSV)

Calculate the value after human review

Our illustrative workload is 300 eligible attempts a month, previously taking 12 minutes each. Assume an average of five human minutes remains after introduction, including review, correction and manual completion of failures. These are assumptions, not model performance measurements.

Hours released: 300 × (12 − 5) / 60 = 35 per month.
Capacity value: 35 × €30 = €1,050.
Less running costs: €1,050 − €250 = €800 per month.
Conditional recovery of setup cost: €2,400 / €800 = 3 months.

The 300 attempts cover the chosen task, not every enquiry the company receives. The five-minute average includes unsuccessful attempts. Include services and maintenance in running costs without counting human time twice. The €2,400 setup figure is fictional, not a Neaptide quote.

Releasing 35 hours does not automatically put €1,050 in the bank. A salaried employee may use that time for other work. That is additional capacity. To claim cash savings, identify expenses that actually disappear; to claim revenue growth, measure the change in revenue.

• Microsoft: connect agent usage with business outcomes

• Try the calculator with your own assumptions

At eleven minutes, running costs exceed the value of released time. Before buying a stronger model, investigate what makes review slow: poor formatting, missing sources or a task that is too broad. Measure the effect of fixing those problems too.

What to bring to a developer

Prepare a one-page brief naming the person responsible, the information the system will receive, the approved sources and the expected output. Explain what should happen when information is missing and which actions the system must never take. Add examples with identifying details removed, current task timings and a maximum budget for the trial.

Repeat the evaluation after changing the model, instructions or sources, and keep measuring rework. The first pilot should produce a defensible next decision. Finding that a simpler form and templates work better is a useful result.

• AI implementation and process automation at Neaptide