Questions to ask before hiring an AI automation developer or agency

Before hiring an AI automation developer or agency, ask questions in five areas: whether they understand your problem, how they will build and test the solution, how they will handle your data and accounts, who owns the result, and what happens after launch. Good answers are specific and honest about limits. Vague promises, pressure to decide quickly, and a fixed price before anyone has looked at your process are warning signs.

Prepare before the first conversation

You will get far better answers if you bring a little homework:

• A short, step-by-step description of the process you want to change. The guide to mapping a workflow before automating it shows how.
• A few real examples with names and numbers removed: customer messages, invoices, reports.
• What a good result looks like: time saved, fewer errors, faster replies to customers.
• The tools you use now, such as your Facebook page, WhatsApp, Google Sheets, accounting software or website.
• The name of the person on your side who will answer questions during the project.

The questions to ask

About your problem

• "What do you need to know about our process before suggesting a solution?" A good developer asks you plenty of questions before proposing anything.
• "Is AI the right tool here, or would simpler automation do?" An honest developer will sometimes tell you that a rule-based workflow or a well-built spreadsheet is enough.
• "What could go wrong with this idea?" Listen for a thoughtful answer rather than reassurance.

About how they will build it

• "Can we start with a small pilot before the full project?"
• "How will you test it with our real examples, including the messy ones?"
• "Where does a person review or approve the output before it reaches a customer?"
• "What happens when the AI is unsure, or wrong?"
• "Can you show similar work and explain what was difficult about it?" It is also fair to ask whether you may speak to someone they have worked with before.

About data, security and access

• "Which of our data will the system see, and where will it be stored?"
• "Which AI providers will you use, and what do their terms say about our data?" They should explain in plain words and point you to the provider's own terms.
• "Whose accounts will the tools run under?" Ideally, accounts in your business's name, with the developer given access rather than owning them.
• "How will passwords and API keys be stored, and how is access removed when the project ends?"

About ownership and handover

• "Who owns the code, workflows and prompts when the project is finished?"
• "What documentation will we receive?"
• "If you become unavailable, could another developer take this over?" A handover checklist is a useful model, even for automation projects.

About cost and life after launch

• "Can you split the quote into phases, with what we receive at each stage?" The article on what affects the cost and timeline of AI projects helps you read the answer.
• "What running costs should we expect, and what makes them go up?"
• "What support do you offer after launch, and how quickly do you respond when something breaks?"
• "How will we measure whether it is working?"

Warning signs

• Promises of exact results or dramatic savings before they understand your process.
• No questions about your data, or a casual attitude to customer information.
• Everything set up on the developer's personal accounts.
• Unwillingness to explain how the system works in plain language.
• Pressure to sign today, or a large payment before any work has been shown.
• No plan for a human check on anything customers will see.

Good signs

• They ask more questions than they answer in the first meeting.
• They suggest starting small and proving the idea first.
• They are clear about what AI cannot do.
• They put scope, deliverables and exclusions in writing.
• They talk about your staff, not only the technology: who will use it, and how they will be trained.

After the conversations

Write down each candidate's answers side by side. Choose clarity and honesty over the lowest quote; a cheaper project that nobody on your team understands becomes expensive later. If you are still deciding what to automate first, read choosing your first AI automation project. And if you would like to put these questions to me about your own process, get in touch.