The number of businesses offering "AI solutions" has exploded. Some are excellent. Many are repackaging off-the-shelf tools at a significant markup and calling it bespoke. A few are outright selling capability they don't have.

If you're evaluating agencies or consultants for an automation project, these are the questions worth asking — and what the answers should tell you.

"Can you show me something you've built that's similar to what I need?"

The most reliable signal is evidence. Not case studies written in marketing language — actual examples. If they can walk you through a workflow they built, explain the decisions they made, and describe what happened when they deployed it, you're talking to someone who has actually done the work.

Vague references to "multiple enterprise clients" without specifics should make you cautious.

"What tools and platforms do you actually build on?"

There's a spectrum here. At one end, agencies build entirely on drag-and-drop platforms like Zapier or Make — legitimate tools, but limited in what they can handle and potentially expensive at scale. At the other end, everything is custom code, which offers more power but higher cost and longer timelines.

The right answer depends on your needs. What you want is an honest explanation of what they use and why, not a pitch about proprietary technology that's really just a branded Zapier account.

"How do you handle the handover and maintenance?"

A lot of automation projects deliver something that works on day one and breaks six months later when a connected tool updates its API or your process changes slightly. Ask specifically what happens after deployment. Do they document everything? Is there a support arrangement? Can your team maintain it without them if needed?

The answer tells you a lot about whether they're building for your long-term success or their next invoice.

"What does your scoping process look like?"

Reputable agencies invest time in understanding your processes before they propose anything. If you receive a quote within 24 hours of a first conversation with no detailed discovery, that's a quote based on assumptions — and you'll pay for those assumptions later in scope creep and change requests.

A proper scoping process should feel thorough, maybe even slightly slow. That's them doing it properly.

"What's the realistic timeline to see something working?"

This is a calibration question. If the answer is "six months minimum," that's either a complex enterprise project or an agency that over-engineers. If the answer is "we can have everything automated in two weeks," be sceptical.

For most B2B automation projects of moderate complexity, four to eight weeks from scoping to a working deployment is realistic. Simpler integrations can be faster; custom AI development takes longer.

One More Thing

The best agencies will ask you as many questions as you ask them. They want to understand your business, your constraints, and your definition of success before they tell you what they'd build. If a sales conversation feels one-sided — lots of talking about their capabilities, not much listening to your specifics — that's worth noting.

We're always happy to have a straight conversation about whether what you need is something we're right for. Sometimes it isn't — and we'll tell you that too.