Should You Build an AI Chatbot or Buy One?

AI chatbots are having a moment, and the vendors know it. Before you sign up for another monthly subscription, or commission a custom build, it's worth being clear about what each path really gives you. The right answer depends less on the technology and more on what you need the bot to do.

What you get off the shelf

Buy-it chatbot platforms are quick to deploy and handle the plumbing for you. They shine when:

• Your needs are standard: answer FAQs, deflect common support tickets, capture basic leads.
• You want something live this week with minimal setup.
• You're fine working within the platform's templates and limits.

The catch: monthly fees that scale with usage or seats, limited control over behavior and tone, your data living in someone else's system, and a ceiling on how deeply it can integrate with your tools.

What a custom build gives you

A custom assistant costs more upfront but buys you control:

• Exact behavior and tone, it represents your brand, not a generic template.
• Deep integration with your own systems (your data, your CRM, your booking flow).
• Ownership of your data and no per-conversation tax that grows with success.
• Honest guardrails tuned to your business, what it should never say, when to hand off to a human.

The assistant on this site's home page is a custom example: it's grounded in how I actually think and work, hands serious prospects to a real conversation, and captures leads, none of which a generic widget would do well. (More on the principles in AI assistants that actually work.)

How to decide

Run it through three questions:

1. Is the job standard or specific? Standard FAQ deflection → buy. Something tied to your unique workflow or brand → build.
2. How central is it? A "nice to have" widget → buy the cheap option. A core part of how you sell or support → owning it pays off.
3. What's the integration depth? If it must plug into your systems and act on your data, custom usually wins; bolt-on widgets struggle here.

A sensible middle path

You don't have to commit blind. A common route: start with an off-the-shelf bot to learn what your visitors actually ask, then build a custom assistant once you know the real use cases and the volume justifies it. The questions people ask the cheap bot become the spec for the good one.

Trying to decide for your business? Tell me what you'd want the bot to do and I'll give you a straight build-or-buy recommendation, including "just buy the simple one" when that's the right call.