Build Or Hire A Chatbot? One Question And 8 Checks

Free checklistFrom the video

Build your chatbot, or hire someone? One question, 8 checks.

There are thousands of YouTube videos promising to teach you how to build a chatbot for your business in 30 minutes. Many of those creators genuinely know what they're doing. But here's what almost none of them tell you.

That person makes it look easy because they are a natural tinkerer: someone who enjoys reading documentation at midnight and figuring out why something broke. If you're a restaurant, salon or gym owner who is brilliant at what you do and simply wants to get on board with AI, that's a completely different starting point. The gap between the two is where the costly mistakes live.

I'm not here to scare you away from chatbots. I want you to implement something, without wasting months or rebuilding from scratch because nobody walked you through the decision first.

The door handleThe thirty-second look behind the plate

A while back I had a door handle that wasn't working properly. I looked it up online, watched a video, decided I needed a new handle, and was about to order one. Then a handyman came by for something else, looked at it for about thirty seconds, reached behind the plate, adjusted one small thing, and it worked perfectly. No new parts. No money spent. Just knowing where to look.

I think about that door handle constantly when I watch business owners approach chatbots. They find a tutorial, pick a platform and start building. Somewhere between week two and month three, they discover the real problem. The one hiding behind the plate the whole time.

Trap 1: DIYThe tutorial teaches A chatbot, not THE RIGHT one

A tutorial won't teach you to build the chatbot for YOUR business, connected to your tools and your customer journeys. And it was probably made by someone who has built many of these. Every shortcut they take is a wall they already hit. You haven't hit those walls yet. You're about to.

DIY bots fail for predictable reasons: poor training, no fallback when the bot doesn't know the answer, trying to make one simple bot do everything over time, and a bot that doesn't connect to anything else you use.

Then there's the platform problem. Most tutorial creators recommend the platform THEY know. Pick the wrong one and six months later you find it only handles WhatsApp: no Instagram, no Facebook, no website chat, no voice. Now you're rebuilding everything. And the platform will change underneath you, so your system can break quietly until a customer gets a wrong answer.

Trap 2: HireHiring the wrong builder costs just as much

If that made you think "I'll just hire someone", slow down. Hiring the wrong person or agency is just as expensive, in a different direction.

Most providers build with whatever is convenient for THEM, and that includes agencies like mine. That isn't dishonesty, it's reality. So the burden is partly on you to ask the right questions up front, and an honest agency will answer them before you sign anything.

First, the one questionDo you want control, or convenience?

This is the framework I use for decisions like this. Marketing and systems are my thing, so there I choose control: I write my own video scripts, even though tools could do it in seconds. Accounting? Plumbing? The door handle? Convenience.

So ask yourself honestly: is building and maintaining a chatbot an area where you, or someone close to your business, wants CONTROL? A team member, a partner or a relative who is a tinkerer AND a marketer, and will put hours in, not just now but ongoing. If that person exists, DIY has a real case. Not the 30-minute version, the ongoing-commitment version. If not, the question isn't whether to hire. It's how to hire smartly.

The 8 checks4 questions to ask a builder, and 4 things true either way

• 1Ask: what happens in six months when I need to change something?

  Your business WILL change. That's not a risk, it's a certainty. The technology is changing too. So ask how the system handles change, and who does it.

  WhyA system that can't change with you becomes the thing you rebuild.

• 2Ask: does it handle every channel my customers use, or just one?

  WhatsApp, Instagram, Facebook, the chat on your website, maybe voice later. Ask which work today, and which can be added without starting again.

  WhyIf it only does one channel, adding the next one means rebuilding everything.

• 3Ask: can it connect to my other systems, or will some connections never be possible?

  Some point-of-sale systems are open to outside software: it may take investment, but you know the path exists. Others are completely closed. If yours is closed, only the people who make it can open it.

  WhyIf a system is closed, no amount of skill opens it from the outside. Better to find out before you commit. And if a builder promises you everything is possible, that is information too.

• 4Ask: is this a one-and-done delivery, or an ongoing relationship?

  Someone in your business, or alongside it, has to own this. Not as a project. As a permanent job.

  WhyA chatbot that is built and then abandoned starts failing quietly, invisibly, the moment your business changes.

• 5You treat version one as a starting point, not a finished product.

  What you launch is a Minimum Viable Product: the start of a journey, not the end of a project. There are always situations the system hasn't met yet, and fixes that only show up once real customers use it.

  WhyAnyone who promises you a finished, set-and-forget chatbot is either inexperienced or not being honest with you.

• 6It's a team of specialist bots, not one bot doing everything.

  I learned this the painful way. My first serious bot was built to know everything. Then it got confused, pulling from the wrong knowledge at the wrong moment, and I had to rebuild it. What works is a boss bot that routes the conversation, with specialists underneath: coupons, bookings, orders, complaints. On the front, it still feels like one person.

  WhyYou wouldn't hire one person to run your marketing, customer service, bookings and technical questions at once. Specialists beat generalists.

• 7Someone retrains the bot every time the business changes.

  Think of your chatbot like a new employee. You launch a product, change a process or run a campaign, and suddenly it's giving customers wrong information. Retraining a bot is fast: update the knowledge, test, deploy. But someone still has to do it. Every time.

  WhyBots give wrong answers not because they're broken, but because nobody briefed them.

• 8Whoever builds it has experience you can't shortcut.

  What compounds is judgement: which setup fits which business, which platform has which ceiling, and which mistakes are coming before they arrive. If the person you hire doesn't understand checks 5 to 8, keep looking.

  WhyYou can start building experience today. But it doesn't download overnight.

If budget decidesA basic bot is better than nothing

If budget really decides it, a basic FAQ bot that answers your common questions, captures a contact and points people in the right direction has real value. Don't let the perfect be the enemy of the good. Just know it's a starting point, not a system. And add up the real cost of doing it yourself, including your time, the learning curve, the rebuilds and the maintenance. The numbers often surprise people.

Next: 9 Guardrails Every AI Chatbot Needs Before Launch and Before You Buy a Chatbot, Ask These 4 Questions. Control or convenience has its own video.