An AI chatbot for a Singapore business usually breaks into three cost layers: the messaging channel, the software that runs the bot, and the build itself. If you rent everything month to month through an off-the-shelf platform, expect roughly SGD 100 to 1,000+ per month, forever. With us there are two ways in: a subscription plan from S$38 a month with no build fee, or a bespoke build from S$4,500 that runs on an Artio Care plan from S$300 a month, so it stays monitored, current, and online. The return is not magic. It comes from two dull, real things: replying faster than you can by hand, and not losing leads while you sleep. Whether that pencils out depends entirely on how many enquiries you get and what a customer is worth to you.
That is the whole answer. The rest of this is me showing my working, because the number you get quoted elsewhere often hides where the money really goes.
The three cost layers, separated
Most quotes blur these together on purpose. Pull them apart and the picture gets clearer.
1. The messaging channel. If you run on WhatsApp, you are on Meta's WhatsApp Business Cloud API. Meta gives you the first 1,000 conversations a month free, then meters after that. For a lot of small businesses, free tier covers the whole month. This is a Meta cost, not a vendor markup, and it is the same whoever builds your bot. The same agent can also sit on Instagram DMs and Facebook Messenger, so one brain answers all three inboxes.
2. The software that runs the bot. This is where the models diverge sharply.
- SaaS chat platforms charge a monthly subscription to host the logic, usually SGD 100 to 1,000+ per month depending on seats, contacts, and features. You are renting. Stop paying and the bot goes dark.
- A build-you-own is a one-time build that lives on your own Meta account and infrastructure. You are not paying rent on the software month after month.
3. The build. Someone has to design the conversation, connect it to your booking or enquiry flow, and test it against the way your customers actually type. With us that is from S$4,500 for a bespoke build, and every live build runs on an Artio Care plan from S$300 a month: monitoring, uptime, Meta and OpenAI change-management, and break-fix, so the bot you paid for keeps working. If you do not need a bespoke build at all, our subscription plans start at S$38 a month with no build fee.
A fuller walk-through of how the WhatsApp side works in practice is in our WhatsApp automation playbook, and the full service is on AI automation.
Rent versus own, over two years
Here is illustrative maths, not a claimed result. Numbers are yours to swap.
Say a SaaS platform costs you SGD 300 a month. Over 24 months that is SGD 7,200, and you own nothing at the end. A bespoke build at S$4,500 plus 12 months of Essential care at S$300 comes to S$8,100 in year one, then S$3,600 a year of care plus messaging after. The rented option looks cheaper early. The build is the asset: designed around your business, with a team keeping it monitored and current rather than a seat licence you rent.
Which wins depends on how long you will run it and how bespoke it needs to be. If you are testing an idea for three months, rent, or start on one of our plans from S$38 a month. If this is going to be part of how your business answers customers for years, a build usually makes more sense. Neither is "the right answer" in the abstract.
How to think about the return, honestly
The return on a chatbot is not a percentage I can hand you. It is a calculation you do with your own two inputs.
Time saved on first response. Add up the hours you or your staff spend typing the same first replies. Opening hours, price ranges, "are you open on Sunday", "do you take walk-ins". If a bot handles those instantly, that time comes back. Value it at whatever an hour of your team costs.
Leads not lost to slow replies. This is the bigger one and the harder one to see, because you never find out about the customer who messaged at 11pm, got nothing, and booked with someone else by morning. A chatbot answers at 11pm.
Illustrative only, plug in your real numbers: if you get 60 enquiries a month, and even a handful of them currently slip away because nobody replied fast enough, and a customer is worth a few hundred dollars to you, then recovering just those few pays for a build in a matter of months. If you get 5 enquiries a month, that same maths does not work, and you should not build this yet. Be honest about your volume.
When it is not worth it
I would rather tell you to wait than sell you something that sits idle.
- Low enquiry volume. If you get a trickle of messages you can comfortably answer yourself, a bot is a solved problem you do not have. Handle it by hand.
- Every enquiry is genuinely bespoke. If nothing repeats and every conversation is a real negotiation, there is little for a bot to automate well. It will frustrate people.
- You have not written down your own answers yet. A chatbot can only be as good as the FAQs, prices, and policies you feed it. If those live only in your head, sort that first. That work is useful even if you never build a bot.
- You are chasing hype. "Everyone has AI now" is not a business reason. Volume and lost leads are.
Where to start
Start by counting. One month of enquiries, roughly how many, roughly how many slip, roughly what a customer is worth. Bring those three numbers and the maths above tells you most of what you need to know before anyone quotes you.
When you are ready, every engagement with us starts with a free consultation and a fixed quote, so you see the real number before you commit to anything. We are Singapore-based, in Paya Lebar. If the honest answer is "not yet", we will tell you that too. You can reach us any time on the contact page, and if you want to compare this against the cost of the website it sits on, we broke that down in how much a website costs in Singapore.
, Zyle