Blog Details

AI chatbot pricing in Singapore falls into two broad models: subscription SaaS platforms starting around S$50 to S$500 a month, and custom-built chatbots typically costing S$8,000 to S$25,000 upfront plus ongoing hosting and maintenance. Which one makes sense depends on how much your chatbot needs to integrate with your existing systems and how specific your use case is.

Most SME owners searching for chatbot pricing are really trying to answer one question: do I need an off-the-shelf tool or a custom build? Below is how to tell the difference and what each option actually costs.

Subscription (SaaS) chatbot pricing

Subscription chatbot platforms let you configure a chatbot through a dashboard, usually with pre-built templates for FAQs, lead capture, or basic booking flows. Pricing is typically tiered by conversation volume or number of channels connected.

Tier	Typical monthly cost (SGD)	What you get
Basic	S$50 - S$150	Single channel, template flows, limited customisation
Growth	S$150 - S$500	Multi-channel (website + WhatsApp), some integrations, more volume
Advanced/Enterprise SaaS	S$500 - S$2,000+	Higher volume, more integrations, priority support

SaaS chatbots are the right call when your use case is genuinely simple: answering common FAQs, capturing a lead's contact details, or routing to a human. They're fast to set up (days, not weeks) and don't require development work. The tradeoff is limited flexibility: if your workflow needs custom logic (checking real-time inventory, pulling a customer's order history, handling a multi-step booking flow across systems), most SaaS platforms will hit a wall.

Custom-built chatbot pricing

A custom-built chatbot is designed around your specific business logic and integrated directly with your systems (CRM, booking software, inventory, WhatsApp Business API). This is where most of Freemansland's conversational AI agent development work happens.

Complexity	Typical build cost (SGD)	Example use case
Simple custom bot	S$8,000 - S$12,000	FAQ + lead capture with one system integration (e.g. CRM)
Mid-complexity	S$12,000 - S$20,000	Booking/reservation flow with calendar and payment integration
Complex/multi-system	S$20,000 - S$35,000+	Multi-step workflows across CRM, inventory, WhatsApp, and human handoff

On top of the build cost, budget for ongoing hosting (often S$50 to S$300 a month depending on volume), API usage costs if the bot uses a language model API, and any WhatsApp Business API messaging fees if that's your channel. Most custom builds also include an initial support and tuning window (commonly 30 to 90 days) to fix issues that only show up once real customers are using it.

What actually drives custom chatbot cost up or down

• Number of system integrations. Each additional system (CRM, POS, booking software) adds integration and testing work.
• Conversation complexity. A bot that just answers FAQs is simpler than one handling multi-turn bookings with confirmations and edge cases (cancellations, changes, no-shows).
• Multilingual support. Adding Chinese or Malay alongside English increases testing scope. See our guide on multilingual AI chatbots.
• Human handoff logic. Deciding when and how the bot escalates to a human agent adds design and testing complexity.
• Channel count. Website widget only is simpler than website plus WhatsApp plus Instagram DM.

SaaS vs custom: which is right for your business?

Factor	SaaS platform	Custom build
Setup time	Days	3-8 weeks typically
Upfront cost	Low to none	S$8,000+
Ongoing cost	Monthly subscription	Hosting + API + maintenance
Customisation	Limited to platform's features	Built around your exact workflow
System integration	Only what the platform supports natively	Whatever your systems allow via API
Ownership	You rent access to the platform	You typically own the build (confirm contractually)

A useful rule of thumb: if you can sketch your ideal chatbot flow and it fits entirely within a SaaS platform's template options, use the SaaS platform and save the budget. If your flow requires the bot to "know" something specific about your business (real-time stock, a customer's booking history, a quote calculation), you're looking at a custom build. For the broader distinction, see chatbot vs conversational AI agent.

Can grants offset chatbot costs?

Some SMEs use grant schemes like PSG (for pre-approved solutions) or EDG (for larger custom projects) to offset chatbot costs, potentially covering up to a percentage of qualifying costs for eligible projects. This requires pre-approval before starting work and is never guaranteed. See our detailed guide on how Singapore SMEs offset AI chatbot costs with grants.

Hidden costs to ask about upfront

• WhatsApp Business API messaging fees, which are billed per conversation by Meta and separate from your chatbot vendor's fee
• Language model API usage costs if the bot uses a generative AI backend (these scale with conversation volume)
• Hosting and uptime monitoring
• Cost of retraining or updating the bot's knowledge base as your products, services, or policies change
• What happens if you want to switch vendors: can you export your conversation flows and data, or are you locked in?

A realistic example

A clinic wanting a chatbot to handle appointment booking, common FAQs, and a handoff to reception for complex cases would typically be looking at a mid-complexity custom build, in the S$12,000 to S$20,000 range, plus WhatsApp integration if that's a preferred channel. A retail shop that just wants FAQ answers and lead capture on their website might get by entirely on a S$150/month SaaS plan. These are illustrative ranges, not quotes, since actual pricing depends on your specific requirements.

What the build process actually looks like for a custom chatbot

Once you commit to a custom build, the typical process runs through a few distinct stages. First, mapping the actual conversations: what customers ask, how staff currently answer, and where the bot should hand off to a human. This usually involves reviewing real chat logs, call transcripts, or WhatsApp history if you have them, since guessing at conversation flows without real examples tends to produce a bot that sounds fine in testing but misses common real-world phrasing.

Next comes the technical build: connecting to your systems, writing the logic for how the bot responds and when it escalates, and setting up the underlying language model configuration. Testing follows, usually with a mix of scripted test cases and, ideally, a soft launch to a small subset of real customers before full rollout. Expect at least one or two rounds of adjustment after the soft launch, since real conversations always surface phrasing and edge cases that internal testing missed.

Why per-conversation and per-message pricing matters

Some SaaS chatbot platforms price by conversation volume rather than a flat monthly fee, which can catch growing businesses off guard if usage scales faster than expected. Before committing to a platform, ask what happens if you exceed your plan's included volume: does it auto-upgrade at a higher price, get throttled, or bill overages per conversation? For custom builds using a language model API directly, usage costs scale with conversation volume and length, so a bot handling 500 conversations a month costs meaningfully less to run than one handling 5,000, even though the build cost is the same. Factor this into your ongoing budget, not just the upfront build price.

Maintenance: the cost most SMEs underestimate

A chatbot isn't a "build once, forget forever" system. Your products change, your policies change, your promotions change, and the bot's knowledge needs updating to match. Some vendors include a certain number of update hours per month in an ongoing support arrangement; others charge per update request. Ask specifically how content updates work: can your own team update FAQs and pricing information through a simple dashboard, or does every change require going back to the vendor? The former is usually worth paying slightly more for upfront, since it avoids a recurring dependency for routine changes.

How to test a chatbot before committing to full pricing

If you're unsure which model fits, a low-cost way to test the waters is starting with a SaaS platform's free or lowest tier to validate that customers actually engage with a chatbot at all before investing in a custom build. Some businesses discover their customers strongly prefer speaking to a human even for simple questions, which is useful (if humbling) information to have before spending S$15,000 on a custom bot nobody uses. If the SaaS trial shows strong engagement but keeps hitting the platform's functional limits, that's a good, evidence-based signal that a custom build is the right next step, rather than guessing upfront.

What "good" chatbot performance actually looks like

It's worth defining success metrics before launch regardless of which pricing model you choose. Common metrics include: percentage of conversations resolved without human handoff, customer satisfaction on chatbot interactions specifically (a short post-chat rating works well), and reduction in repetitive enquiries reaching your human team. A chatbot that resolves 60-70% of routine enquiries without escalation is generally considered solid performance for a well-scoped SME bot; expecting 100% automation from day one is unrealistic and usually a sign of an overly ambitious initial scope rather than a technology shortfall.

Ready to see what AI can do for your business?

Not sure whether your use case needs a SaaS tool or a custom build? That's exactly what a scoping conversation is for. Freemansland's conversational AI agent development service covers everything from simple FAQ bots to multi-system booking flows. Request a quote or contact us to talk through your specific use case. WhatsApp +65 9184 9908 or email glenn@freemansland.co.

Frequently Asked Questions