The build-vs-buy decision for a chatbot follows the same logic as any software decision — but with a chatbot-specific twist: the gap between a good platform chatbot and a bad custom one is much larger than the gap between a good and bad CRM.
The case for buying a chatbot platform
Platform chatbots (Intercom, Freshdesk, Zendesk AI, Drift) win in specific situations:
- Speed to market: A platform chatbot can be live in days. A custom build takes weeks to months.
- Proven infrastructure: The platform handles uptime, security, compliance certifications, and model updates. You don't.
- Predictable cost: Monthly SaaS pricing is easier to budget than engineering project costs plus ongoing infrastructure.
- Existing integrations: Major platforms integrate with Shopify, Salesforce, HubSpot, Zendesk, and most common tools out of the box.
Platform chatbots make sense when your support workflows are standard, your tech stack is mainstream, and your primary goal is reducing inbound contact volume rather than differentiating the customer experience.
The case for building a custom chatbot
A custom build wins when:
- Data sensitivity: Your knowledge base contains confidential product, pricing, or client information you cannot pipe through a third-party SaaS platform's AI model.
- Proprietary integrations: Your core systems (bespoke ERP, legacy order management, custom CRM) don't have connectors in any major platform.
- Differentiated experience: Your chatbot is a product feature, not a support tool. The conversation quality is a competitive differentiator, not just a cost-reduction measure.
- Platform pricing at scale: Many platforms charge per conversation or per seat. At high volume, custom infrastructure becomes cheaper within 12–24 months.
The decision matrix
| Factor | Buy wins | Build wins |
|---|---|---|
| Time to market | Under 4 weeks needed | 3+ months acceptable |
| Integrations needed | Standard tools (Shopify, SF, HubSpot) | Custom or legacy systems |
| Data sensitivity | Non-sensitive product content | Confidential or regulated data |
| Monthly conversation volume | Under 50,000 | Over 50,000 |
| Internal dev resource | No engineering team | In-house or outsourced dev |
| Differentiation need | Cost reduction only | Core product feature |
Total cost of ownership: a 24-month view
Most build-vs-buy analyses compare upfront cost only. The correct comparison is 24-month TCO:
Platform (10,000 conversations/month):
- Monthly licence: £400–£800
- Setup and configuration: £500–£2,000 (one-time)
- 24-month total: £10,100–£21,200
Custom build (same volume):
- Initial build: £8,000–£15,000
- Monthly hosting + API: £120–£300
- Maintenance: £300–£600/month
- 24-month total: £18,760–£38,200
At moderate volume, a platform often wins on pure cost. The build becomes economical above ~50,000 conversations/month, or when the platform's limitation forces you to build custom solutions on top of it anyway.
Use our Build vs Buy Decision Tool to run your specific numbers, and our Chatbot ROI Calculator to model the payback period before committing either way. See also: How Much Does a Business Chatbot Cost in 2026?