Ranges below are typical planning estimates for custom AI chatbot builds in 2026. They are starting points for budgeting — discovery is required before any fixed quote.
Типичные диапазоны стоимости
Диапазоны — только ориентиры для планирования. Итоговая цена зависит от scope, интеграций, дизайна и сроков — требуется discovery.
| Scope | Typical range | Notes |
|---|
| Guided FAQ / widget assistant | $8,000 – $25,000 | Curated prompts, limited sources, web embed, basic analytics. |
| Knowledge-base connected bot | $25,000 – $60,000 | Document ingestion, retrieval tuning, admin review, channel integrations. |
| Workflow agent with integrations | $60,000 – $150,000+ | CRM/helpdesk actions, role-aware access, monitoring, escalation paths. |
| Ongoing improvement retainer | $3,000 – $15,000 / month | Prompt ops, evals, content updates, and safety reviews as usage grows. |
AI chatbot pricing factors
Chatbot estimates track data, risk, and operational needs — not model hype alone:
- Source documents, freshness, and permission boundaries
- Languages, tone, and brand-safe response policies
- Channels (web, Slack, WhatsApp, etc.) and SSO requirements
- Human handoff, ticketing, and logging expectations
- Evaluation harnesses and regression testing for prompts
- Compliance considerations for PII and retention
Typical chatbot delivery timeline
FAQ-style assistants may reach pilot in roughly 4–8 weeks; retrieval-heavy bots with integrations often need 8–16 weeks including content preparation and review cycles.
Timelines depend on how quickly stakeholders approve knowledge sources and escalation rules — discovery surfaces those dependencies early.
Common chatbot architecture choices
DEWEB favors maintainable patterns with observable behavior:
- Retrieval-augmented generation over approved content stores
- Orchestration layers for tools, APIs, and structured workflows
- Vector or hybrid search depending on corpus size and update frequency
- Admin consoles for transcript review and content updates
- Rate limits, caching, and fallbacks when models or APIs fail
DEWEB process for AI assistants
We begin with use-case interviews, data inventory, and safety requirements — then prototype against real questions from your team or customers.
For broader automation, see our AI business automation services; chatbots are often one component of a larger workflow estimate.
Когда выбрать этот подход
When a lighter chatbot budget fits
- Answers come from a small, well-maintained knowledge base
- Human agents remain the primary resolution path for complex cases
- One channel (e.g., marketing site widget) is enough for phase one
When to plan for higher chatbot investment
- The bot must perform authenticated actions in CRM or billing systems
- Regulated industries require detailed logging, review, and access control
- Multiple languages or high-stakes advice domains need rigorous evals
Часто задаваемые вопросы
- How much does it cost to build an AI chatbot?
- Many knowledge-base chatbots fall in a typical range of $25,000–$60,000, with simpler FAQ widgets starting lower and integration-heavy agents costing more. Discovery defines data sources and guardrails that drive the estimate.
- Do chatbot costs include LLM usage fees?
- Build estimates usually cover engineering, integration, and launch readiness. Ongoing model and hosting usage is often billed separately based on traffic — DEWEB can outline expected ranges after architecture review.
- Can we start with a pilot before full rollout?
- Yes. Phased pilots with limited sources and channels are a common way to validate usefulness before expanding scope.
- How does DEWEB estimate AI chatbot work?
- We scope use cases, content readiness, integrations, and review workflows. Estimates are transparent assumptions — not promises about automation percentages or revenue lift.
Planning an AI assistant?
Describe your users, knowledge sources, and handoff rules to DEWEB. We will propose a phased pilot and share an estimate after discovery.
