Customer support decision
Zendesk
A small team can build a basic ticketing + RAG-backed reply system, but Zendesk’s durable advantages — large proprietary interaction data, expansive integrations marketplace, and production-grade omnichannel/voice scale — make a full replacement impractical for most teams.
Visit website↗$19/mo
$228/yr
Per seat. Read off the official pricing page.
$100one-off80 h to build
$300/mo20 h/mo upkeep
On cash alone, building overtakes the subscription at 17 seats.
The code exists. It is not what you are paying for.
These 2 projects are real, published, and do the core job — and this page still says keep paying. What the subscription buys is integration maintenance, proprietary data and infrastructure at scale, and none of that ships in a repository. Fork one anyway if you want to. Go in knowing what it does not carry. What stays hard ↓ · All Zendesk alternatives, with the arithmetic →
What a replacement has to do
- Ingest customer messages into ticket records, provide an LLM-backed conversational assistant that suggests or returns answers, search a unified knowledge base (RAG), route tickets to agents, and basic analytics/dashboarding for tickets and agent performance.
What it still won’t have
- Proprietary training signal from Zendesk’s large installed base and interaction corpus
- Marketplace of 1,800+ prebuilt apps and deep third-party integrations
- Scale, reliability and operations for voice and omnichannel contact center at enterprise scale
- Enterprise-grade governance, sandboxing and advanced security features
- Prebuilt agentic AI features (self-improving AI agents, auto-triage, generative voice) out of the box
What remains hard
- Integration maintenance
Marketplace Find 1,800+ apps, partners, and integrations to serve customers and employees
- Proprietary data
Trillions of data points turned into billions of successful outcomes.
- Infrastructure at scale
22K+ AI customers 830M AI interactions 4.8B Resolutions delivered
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 17 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal self-hosted customer support platform in Node.js + Postgres + Redis + vector DB (Milvus or pgvector) with a React agent UI. Implement: (1) ticket ingestion endpoints (email/webhook/chat) storing tickets in Postgres; (2) knowledge-base ingestion and embedding pipeline that indexes documents into the vector DB; (3) an API that calls an external LLM for RAG responses and suggestion generation; (4) ticket routing rules (skill tags, round-robin) and an agent acceptance flow; (5) a simple reporting page (tickets/day, automation rate, CSAT input). Out of scope: telephony infrastructure, marketplace integrations, multi-tenant billing, and enterprise governance features. Include error handling, rate limiting, retry logic for LLM and vector DB calls, and automated tests for ingestion, RAG search, and reply generation.
How we checked
How the score was reached
- Pay verdict base20
- An open-source build was found+5
- 5 cited sources+3
- Price verified on pricing page+3
- 3/3 assessment runs agreed+4
- Hard moats found in the evidence-6
- Evidence score29
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 5
Every page the run actually retrieved.
- official productZendesk product
- official pricingZendesk pricing
- official docsZendesk platform
- open sourcechatwoot/chatwoot
- open sourcezammad/zammad
Integrity checks
What held up, and what did not.
