Fini
RAGless customer-support AI agent that automates refunds, KYC, and plan changes.
Fini is a RAGless customer-support AI agent that automates refunds, KYC, and plan changes, confirmed across 2026 support-tooling roundups. It prices around $0.69 per resolution.
Vendor details
Canonical URL
https://fini.ai
Category
Customer support agent
Subcategory
Support — AI agent
Use cases & customers
Primary use cases
Target customers
Deployment options
In practice
Your support team spends its day on refunds, KYC checks, and plan changes. Fini automates those actions directly, so routine requests resolve without a rep.
A customer wants a refund now, not a ticket number. Fini's agent processes the refund itself, closing the loop in the conversation.
Account changes like a plan upgrade bounce between support and billing. Fini handles plan changes as an action, so the customer's request gets done in one place.
Agentic Index coverage score
8.5 / 14 capabilities · 61%
| Integrations & Tool CallingLayers onto existing helpdesks including Zendesk, Intercom, Freshdesk and Salesforce rather than replacing them, and executes real actions in connected payment systems including Stripe and Adyen; corroborated by two independent competitor reviews, though no full integration catalogue was retrieved, Fini product pages and independent 2026 reviews 2026-07-30 | Full |
|---|---|
| Workflow OrchestrationSophie connects knowledge modules, action modules and an LLM supervisor into an orchestration layer that plans through a request, checks its logic against company policy and resolves multi step issues end to end, executing refunds, cancellations and account changes within defined business rules, Fini platform and product documentation 2026-07-30 | Full |
| Knowledge Grounding & RAGGrounds on the existing knowledge base and past tickets from day one, but via a REASONING FIRST, EXPLICITLY RAGLESS architecture: rather than fetching the nearest knowledge snippet and generating from it, the engine follows the company's stated rules and policies and traces every decision back to approved internal knowledge, Fini platform documentation 2026-07-30 | Full |
| Human Oversight & GuardrailsGuardrails enforce safety checks in real time, company specific rules are applied throughout the conversation, actions are bounded by defined business rules, and CONFIDENCE BASED ESCALATION is built specifically for high risk scenarios where acting incorrectly carries regulatory or financial consequences, handing off to a human with full context, Fini platform documentation 2026-07-30 | Full |
| Security, Identity & GovernanceThe broadest compliance stack in the customer support lane: SOC 2 Type II, ISO 27001, ISO 42001, GDPR, PCI DSS LEVEL 1 and HIPAA, covering payments, health data and AI management in one set; PCI DSS Level 1 is the highest transaction volume tier and allows cardholder data to be handled inside support interactions, and PII Shield applies always on redaction before data reaches the model, Fini compliance documentation 2026-07-30 | Full |
| Observability & AuditabilityA traceability layer audits every interaction and each response carries a traceable reasoning path pointing to the specific policies and decision logic used, producing an AUDIT READY EXPLANATION FOR EVERY ACTION THAT IS EXPORTABLE in formats meeting SOC 2 and PCI DSS examination standards; exportable examination ready audit artefacts are unique in this lane, Fini platform documentation 2026-07-30 | Full |
| Memory & State PersistencePast tickets are used as a grounding source and the agent is described as self improving, but both are retrieval and configuration refinement rather than a persistent memory layer, and no cross conversation customer state store was documented on the pages reviewed, Fini platform documentation 2026-07-30 | Unable to verify |
| Deployment & Data ResidencyNo deployment model choice, region selection or data residency commitment was documented on the pages reviewed, which is notable given PCI DSS Level 1 and HIPAA scope; the trust and DPA pages were not retrieved and are the re verification target for this cell, Fini platform documentation 2026-07-30 | Unable to verify |
| Prebuilt Agents, Templates & PacksSophie ships as a single named prebuilt agent covering high volume support journeys out of the box including cancellations, refunds and account changes, and goes live on day one against an existing knowledge base, but no template library, agent catalogue or marketplace was documented, Fini product pages 2026-07-30 | Partial |
| Triggers & Channel CoverageResolves across chat, email and voice with voice native to the platform rather than integrated through a third party, operating in over 50 languages, Fini product pages and voice agent documentation 2026-07-30 | Full |
| Model Flexibility & RoutingAn LLM supervisor sits in the architecture but the underlying model is not named and no model selection, routing or bring your own model capability was documented on the pages reviewed, Fini platform documentation 2026-07-30 | Unable to verify |
| APIs, SDKs & MCP ExtensibilityAction modules connect outward into helpdesks and payment systems, implying a programmatic surface, but no public API reference, SDK or MCP implementation exposed by Fini for others to build against was retrieved on the pages reviewed, Fini platform documentation 2026-07-30 | Partial |
| Testing, Debugging & OptimizationExtensive internal benchmarking is published and the agent is described as self improving, but no customer facing test environment, simulation harness or scored evaluation suite for validating agent changes before release was documented; the published accuracy and resolution figures are vendor reported and unverified, Fini published research and product pages 2026-07-30 | Partial |
| Browser & Computer UseNo browser control or computer use capability documented; the agent acts through connected helpdesk and payment system APIs rather than by operating a browser or a human interface, Fini platform documentation 2026-07-30 | Unable to verify |
The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Usage-based · from $0.69/resolution
resolutions
Included quota
Growth ($1,799/mo minimum): ~2,600 resolutions covered by the minimum at $0.69 each, agent actions (e.g., refunds via Stripe/Adyen), 20+ helpdesk/KB/CRM integrations, and PII Shield redaction. Free Starter is capped (~50 questions/mo, older model).
What is public
Fini (YC-backed; its agent is 'Sophie') uses resolution-based pricing. Published structure: a free Starter tier (severely limited — ~50 questions/mo, older model), then Growth at $0.69/resolution with a $1,799/mo minimum (the minimum covers ~2,600 resolutions), and custom Enterprise. There is no mid-tier between free and the $1,799 Growth floor.
Billing mechanics
Outcome-based: you pay $0.69 per AI-resolved ticket on Growth, billed against a $1,799/mo minimum (you pay the full minimum even at low volume). No per-seat fee for the AI agent — cost tracks resolutions, so as the resolution rate climbs, total spend rises. Enterprise is custom with the full compliance stack.
Cost watchouts
$1,799/mo Growth minimum regardless of volume, no mid-tier (steep free-to-Growth cliff), spend rises as resolution rate improves, headline 98% accuracy / 80%+ resolution claims rely on Fini's own data, Enterprise pricing custom
Variable cost rationale
Pure outcome-based above a high fixed minimum — spend scales directly with resolution volume, and the better the AI performs the more you pay, making budgets volume-sensitive
Additional watchouts
The $0.69 headline is real only at high volume — the $1,799/mo minimum dominates cost for smaller teams; outcome-based billing means cost grows with success
Overage / add-ons
Above the minimum, every resolution bills at $0.69; below it you still pay the full $1,799 (so the effective per-resolution cost is far higher at low volume — e.g., ~$3.60 at 500 tickets/mo).
Sales call required
Mixed (some tiers require a call)
Lowest paid plan
~$0.69/resolution
Commercial notes
YC-backed; agent takes actions (refunds via Stripe/Adyen); reasoning-first architecture; claims 98% accuracy / zero hallucinations and 2M+ queries processed (vendor data); ~$0.69/resolution is the lowest published rate among major outcome-priced rivals
Key ambiguities
How 'resolution' is defined/counted drives the bill; the $1,799 minimum and exact Enterprise terms aren't always prominent on the official site; performance claims are vendor-reported
Cancellation / refund
Growth carries a $1,799/mo minimum commitment; Enterprise terms negotiated; refund/cancellation terms not publicly detailed
Support SLA / resale
Enterprise adds SLA-grade compliance (SOC 2 Type II, ISO 27001/42001, GDPR, PCI-DSS Level 1, HIPAA), always-on PII Shield, and white-glove onboarding (~48-hour deployment)
Missing data
Exact Enterprise pricing, the resolution-counting definition, and any volume discounts aren't fully public; the $1,799 minimum is the key cost fact often omitted from the headline rate. Seed 'from $0.69/resolution' is accurate but omits the $1,799/mo Growth minimum.
Related vendors
- Decagon — AI agent platform purpose-built for customer support, with deep…
- 5.Y — Early stage Singapore platform (GLUCOSE) deploying goal driven AI…
- Ada — Agentic customer experience platform whose AI agents autonomously…
- Agency AI — Agentic AI for customer success from Drift co-founder Elias Torres,…
- Aisera — Enterprise agentic AI platform that orchestrates specialized agents…
- ASAPP — Generative AI for enterprise contact centers, with human-in-the-loop…
Alternatives to Fini
The closest documented capability profiles to Fini among customer support agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- KODIF9.5 / 14Adds documented Memory & State Persistence and Deployment & Data Residency
- Zingtree9.5 / 14Adds documented Memory & State Persistence and Deployment & Data Residency
- Assembled10.0 / 14Adds documented Memory & State Persistence
- Gorgias9.0 / 14Fuller documented coverage on Prebuilt Agents, Templates & Packs and APIs, SDKs & MCP Extensibility
- Nurix AI8.0 / 14Adds documented Memory & State Persistence
- Parahelp9.0 / 14Fuller documented coverage on Prebuilt Agents, Templates & Packs and Testing, Debugging & Optimization
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded