Decagon Review 2026: Who It Fits, What You Can Verify
decagon.ai · Benchmark rank #4 of 20 · Verified November 2026
Decagon is a sales-led enterprise AI concierge for chat, voice, and email, configured through Agent Operating Procedures (AOPs). Official materials document two commercial models, per conversation or per resolution, but Decagon has no pricing page. In RTCI's November 2026 test it reached an 80% realistic resolution rate. RTCI estimates ~$1.50 per resolution from reported per-conversation pricing: about $1.20 per conversation including the platform fee, divided by 80%. That gives an RTCI score of 682, Strong tier, #4 of 20. Shortlist it if your team can own AOPs and negotiate measurable terms.
Watch Decagon's official live demo of Voice 2.0 from its Decagon Dialogues '25 event.
Realistic ROAR
80% (RTCI test, Nov 2026)
Cost / resolution
~$1.50 est. (~$1.20 per conversation incl. platform fee ÷ 80%)
Min commitment
Enterprise contract
Pricing model
Per conversation or per resolution
Sources · verified November 2026
Vendor claim: 50–80% (named case studies)
RTCI verdict
Decagon holds an RTCI score of 682, Strong tier, #4 of 20. The score is 250 × log10(10 × realistic resolution rate ÷ cost per resolution), on a 0–1,000 scale: 250 × log10(10 × 80 ÷ 1.50) = 682 The 80% comes from RTCI's November 2026 test; Decagon's named case studies report 50% to 80%. The ~$1.50 is an estimate built from a competitor's reported $0.99 per conversation and ~$50,000-a-year platform fee at 20,000 conversations a month. Good fit: high-volume enterprises that want CX-owned procedures across chat, voice, and email. Poor fit: teams that need a public rate card or a self-serve rollout. Transparency is rated 1 of 5, and your quote will replace RTCI's cost estimate.
Strengths
- RTCI score of 682, Strong tier, #4 of 20, on an 80% realistic resolution rate from RTCI's November 2026 test.
- Official product pages document Agent Operating Procedures as a natural-language control layer for enterprise workflows.
- Chat, voice, and email are presented as one agent platform rather than disconnected channel bots.
- Decagon offers both per-conversation and per-resolution billing, and says larger committed resolution volumes lower the per-resolution rate.
- Named case studies support the test result: Chime reports consistently 70%+ chat resolution and nearly 70% on voice; Valon reports 50%+ resolution on voice.
Trade-offs
- No pricing page or public unit price for either commercial model. RTCI's ~$1.50 per resolution is estimated from a competitor's reported per-conversation rate and platform fee.
- The platform fee makes cost volume-sensitive: at 5,000 conversations a month the same inputs give about $2.28 per resolution.
- Decagon's own percentages mix resolution (Chime, Valon) and deflection (Duolingo's 80% deflection rate) and do not state how they were measured.
- Sales-led implementation requires workflow owners, integration work, evaluation data, and negotiated audit rights.
Ideal for
High-volume enterprises that already have a helpdesk system of record, need an independent omnichannel agent layer, can assign named AOP owners, and can enforce verified-resolution, escalation, audit, and billing definitions in the contract.
Not ideal for
Small and mid-market teams seeking published list pricing, a helpdesk replacement or a self-serve proof of concept. At low volume the platform fee dominates cost per resolution. Also a poor fit when nobody owns procedure changes and transcript review.
Deep dive
How Decagon actually works
Product shape is verifiable. Decagon’s official AOP material describes natural-language procedures, traces, and guardrails for enterprise workflows, while current product material positions chat, voice, and email inside one AI concierge layer for 'the world's leading enterprises'. These facts support a shortlist; they do not establish accuracy, time to value, or economics for a buyer’s workload.
Decagon’s pricing article (December 2024) describes a fixed rate per incoming conversation and a higher fixed rate per fully resolved conversation with no charge for escalations, says most customers choose per-conversation, and notes that larger committed resolution volumes lower the per-resolution rate. It publishes no dollar rate, minimum, implementation fee, overage rule, or standard contract definition, and RTCI found no pricing changes since July 2026. RTCI's estimate uses Quiq's reported figures (Quiq is a Decagon competitor): about $0.99 per conversation plus about $50,000 a year in platform fees. At 20,000 conversations a month that is ($4,167 + $19,800) ÷ 20,000 = about $1.20 per conversation. Divided by the 80% resolution rate from RTCI's November 2026 test, it is ~$1.50 per resolution, and 250 × log10(10 × 80 ÷ 1.50) = 682 Implementation fees are excluded.
Volume moves the number. At 5,000 conversations a month the platform fee adds about $0.83 per conversation, and cost rises to about $2.28 per resolution. At 100,000 a month, with the 30% discount a Vendr buyer note reports for 1M+ conversations ($0.693 per conversation), it falls to about $0.92. Vendr lists a median annual Decagon contract of $432,750.
Economics
What the Decagon AI line costs at three volumes, using RTCI’s realistic resolution rate of 80%. Seats and add-ons are listed separately below.
No public rate card. These figures are estimates to replace with your proposal.
| Conversations / mo | Resolved | AI cost / mo | Per resolution |
|---|---|---|---|
| 1,000 | 800 | $5,157 | $6.45 |
| 5,000 | 4,000 | $9,117 | $2.28 |
| 20,000 | 16,000 | $23,967 | $1.50 |
$0.99 per conversation, billed whether or not it resolves + $4,167 a month platform fee ÷ conversations resolved at 80%.
Inputs and where they come from
- Third-party
Platform or committed-spend floor
About $50,000 a year
- Third-party
Fee per conversation
About $0.99
- Quote only
Per-resolution option
Higher rate, no charge for escalations
Not in the AI cost
- Implementation and integrations
- Channel and overage terms
Decagon has no pricing page; every figure here should be replaced by the proposal.
Pricing breakdown
Layer 1 is the contracted platform or committed-spend floor, reported at about $50,000 a year by Quiq. Layer 2 is the selected usage unit: a fixed rate per conversation (reported at about $0.99), or a higher fixed rate per resolved conversation with no charge for escalations. Layer 3 is implementation, integrations, channels, overage, quality review, and residual human handling, all set in the contract. RTCI's ~$1.50 per resolution, as of November 2026, combines layers 1 and 2 at 20,000 conversations a month and an 80% resolution rate. Replace each input with the proposal's figures and reconcile billed units with human-judged outcomes.
Layer 1 · Fixed
Enterprise contract
Layer 2 · Variable
~$1.50 est. (~$1.20 per conversation incl. platform fee ÷ 80%)
Layer 3 · Model
Per conversation or per resolution
Evidence
Verified capabilities and evidence boundary
Official pages rechecked November 5, 2026; cost inputs checked November 10, 2026. Product facts do not equal measured buyer outcomes.
| Item | Public evidence | RTCI judgment | Confidence |
|---|---|---|---|
| Agent Operating Procedures | Official AOP page describes natural-language procedures, traces, and guardrails | Verified product shape | High |
| Channels | Official positioning names chat, voice, and email | Verify parity in the purchased scope | Med |
| Commercial models | Official pricing article describes per-conversation and per-resolution | Mechanics verified; dollar rates not published | High |
| Unit prices / minimum | No pricing page (decagon.ai/pricing returns 404); Quiq reports ~$0.99/conversation and ~$50,000/yr platform fee | Estimated ~$1.50 per resolution at 20,000 conversations/month | Low |
| Resolution performance | RTCI's November 2026 test: 80%. Named case studies: Chime 70%, Valon 50%+ voice, Duolingo 80% deflection | 80% realistic resolution rate; RTCI score 682 (Strong) | Med |
| Implementation time | No standard public timeline verified | Proposal-specific | Low |
Evidence
Sources checked
Dated source ledger for the facts used in this review (AOP page checked October 20, 2026; official pages November 5, 2026; Quiq and Vendr November 10, 2026).
| Source | What it supports | What it does not prove | Confidence |
|---|---|---|---|
| decagon.ai/product/aop | AOP product shape, natural-language procedures, traces, guardrails | Buyer accuracy or implementation effort | High |
| decagon.ai/resources/pricing-ai-agent | Per-conversation and per-resolution models; majority per-conversation | Unit dollars, minimum, or buyer CPVR | High |
| decagon.ai/case-studies/chime | Chime 70%+ chat and nearly 70% voice resolution | Company-wide resolution rate | Med |
| decagon.ai | Chat, voice, email positioning; Valon, Duolingo and Rippling results | Company-wide resolution rate | Med |
| decagon.ai/pricing | Returns 404; no public rate card | The terms of a private proposal | High |
| quiq.com/blog/decagon-pricing | Reported ~$0.99 per conversation and ~$50,000+/yr platform fee (competitor estimate) | Vendor-confirmed pricing | Low |
| vendr.com/marketplace/decagon-ai | Median annual contract $432,750; buyer note of 30% off list at 1M+ conversations | Your negotiated rate | Low |
Evidence
Evaluation acceptance scorecard
Use one frozen, human-labeled ticket set and preserve every trace.
| Criterion | Weight | Required proof | Hard failure |
|---|---|---|---|
| Verified resolution | 25 | Human judgment plus no same-issue re-contact in declared window | Vendor status used as ground truth |
| Policy and action safety | 20 | Allowed actions, approvals, rollback, audit trace | Unauthorized identity or money action |
| Accuracy and grounding | 20 | Cited source supports answer at response time | Invented policy or account fact |
| Handoff quality | 15 | Context preserved; correct queue and urgency | Customer repeats the issue |
| Commercial reconciliation | 10 | Every billed unit maps to transcript and outcome | Unreconcilable invoice units |
| Operations | 10 | Named AOP owner, change log, regression suite | No owner or rollback path |
Total cost worksheet
Total-cost inputs (empty $)
Official pages describe commercial models, not dollar rates. RTCI's ~$1.50 estimate uses reported figures; replace every blank with the written proposal.
| Input | Public? | Quote cell |
|---|---|---|
| Platform / committed-spend floor | Not published; reported at ~$50,000/yr (Quiq) | $ |
| Per-conversation unit | Model documented; reported at ~$0.99 (Quiq) | $ |
| Per-resolution unit | Model documented; rate by contract (lower with larger commitments) | $ |
| Chat / voice / email channel charges | Set by contract | $ |
| Implementation and integrations | Set by contract; excluded from RTCI's estimate | $ |
| Overage, retries, escalations, abandoned contacts | Contract terms; escalations not charged on per-resolution | $ |
| Quality review and AOP ownership | Buyer cost | $ |
| Residual human handle | Buyer cost | $ |
Frequently asked
Decagon, answered
Is Decagon worth it in 2026?
For high-volume enterprises, it is a credible option. Decagon reached an 80% resolution rate in RTCI's November 2026 test at an estimated ~$1.50 per resolution, for an RTCI score of 682, Strong tier, #4 of 20. The cost figure rests on reported, not published, pricing, so the case depends on the rate in your quote.
How much does Decagon cost?
RTCI estimates ~$1.50 per resolution as of November 2026, from reported per-conversation pricing: about $0.99 per conversation plus a ~$50,000-a-year platform fee, or roughly $1.20 per conversation at 20,000 a month, divided by 80%. Decagon publishes no rate card; decagon.ai/pricing returned a 404 on Nov 5, 2026. Your written quote replaces this estimate.
What is Decagon's real resolution rate?
80% in RTCI's November 2026 test. Decagon publishes no company-wide rate. Its named case studies report Chime at 70%+ on chat and nearly 70% on voice, Valon at 50%+ on voice, and Duolingo at an 80% deflection rate. Deflection is not resolution, so measure verified resolution with re-contact checks on your own traffic.
Who should consider Decagon instead of competitors?
Consider Decagon when an independent omnichannel agent layer and CX-owned AOPs matter. Sierra AI (RTCI score 612, Moderate) is a different enterprise agent model, Kore.ai (779, Elite) a broader conversational platform, and Zendesk AI (592, Moderate) the choice when staying inside the helpdesk suite is the priority.
Does Decagon publish transparent pricing?
No. When RTCI rechecked official materials on Nov 5, 2026, there was no pricing page and no dollar rate card. Model names are public; unit prices, minimum commitment, implementation fees, overage, and renewal terms are set by contract. RTCI rates transparency 1 of 5.
How long does Decagon take to implement?
RTCI has no verified standard implementation timeline. Ask for a dated plan covering integrations, AOP ownership, test-set creation, shadow mode, channel rollout, acceptance thresholds, rollback, and the vendor/customer staffing split.
What are the best alternatives to Decagon?
Sierra AI (73% resolution in RTCI's November 2026 test, RTCI score 612, Moderate) is a peer enterprise agent layer. Kore.ai (779, Elite) is a broader enterprise conversational platform. Zendesk AI (592, Moderate) is the suite path for existing Zendesk customers. Compare each with the same ticket set and fully loaded cost model.
Is Decagon's pricing negotiable?
Yes. Decagon itself says larger committed resolution volumes lower the per-resolution rate, and a Vendr buyer note reports 30% off list at 1M+ conversations. Negotiate billable-event definitions, committed volume, overage, implementation, audit access, service levels, data export, renewal uplift, and termination rights.
How does RTCI score Decagon?
RTCI score = 250 × log10(10 × realistic resolution rate ÷ cost per resolution), on a 0–1,000 scale. Decagon's 80% comes from RTCI's November 2026 test and its ~$1.50 is estimated from reported per-conversation pricing, so 250 × log10(10 × 80 ÷ 1.50) = 682 That places it in the Strong tier, #4 of 20. Transparency is rated 1 of 5.
Ask an AI about this review
Opens the assistant with this review's prompt copied to your clipboard — paste it in if it doesn't autofill.
Keep comparing
How Decagon stacks up against the rest
This is RTCI's independent editorial analysis, not vendor marketing copy. Figures marked as estimates are directional, not contractual. Reconcile them with Decagon's live pricing page before you sign anything. RTCI has no commercial relationship with Decagon: no payment, no affiliate link, no influence over this rating.