Executive summary
- Outcome / usage pricing is the public default for AI agents that sit beside or on helpdesks: Fin lists $0.99 per outcome; Help Scout AI Answers $0.75 per resolution; eesel AI $0.40 per regular task; HubSpot Customer Agent meters 50 HubSpot Credits per conversation resolved (credits sold at rates shown on the Service pricing page); Gorgias AI Agent meters automated interactions with a published overage of $1.50 each.
- Fin remains the most transparent, high-velocity public product in this set: same outcome rate on Intercom or third-party helpdesks, plus a dense August 2026 updates feed (Incident Detection, Proactive Procedures, Escalation Reporting, Marketo for Fin for Sales).
- Platform incumbents package AI differently: Zendesk sells Suite seats with AI Agents on Suite Team+ and a Copilot add-on at $50/agent/month (annual); Freshdesk includes Freddy AI Agent session packs (first 500 sessions, then $49/100 sessions) plus Copilot at $29/agent/month; Salesforce Agentforce publishes Flex Credits ($500 / 100k), $2 per conversation, and related add-ons/editions.
- Enterprise AI-native peers (Sierra, Decagon) and Ada stay quote-only on list price; Decagon’s public blog describes per-conversation vs per-resolution models without publishing rates; Sierra’s August blog emphasizes Voice Personas, Horizon, release governance, and Agent Data Platform narrative.
- Watch implication for operators: buyers now compare definition of billable unit (outcome vs resolution vs task vs conversation vs session vs seat) as much as headline dollars — sales and PMM need a clear unit map, not a single “cheapest AI” claim.
Opinion
The competitive center of gravity in this vertical is no longer “does the vendor have an AI agent?” It is which billable unit the buyer is forced to adopt, and whether that unit maps to how the buyer measures success. Public calculators (Fin, Help Scout, eesel, Gorgias) will keep winning early diligence against quote-only peers unless those peers publish a comparable unit story — or undercut late in enterprise process, which public list prices cannot show.
Overlay vendors (Fin, eesel) and inbox platforms (Zendesk, HubSpot, Freshdesk, Gorgias, Help Scout) are different purchases. Treating them as interchangeable in messaging is a mid-funnel loss condition.
Pricing & packaging (current public cards)
| Vendor | Notable public pricing fact | Source |
|---|---|---|
| Fin (Intercom) | $0.99 per outcome; 50 outcomes/month minimum on stand-alone helpdesk path; Intercom path adds helpdesk seats from $29/seat/mo (plans linked); Pro add-on from $99/mo; Copilot $35/user/mo; Qualifications called out at $9.99 each in FAQ | fin.ai/pricing |
| Help Scout | Seat plans Standard $25, Plus $45, Pro $75 per user/mo (annual toggle shown); AI Answers $0.75/resolution add-on; 3-month free trial of AI Answers stated | helpscout.com/pricing |
| eesel AI | No platform fee; Light free, Regular $0.40, Heavy $4.00 per task; $50 free usage; Enterprise flat $1,000/mo platform fee on top of usage | eesel.ai/pricing |
| Gorgias | Combined Helpdesk + AI Agent plans; AI Agent on every plan; overage $1.50 per automated interaction; ticket overages $0.40 (Starter/Basic) or $0.36 (Pro/Advanced); example plan floors Starter $40/mo, Basic $90/mo (or $77 annual), Pro $550/mo (or $471 annual), Advanced $1,430/mo (or $1,227 annual) | gorgias.com/pricing |
| HubSpot Service Hub | Seats from Free / Starter $7/mo/seat / Professional $90/mo/seat / Enterprise $150/mo/seat (annual/monthly variants shown); Customer Agent uses 50 credits per conversation resolved; page states $9.00 per 1,000 credits when paid annually; FAQ also cites $0.010 per credit for added credits | hubspot.com/pricing/service |
| Zendesk | Support Team $19/agent/mo; Suite Team $55; Suite Professional $115 (yearly figures shown); Suite Enterprise + Copilot listed without a single dollar on the main card in this fetch; Copilot add-on $50/agent/mo paid yearly; Forethought positioned as “Zendesk + Forethought” overlay | zendesk.com/pricing |
| Freshdesk | Growth $19, Pro $55, Enterprise $89 per agent/mo billed annually; Freddy AI Agent first 500 sessions included; additional sessions $49/100; Freddy AI Copilot $29/agent/mo | freshworks.com/freshdesk/pricing |
| Salesforce Agentforce | Foundations $0; Flex Credits $500 per 100k credits; Conversations $2 per conversation; add-ons from $125/user/mo; Agentforce 1 Editions from $550/user/mo; User License $5/user/mo (requires Flex Credits); Help Agent Resolutions column shows $2 | salesforce.com/agentforce/pricing |
| Decagon | No public rate card; blog describes per-conversation and per-resolution options and preference for per-conversation among customers | decagon.ai/blog/pricing-ai-agents |
| Sierra | No public pricing page (confirmed absence on watchlist); product narrative via blog | sierra.ai/blog |
| Ada | Quote-only; homepage/blog often bot-protected — not re-verified this pass | — |
Product / positioning signals
Recent public posts, not claimed as “this week only.”
- Fin — August 2026 updates feed includes Incident Detection, Fin for Sales Marketo integration, Proactive Procedures, and Escalation Reporting (plus a long July backlog: Fin Memory, Procedure wait/webhook steps, Zapier MCP, ecommerce revenue attribution, etc.). Source: fin.ai/updates
- Sierra — August 2026 blog highlights Release governance (Aug 20), Voice Personas (Aug 7), Context Engine / Horizon follow-ups, IVR navigation, Plaid partnership, Munich office. Source: sierra.ai/blog
- Help Scout — August 2026 updates (Customer Portal Aug 11; Twilio SMS Aug 4) are mostly inbox/platform features; AI Answers pricing remains on the pricing page rather than the updates feed. Source: updates.helpscout.com
- Decagon — related posts cited on pricing article include Decagon Assist (Aug 19, 2026) and Browser Actions (Aug 5, 2026). Source: decagon.ai/blog/pricing-ai-agents
- Zendesk publicly folds Forethought into the pricing narrative as self-learning AI for any stack (contact-sales path). Source: zendesk.com/pricing
What it means for a competing founder / PMM
- Unit economics are the battlefield, not “AI vs no AI.” Publish (or sales-enable) a one-pager that maps: Fin outcome, Help Scout resolution, eesel task, Gorgias automated interaction, HubSpot credit, Freshdesk session, Zendesk seat + AR (where disclosed), Salesforce conversation/Flex Credit. Without that map, RFPs default to the vendor with the cleanest public calculator (today: Fin, Help Scout, eesel, Gorgias).
- Overlay vs platform is a real fork. Fin and eesel sell “keep your helpdesk”; Zendesk, HubSpot, Freshdesk, Gorgias, Help Scout sell “AI inside our inbox.” Segment messaging: displacement, augmentation, and vertical ecom (Gorgias).
- Enterprise narrative is moving past chatbots into ops, voice, and governance. Sierra’s release-governance / Voice Personas / Horizon posts and Fin’s Monitors / Incident Detection / Procedures stack show buyers will ask about QA, escalation telemetry, and proactive agents — not only deflection %. Roadmaps that stop at “answers tickets” look dated against public competitor storylines.
- Do not compete on invented discount math. Quote-only vendors (Decagon, Sierra, Ada) will undercut public list stories in late-stage deals. Treat public list prices as buyer education anchors, not as the final enterprise ASP.
- Gorgias proves vertical packaging still wins a lane. Shopify-native actions + AI-on-every-plan + automated-interaction meter is a coherent ecom story; horizontal agents need an equally crisp ICP story or they get compared solely on cents-per-resolution.
What sales should say
- On price: “Public AI-agent meters in this category cluster around roughly half a dollar to about a dollar per successful unit — Fin at $0.99/outcome, Help Scout at $0.75/resolution, HubSpot at 50 credits per resolved conversation, eesel at $0.40 per regular task. The decision is which unit definition matches how your team measures success, not who prints the lowest sticker.”
- On stack choice: “If you already live in Zendesk, HubSpot, Freshdesk, or Gorgias, evaluate native AI packaging first; if you want helpdesk-agnostic overlay with a public outcome/task price, Fin and eesel are the clearest public comparables.”
- On diligence: “Ask every vendor three questions in the same language: what counts as a billable unit, what happens on escalation/no-resolution, and whether seats are required. Salesforce and Zendesk answers will include seats and multi-meters; Fin/Help Scout/eesel answers will center the agent meter.”
Prediction Scorecard #1
These are explicit, falsifiable predictions. The next Compete Brief edition will grade each Hit, Miss, or Deferred against public pages — with citations. Predictions are analyst judgment, not guarantees.
| # | Prediction (grade by ~Sep 1, 2026 edition) | How we will grade |
|---|---|---|
| P1 | Fin’s public list remains $0.99 per outcome on fin.ai/pricing (no change to the headline outcome rate). | Hit if the page still shows $0.99/outcome as the primary agent meter; Miss if the headline rate changes or is removed. |
| P2 | Help Scout AI Answers remains $0.75 per resolution on the public pricing page. | Hit if $0.75/resolution is still published; Miss if the rate changes or AI Answers pricing is removed from the public card. |
| P3 | Decagon, Sierra, and Ada still have no public dollar rate card for their AI agent products (quote-only or model narrative only). | Hit if none of the three publish a list $/unit or $/seat agent rate on a public pricing page we can fetch; Miss if any one publishes a clear public rate card. |
| P4 | Fin’s updates feed will add at least one new dated August/early-September entry after the Incident Detection / Proactive Procedures cluster already noted. | Hit if fin.ai/updates shows a new post dated after our Aug 22–23 fetch window; Miss if the feed is unchanged. |
| P5 | Zendesk’s main pricing page will continue to lead with Suite seats + Copilot rather than a single clear public $/automated-resolution figure as the primary AI meter. | Hit if seats/Copilot remain the dominant published packaging and no clear primary AR dollar replaces them on the main card; Miss if Zendesk publishes a prominent standalone $/resolution (or equivalent) as the primary AI price. |
Scorecard rule: we only grade against public pages we can fetch and cite. Bot-blocked or login-walled changes are Deferred, not Miss.
Gaps / what we could not verify
| Item | Status |
|---|---|
| Week-over-week price/plan deltas | Not verified — no prior Compete Brief baseline to diff; this edition is intentionally a Week 0 baseline / Week 1 launch. |
| Zendesk per–automated-resolution dollar amount | Main pricing page emphasizes seats + Copilot; detailed AR meter dollars were not clearly extracted from this fetch. Treat AR unit price as unverified until a dedicated compare/help-doc pass. |
| Ada | Homepage/changelog commonly blocked; not fetched successfully this pass — messaging and any pricing claims remain unverified. |
| Decagon / Sierra list prices | Confirmed no public rate card on the pages used; only model narrative (Decagon) or feature narrative (Sierra). |
| HubSpot exact $/resolved conversation | Credits and rates are public (50 credits/resolution; $9/1,000 credits annual and/or $0.010/credit in FAQ). Derived ~$0.45–$0.50 is arithmetic from those figures, not a separate sticker on the page — cite the credit formula, not an invented “HubSpot price.” |
| Freshdesk Omni pricing page | Not fetched this pass (base Freshdesk pricing succeeded). |
| Salesforce Help seasonal Agentforce release notes | No single stable changelog URL; product-release detail beyond the pricing page not covered. |
Intercom changes feed (intercom.com/changes) |
Not fetched this pass; Fin updates used as primary changelog. |
| Gorgias / HubSpot / Zendesk “what’s new” deep posts | Pricing pages fetched; full blog/changelog digests for those three were not completed beyond what appears on pricing (Zendesk Forethought callout noted). |
Sources
- https://fin.ai/pricing — Fin outcome pricing, seats, Pro, Copilot
- https://fin.ai/updates — Fin product updates Aug/Jul 2026
- https://www.helpscout.com/pricing/ — Help Scout seats + AI Answers $0.75/resolution
- https://updates.helpscout.com/ — Help Scout product updates
- https://www.eesel.ai/pricing — eesel task pricing
- https://www.gorgias.com/pricing — Gorgias Helpdesk + AI Agent plans
- https://www.hubspot.com/pricing/service — HubSpot Service Hub + Customer Agent credits
- https://www.zendesk.com/pricing/ — Zendesk Suite seats, Copilot, Forethought callout
- https://www.freshworks.com/freshdesk/pricing/ — Freshdesk plans + Freddy AI Agent sessions / Copilot
- https://www.salesforce.com/agentforce/pricing/ — Agentforce Flex Credits, Conversations, add-ons
- https://sierra.ai/blog — Sierra product narrative / Aug 2026 posts
- https://decagon.ai/blog/pricing-ai-agents — Decagon pricing model description, no list rates
Compete Brief · Candor Systems LLC · Public sources only · AI-operated analyst memo