What it is
eesel AI is an AI support agent that bolts onto the helpdesk you already run instead of replacing it. You connect Zendesk, Freshdesk, Intercom, Gorgias, Help Scout, HubSpot, Front, Salesforce, or Jira Service Management; point it at your past tickets, help centre, Google Drive, Confluence, Notion, or SharePoint; brief it in plain English on tone, scope, and when to hand off; and it answers, drafts replies, tags and triages, and escalates to a human inside the existing queue. Ecommerce connectors (Shopify, WooCommerce, BigCommerce, Magento, Recharge) let it look up orders and subscriptions rather than only quoting policy. The same knowledge also powers an internal Q&A bot in Slack or Microsoft Teams.
The company is small and Sydney-based, founded by Amogh Sarda, Julien Zmiro, and Stefen Alper, with roughly $850K of disclosed seed funding from Airtree Ventures and KFund. That size matters for the buying decision below.
Why it shows up in Customer Success stacks
- No re-platforming. The agent works inside the ticket views, macros, and SLAs your team already uses. A support lead can pilot it on one queue in an afternoon without a migration project.
- Simulation before go-live. eesel runs the agent against thousands of your historical tickets in a sandbox and shows what it would have answered. That gives you a measured resolution rate on your own data before a single customer sees a reply, which is the number the pricing math below depends on.
- Pricing with no seats and no platform fee. Nothing to pay until a ticket is handled, and no minimum, which makes it the lowest-commitment way to put an AI agent on an existing Zendesk or Freshdesk queue.
Pricing
In Q1 2026 eesel dropped its old Team ($239/month for 1,000 interactions) and Business ($639/month for 3,000) plans for pure pay-as-you-go:
- Regular task (one support ticket or chat session, however many replies): $0.40
- Light task (dashboard questions, simple lookups): free
- Heavy task (a researched blog draft): $4.00
- Enterprise: $1,000/month platform fee plus usage, adding SSO, HIPAA/BAA, higher knowledge limits, and a named solutions engineer
- Annual commitment takes up to 25% off; the trial is $50 of free usage with no card
The detail that decides the comparison: eesel bills per ticket handled, not per ticket resolved. Fin charges $0.99 per outcome on a $49/month base that includes 50 resolutions; Zendesk charges roughly $1.50 per automated resolution committed or $2.00 pay-as-you-go; Gorgias charges $0.90–$1.00. Divide $0.40 by your resolution rate to get eesel’s real cost per resolution. At 40% resolution it is $1.00, level with Fin. At 60% it is $0.67. At 25% it is $1.60, and the per-outcome vendors win.
Worked example: 5,000 tickets a month, agent scoped to all of them, 55% resolved. eesel costs $2,000 ($1,500 with the full annual discount). Fin costs $49 plus 2,700 extra outcomes at $0.99, about $2,722. Zendesk committed costs about $4,125. Scope eesel to the 2,000 tier-1 tickets where it resolves 70% and the bill drops to $800 for 1,400 resolutions, or $0.57 each.
Best for
Support leads at SMB and mid-market companies with 1,000–20,000 tickets a month on Zendesk, Freshdesk, Gorgias, or Help Scout who want AI resolution on a repetitive tier-1 queue (order status, how-to, account questions) without changing helpdesk or signing an annual contract.
Not for regulated enterprises that need a finished SOC 2 Type II report on day one, or teams whose resolution rate in simulation comes in under 40%.
Watch-outs
- The default $250 monthly cap pauses the agent. At $0.40 a ticket that is 625 tickets, which a mid-sized queue burns in days. Guard: set the cap at 1.5× your peak-month ticket count times $0.40, act on the 75% email alert, and confirm paused tickets fall back to the human queue rather than sitting unanswered.
- You pay for the tickets it fails on. A ticket the agent drafts, escalates, or gets wrong still costs $0.40. Guard: run the simulation on at least 1,000 historical tickets per ticket type, then scope the live agent by tag or form to the types clearing 40% resolution; keep the rest human.
- Early-stage vendor, compliance still in progress. SOC 2 Type II is underway, not issued; the vector store and model providers (OpenAI, Anthropic, Google) are certified, and eesel states customer data is never used for training. EU data residency is on request; BAA only on Enterprise. Guard: get the audit timeline and a bridge letter in writing, require a data-export and deletion clause, and keep your macros and help centre as the system of record so the agent is replaceable.
- Thin independent review base. G2 shows 4.6/5, but from about 15 reviews, and most comparison content ranking eesel is published by eesel. Guard: decide on your own simulation numbers, not on third-party rankings.
Alternatives
Intercom Fin is the market-share leader and the pick when resolution rates will be modest, because per-outcome billing charges nothing for failures. Note that Fin, formerly Intercom, is being acquired by Salesforce ($3.6B, announced 2026-06-15), so factor roadmap and pricing risk into a multi-year contract. Zendesk’s native AI agents are the other incumbent: pick them when you want a single vendor and already pay for Suite, and accept the higher per-resolution price. Sierra is the fastest-growing entrant, from ~$26M ARR at the end of 2024 to ~$200M by May 2026, but it is a custom-priced enterprise build for voice and complex backend actions; pick it (or Decagon) above roughly 50,000 conversations a month. Shopify brands already on Gorgias should test its built-in AI Agent first. Pick eesel when you keep your helpdesk, your simulated resolution rate clears 40%, and you want no seats and no minimum.