# KimonDesk > AI-native helpdesk with flat per-tier pricing. Every feature, every tier. > Free forever, paid from £49/mo ex VAT. Built by Kimon Services. ## Compare pages — TL;DRs ### vs Zendesk Choose Zendesk if you need its Salesforce-deep sync or 1500+ marketplace today. Choose KimonDesk if per-agent pricing is killing your unit economics. On a typical 8-agent team, KimonDesk Growth is roughly £17,300 a year cheaper and includes AI, SSO, and audit log without surcharge. Read more: https://kimondesk.com/compare/zendesk ### vs Freshdesk Choose Freshdesk if you need deep Hindi NLU, an existing Freshworks bundle, or a 2-3 agent team where Pro maths still works. Choose KimonDesk if AI as a core feature matters. On 8 agents, KimonDesk Growth is roughly £7,000 a year cheaper. Read more: https://kimondesk.com/compare/freshdesk ### vs Intercom Choose Intercom if in-app messaging or product analytics is core to your product. Choose KimonDesk if per-resolution AI charging is killing your budget predictability. On 8 agents at typical SME volume, KimonDesk Growth is roughly £11,900 a year cheaper. Read more: https://kimondesk.com/compare/intercom ### vs HubSpot Choose HubSpot if you already pay for the CRM and want native unification across sales, marketing, and support. Choose KimonDesk if you do not already run on HubSpot. On 8 seats, KimonDesk Growth is roughly £10,200 a year cheaper and includes AI without surcharge. Read more: https://kimondesk.com/compare/hubspot ### vs Zoho Desk Choose Zoho Desk if you already run on Zoho One and the bundle economics work for you. Choose KimonDesk if you are not in the Zoho ecosystem. On 8 agents with Zia AI and SalesIQ, KimonDesk Growth is roughly £4,600 a year cheaper. Read more: https://kimondesk.com/compare/zoho-desk ### vs Help Scout Choose Help Scout if you are a pure email support team and minimalism matters more than channel breadth. Choose KimonDesk if you need WhatsApp, voice, or social DMs now or within a year. On 8 users, KimonDesk Growth is roughly £4,260 a year cheaper. Read more: https://kimondesk.com/compare/help-scout ### vs Salesforce Choose Salesforce if you run on Sales Cloud at 200+ seats with enterprise certification needs. Choose KimonDesk if you are SME or mid-market without Sales Cloud. On 8 users with Einstein, Digital Engagement, and partner implementation, KimonDesk Growth is roughly £26,000 a year cheaper in year one. Read more: https://kimondesk.com/compare/salesforce ## Glossary ### AI auto-resolution AI auto-resolution is the practice of letting a language model answer a customer ticket end-to-end when the model is confident the response will resolve the question without a human touch. The helpdesk closes the ticket if the customer confirms, or escalates otherwise. Read more: https://kimondesk.com/glossary/ai-auto-resolution ### Audit log An audit log is a tamper-resistant record of every meaningful action inside the helpdesk: who logged in, what they changed, which tickets they viewed, when they exported data. Auditors, regulators, and security reviews require it as evidence of access controls. Read more: https://kimondesk.com/glossary/audit-log ### Brand voice training Brand voice training is the process of teaching an AI helpdesk model to reply in your brand's tone, vocabulary, and formatting conventions. Done well, customers cannot tell whether the response came from an agent or the model. Done poorly, the gap is jarring. Read more: https://kimondesk.com/glossary/brand-voice-training ### Collision detection Collision detection is a helpdesk feature that warns agents when a colleague is already typing a reply on the same ticket they just opened. It prevents two agents from sending duplicate or conflicting replies to the customer at the same time. Read more: https://kimondesk.com/glossary/collision-detection ### CSAT CSAT, short for Customer Satisfaction Score, is a measurement that captures how a customer felt about a single support interaction, usually as a 1 to 5 rating gathered immediately after a ticket is resolved. Most helpdesks express it as the percentage of positive responses. Read more: https://kimondesk.com/glossary/csat ### Customer portal A customer portal is a self-service web surface where customers can search the knowledge base, raise a ticket, track existing tickets, and update their account details without contacting an agent. A good portal deflects 20 to 30 percent of routine inbound queries. Read more: https://kimondesk.com/glossary/customer-portal ### Escalation An escalation is the act of routing a ticket from a junior agent to a senior agent, team lead, or specialist queue when the original owner cannot resolve it within scope, time, or authority. Good escalation paths shorten resolution time without overloading senior staff. Read more: https://kimondesk.com/glossary/escalation ### FRT First-Reply Time, or FRT, is the elapsed time from when a customer submits a ticket to when an agent sends the first human or AI-drafted reply. It is the metric customers feel most acutely and the easiest to influence with staffing or automation. Read more: https://kimondesk.com/glossary/frt ### Intent detection Intent detection is a machine-learning step that reads an inbound customer message and labels it with the underlying purpose: refund request, password reset, billing query, complaint, and so on. The label drives routing, prioritisation, and whether AI auto-resolution attempts a reply. Read more: https://kimondesk.com/glossary/intent-detection ### Internal note An internal note is a comment on a ticket visible to agents and admins but never to the customer. Internal notes are how a team coordinates handoffs, asks colleagues for input, and records context that should not appear in the customer-facing thread. Read more: https://kimondesk.com/glossary/internal-note ### Macro A macro is a pre-written reply template an agent can insert into a ticket with one click, optionally combined with actions like changing status, adding tags, or assigning to a team. Macros speed up common replies without forcing copy-paste from a Google Doc. Read more: https://kimondesk.com/glossary/macro ### MTTR MTTR, short for Mean Time to Resolution, is the average elapsed time from when a customer ticket opens to when it is marked solved. It measures the throughput of a support team rather than the speed of a single first reply. Read more: https://kimondesk.com/glossary/mttr ### Omnichannel Omnichannel support means a single helpdesk that handles every customer channel, email, chat, WhatsApp, social, voice, in one unified inbox. Agents see the full conversation history across channels in one thread, so the customer never has to repeat themselves. Read more: https://kimondesk.com/glossary/omnichannel ### Routing rule A routing rule is a saved condition that automatically assigns inbound tickets to a team, agent, or queue based on attributes like channel, customer tag, intent, or business hours. Routing rules replace the senior agent who used to triage by hand at 9am. Read more: https://kimondesk.com/glossary/routing-rule ### Sandbox A sandbox is a separate copy of your helpdesk where you can test workflow changes, AI prompt edits, integrations, and migrations without affecting the production environment your agents and customers rely on every day. The sandbox is wiped clean on demand. Read more: https://kimondesk.com/glossary/sandbox ### Sentiment detection Sentiment detection is a model that scores each customer message for emotional tone, typically as positive, neutral, or negative. The score lets a helpdesk prioritise angry customers, route complaints to a senior agent, and flag at-risk relationships before they churn. Read more: https://kimondesk.com/glossary/sentiment-detection ### SLA A Service Level Agreement, or SLA, is a written commitment to respond to and resolve customer tickets within defined timeframes. Helpdesks track SLAs against business hours, ticket priority, and customer tier so support teams can spot breaches before customers complain. Read more: https://kimondesk.com/glossary/sla ### SSO Single Sign-On, or SSO, lets your agents log in to the helpdesk using your existing identity provider, such as Google Workspace, Microsoft Entra, or Okta. SSO removes a separate password to manage and lets IT enforce one access policy across the stack. Read more: https://kimondesk.com/glossary/sso ## Blog posts — titles and TL;DRs ### The True Cost of Zendesk in 2026: What an 8-Agent Team Actually Pays Published 2026-04-26 by Michael Kitt. Itemising what an 8-agent team on Zendesk Suite Professional with Advanced AI pays per year, using public list pricing as of Q2 2026, and what the same team pays on KimonDesk. An 8-agent team on Zendesk Suite Professional with Advanced AI pays around £18,800 a year on list pricing. KimonDesk Growth costs £1,428 a year for the same team with AI included. The difference is roughly £17,400 a year, itemised below. Read more: https://kimondesk.com/blog/true-cost-zendesk ### What 'AI-Native Helpdesk' Actually Means in 2026 Published 2026-04-25 by Michael Kitt. Helpdesks marketed as AI-native have automated drafts, summarisation, sentiment, and auto-resolution baked into core. Bolt-on AI is sold as a premium tier add-on. An AI-native helpdesk has automated draft replies, ticket summarisation, sentiment detection and auto-resolution as core features rather than premium add-ons. In 2026, traditional vendors charge $50/agent/month extra for AI; AI-native vendors include it at every tier. Read more: https://kimondesk.com/blog/ai-native-helpdesk ### Why Your Support Shouldn't Run on Gmail (And What to Switch To) Published 2026-04-24 by Michael Kitt. Shared inboxes break at 3 agents: collision, no SLA, no audit, broken search. A migration anatomy plus what a basic helpdesk gives you. A shared Gmail inbox stops working at three agents. You cannot measure first-reply time, enforce SLAs, avoid duplicate replies, or search threads after a few months. Move to a helpdesk when you cross your second support hire, not your fifth. Read more: https://kimondesk.com/blog/why-not-gmail ### Zendesk Alternatives for SMEs in 2026: An Honest Buyer's Guide Published 2026-04-23 by Michael Kitt. Where to look when Zendesk Suite Professional gets too expensive. Honest tradeoffs across the four most-named alternatives plus KimonDesk. The best Zendesk alternative depends on team size and channel mix. For 2 to 15 agents wanting flat per-tier pricing with AI included, KimonDesk fits closest. For sales-heavy teams, HubSpot Service Hub. For minimalists, Help Scout. Honest tradeoffs below. Read more: https://kimondesk.com/blog/zendesk-alternatives-sme-2026 ### How AI Auto-Resolution Works (and Where It Should Never Be Used) Published 2026-04-22 by Michael Kitt. AI auto-resolution explained: confidence thresholds, the 5 ticket types it handles well, and 5 where it must not auto-act. AI auto-resolution works by matching incoming tickets to your knowledge base and replying when confidence is high. It works for refunds, password resets and shipping queries. It should never auto-resolve emotional escalations, abuse reports or anything financial above your refund threshold. Read more: https://kimondesk.com/blog/ai-auto-resolution ### First-Reply Time Benchmarks for SME Support Teams Published 2026-04-21 by Michael Kitt. What median, top-quartile, and bottom-quartile first-reply times look like for SME helpdesks in 2026, and the levers that move the dial. Median first-reply time for SME support in 2026 is around two hours during business hours and fourteen hours overnight. Top-quartile teams hit thirty minutes business-hours and four hours overnight. Pre-drafted templated replies and automated routing are the biggest two levers. Read more: https://kimondesk.com/blog/first-reply-time-benchmarks ### Building a Support Team from 1 to 10 Agents Without Burnout Published 2026-04-20 by Michael Kitt. Reactive hiring beats proactive hiring. The thresholds that signal each new agent, plus the tool ladder that gates expansion. Hire support reactively, not proactively. The first hire goes when one founder spends over 50 percent of their time replying. The second when sustained nights or weekends. Beyond five, hire team-leads alongside agents. Tooling investment gates expansion: agent number three needs a real helpdesk. Read more: https://kimondesk.com/blog/scaling-support-1-to-10 ### The Hidden Cost of 'Enterprise Call Us' Pricing Published 2026-04-19 by Michael Kitt. Hidden enterprise pricing is a sales tactic, not a buyer benefit. Four ways it hurts you, the vendor incentives that drive it, and the transparent alternative. Hidden enterprise pricing is a sales tactic, not a buyer benefit. It defers your decision past your evaluation window, lets sales reps anchor high, and obscures who is getting what discount. The transparent alternative, flat per-tier pricing, saves twenty to fifty percent in negotiation overhead alone. Read more: https://kimondesk.com/blog/enterprise-call-us-pricing ## Key facts - Founded by Michael Kitt, Co-Founder of Kimon Services (kimonservices.com). - Pricing model is flat per-tier (£0 / £49 / £149 / £499 ex VAT). No per-agent surcharge. - AI-native, auto-resolution, drafts, summaries, sentiment included on every tier (not an add-on). - Self-hosted on a VPS in the UK. EU-resident hosting guarantee on the Scale tier. ## Blog posts — full bodies ### The True Cost of Zendesk in 2026: What an 8-Agent Team Actually Pays An 8-agent team on Zendesk Suite Professional with Advanced AI pays around £18,800 a year on list pricing. KimonDesk Growth covers the same team for £1,428 a year with AI included. The headline gap is roughly £17,400 a year, every line item itemised below from public vendor pricing pages accessed in Q2 2026. ## Why this question keeps coming up Almost every conversation I have with a small support leader starts the same way. They have grown to seven, eight, nine agents. The Zendesk renewal lands. The number on the quote does not match the number in their head, and the gap between the two has been widening for three years running. The marketing pages talk about per-agent pricing in clean round figures. The renewal quote talks about Suite Professional, plus Advanced AI, plus implementation top-ups, plus the SSO upgrade that someone in security insisted on, plus the sandbox the developers asked for so they could test workflow changes without breaking production. None of those line items are surprising on their own. Together, they roughly double the headline number. The point of this post is not to attack a company. It is to walk you through the maths, line by line, with public pricing, so you can do the same calculation against your own renewal quote. If your numbers come out lower than mine, fair enough, send me your maths and I will publish the correction. If they come out higher, you are in the same boat as the rest of the SME support world right now. ## The 8-agent benchmark I picked 8 agents because that is the median size of the support teams switching off legacy helpdesks onto KimonDesk in the last quarter. Two of them are part-time. One does live chat only. One owns escalations. The rest split inbound by channel. None of this is exotic. It is the size that justifies a real helpdesk over a shared inbox, but does not yet justify a dedicated buyer's role to negotiate vendor contracts. For the rest of the post, I am pricing the same team in two ways. First on Zendesk Suite Professional with Advanced AI, which is the configuration most teams of this size land on once they realise the AI add-on is the only way to get auto-resolution and reply drafting. Second on KimonDesk Growth, which is our flat-tier price covering up to 25 agents. ## Zendesk: what the public pricing page says Suite Professional sits at $115 per agent per month on the published list, billed annually. That is the entry point that includes SLA management, multi-channel ticketing and the basic skills-based routing most growth-stage teams need. There is a cheaper Suite Team tier at $69 per agent per month, but it omits SLA and most of the routing features, so almost nobody at the 8-agent mark stays on it. For our 8 agents, that is $115 multiplied by 8 multiplied by 12, which works out to $11,040 per year. At the fixed 0.79 GBP exchange rate I have been using for all 2026 vendor comparisons, that is £8,722. This is the number people quote when they say "Zendesk is about £900 per month for our team." It is also the number that bears almost no relationship to what they actually pay on the renewal. ## The first add-on: Advanced AI Reply drafting, automated triage, sentiment detection and the auto-resolver all sit behind the Advanced AI add-on. Zendesk publishes it at $50 per agent per month. For 8 agents that is $4,800 per year, or £3,792. The add-on is worth paying for. AI auto-resolution, when configured well, will close 30 to 60 percent of repetitive tickets before a human sees them, which is what makes the maths of an 8-agent team work in the first place. The point is not that it is bad. The point is that it is sold as a separate line item on top of the seat price. ## The second add-on: workforce management or QA Once you cross five agents you start needing scheduling and quality assurance. Zendesk's WFM and QA modules are priced separately, typically around $42 per agent per month combined for the lighter tier. For 8 agents that is $4,032 per year, or £3,184. You do not strictly need WFM and QA on day one. Most teams add them in year two when they realise they have no consistent way to evaluate agent quality or plan shift coverage. By year three, both modules are standard. ## The third add-on: sandbox, SSO and audit log Suite Professional does not include a sandbox environment, SAML SSO or a fully exportable audit log. To get all three you upgrade to Suite Enterprise, which adds roughly $25 per agent per month above Professional. For 8 agents that is $2,400 a year incremental, or £1,896. Some teams sidestep this by negotiating bundled pricing, but the published list price is the published list price. If you would rather not upgrade, the alternative is to operate without a sandbox (risky for any kind of automation work), without SSO (a security audit failure for most regulated SMEs), or without the audit log (compliance friction the moment you handle anything sensitive). ## Implementation: the line item nobody talks about Zendesk's self-serve onboarding is free. Guided implementation runs from $2,500 for a basic engagement up to mid five figures for a partner-led rollout. Most 8-agent teams pick something in the middle, around $2,500 to $5,000. I am taking the lower end and amortising it across the first year. That puts implementation at $2,500, or £1,975 in year one. In subsequent years it falls to zero unless you are doing a major reconfiguration. For a fair comparison over multiple years, divide by three to amortise. ## Adding it up Subtotal in dollars: $11,040 base + $4,800 AI + $4,032 WFM/QA + $1,440 sandbox/SSO incremental allocation + $2,500 implementation. That is $23,812 in year one. At 0.79 GBP per dollar, that is £18,811 a year. Two caveats. First, this is list pricing. Most teams negotiate 10 to 25 percent off if they sign a multi-year contract, which would bring the figure down to the £14,000 to £17,000 range. Second, currency moves. At the time of writing the rate is favourable; six months ago at 0.78 the same calculation produced about £18,500, so the answer is range-bound by a few hundred pounds either way. The headline is unchanged. An 8-agent team on Zendesk Suite Professional with Advanced AI is paying somewhere between £14,000 and £19,000 a year, depending on negotiation skill and currency timing. ## KimonDesk Growth: the same team Our Growth tier is £149 a month, billed monthly. It covers up to 25 agents, every channel, every AI feature, SSO, audit log, sandbox and the full automation builder. There is no per-agent surcharge, no Advanced AI add-on, no Suite Enterprise upgrade. Twelve months at £149 is £1,788. Annual prepay knocks 20 percent off, taking the year-one cost to £1,428. Implementation is self-serve and free, with optional founder-led onboarding for the first 50 customers (also free). Compare directly at the [pricing page](/pricing). The calculator there lets you toggle agent count and currency, and the output is a shareable URL you can drop into a Slack thread for your finance lead. ## The receipt, side by side | Line item | Zendesk (8 agents) | KimonDesk Growth (8 agents) | | ------------------------------------ | ------------------------------ | ---------------------------------- | | Base seat licence, per year | £8,722 (Suite Professional) | £1,788 (flat tier) | | Advanced AI add-on | £3,792 | included | | WFM / QA add-on | £3,184 | included | | Sandbox / SSO upgrade | £1,896 | included | | Implementation, year one | £1,975 | £0 | | Annual prepay discount | none on this tier | −£360 | | **Year-one total** | **£18,811** | **£1,428** | | **Difference, year one** | | **−£17,383** | That £17,383 figure is the receipt I keep referring to in talks and on the [compare hub](/compare). It is also the same number that appears across the Receipt War sections on Home, Pricing and the comparison pages, derived from this exact methodology. ## What the £17,383 actually buys you back I find it helpful to translate the saving into operational reality. £17,383 a year is roughly: - A part-time support hire on £12 an hour for 28 hours a week. That extra hand is what closes the queue at 6pm instead of letting it drift overnight. - A year of decent customer research. Real interviews, transcribed, analysed, with a researcher you do not have to share with the product team. - A two-day off-site for the support team. The kind that builds the muscle memory to handle a Black Friday week without the leader having to be in the chat every hour. The point is not that you should spend it on any of these. The point is that the gap is large enough that "we have always paid it, the renewal is what it is" stops being a defensible answer. Even a 50 percent negotiated discount on Zendesk leaves a £7,000 a year delta on the table. ## Common objections > Zendesk is the established enterprise tool. We can't risk switching. For an 8-agent team, "enterprise" is overkill. The features your team actually uses on a daily basis (multi-channel inbox, SLAs, macros, reporting) are commodity in 2026. KimonDesk ships them all on the £49 Pro tier. Switching helpdesks for a team this size is a one-week migration, not a quarter-long programme. > The list price is not what we actually pay. We negotiated 30 percent off. Apply 30 percent off the £18,811 and the answer is still £13,168, against £1,428. The delta is still over ten grand. Bring me a renewal quote that is below £5,000 a year for 8 agents on Suite Professional with Advanced AI and I will publish the correction. > Migration is too painful. Importing tickets, contacts and macros is a 30-minute job for any helpdesk that takes the API design seriously. Routing rules and SLA policies take longer because they are bespoke. We have a migration playbook and we run it for you in the first 50 accounts. The honest answer is that it is two days of careful work, not a quarter. > What about [glossary term, e.g. SSO](/glossary/sso)? SSO is included on every KimonDesk tier from Free upwards. Same for [audit log](/glossary/audit-log), full webhooks, and the developer sandbox. There is no Enterprise gate on KimonDesk because there is no Enterprise tier. The pricing model is flat. ## What this post is not This is not a "Zendesk is bad" post. Zendesk is a solid product built for a specific kind of buyer (large support organisations with dedicated procurement). The pricing model that makes sense at 200 agents stops making sense at 8. It is also not a "switch tomorrow" post. Migration is real work and the right time to do it is when your renewal is at least 90 days out. If you are mid-contract and the maths still hurts, plan the switch for the next renewal window and use the saved cost to fund the project. What it is, is a public, sourced, line-item walkthrough of the actual numbers, so you do not have to do the calculation alone. If your finance lead wants to challenge it, send them this URL. ## References - Zendesk Suite pricing, public page: list prices accessed Q2 2026. Suite Professional $115/agent/mo, Suite Enterprise $169/agent/mo, Advanced AI add-on $50/agent/mo. - Zendesk WFM and QA pricing pages, accessed Q2 2026. - KimonDesk pricing: see the [/pricing](/pricing) page for the up-to-date tier table. - Currency: fixed 0.79 USD-to-GBP rate used across all 2026 KimonDesk vendor comparisons. The rate is documented in our [comparison hub methodology](/compare). - Receipt War source: derived from this same calculation, used on Home, Pricing and Compare. Read more: https://kimondesk.com/blog/true-cost-zendesk ### What 'AI-Native Helpdesk' Actually Means in 2026 An AI-native helpdesk treats automated draft replies, ticket summarisation, sentiment detection and auto-resolution as core features that ship at every tier. AI-bolted-on means the same features exist but sit behind a separate paid add-on, typically charged at $30 to $80 per agent per month on top of the base seat licence. ## Why the term matters "AI-native" is a marketing label, and like every marketing label it gets stretched. Every helpdesk vendor in 2026 says they are AI-native because every helpdesk vendor in 2026 has shipped some kind of AI feature. The label has lost meaning at the surface. What still has meaning is the underlying architectural and commercial choice the vendor made. I want to make the distinction concrete because, when we built KimonDesk, the choice we made was a commercial one as much as a technical one. AI features are core in our product not because we are clever, but because we made an early decision that you should not have to pay extra for the part of the product that does the heaviest lifting on a busy week. This post walks through what the term actually means, the five things AI does in customer support today, and the buyer's checklist for telling apart genuine AI-native vendors from AI-enabled vendors that have learned the right marketing copy. ## The architecture distinction The honest version of "AI-native" has a specific technical signature. AI inference happens at the same layer as the rest of the product, not in a separate service that the rest of the product asks for permission. Drafts appear inline in the agent's reply box, not in a side panel that requires a click. Auto-resolution runs as part of the inbound ticket pipeline, not as a separately licensed automation. The model can be swapped (Claude or GPT, perhaps a fine-tuned in-house model later) without touching the agent UI, because the UI was designed for AI from the start. The bolt-on alternative looks different in the codebase. The original product was built for human-only handling. AI was retrofitted as a separate module, often by acquiring a smaller AI vendor or shipping a partner integration. The agent UI shows AI suggestions in a panel that sits next to the reply box. The auto-resolver lives in a separate workflow module and has to be configured to hand tickets back to the human inbox if it gives up. Neither approach is technically wrong. Bolt-on AI can be excellent. The reason the distinction matters is that bolt-on AI is almost always priced as a separate SKU, which translates into a separate line item on your renewal quote. ## The five things AI does in customer support To make the architectural distinction tangible, here are the five concrete AI capabilities that matter today, in priority order for an SME team. ### 1. Reply drafts Inbound ticket arrives. The system reads the ticket, the customer history, your knowledge base, and produces a draft reply for the agent to edit. Good drafts get accepted with one or two small changes. The lift is roughly 30 to 50 percent off median handle time once agents trust it. This is the single biggest productivity win and the easiest to evaluate. If a vendor's draft quality is poor on your specific tickets, no amount of marketing copy fixes it. Test it during the trial. ### 2. Ticket summarisation Long threads (10 plus messages, multiple channel switches, escalations) get a one-paragraph summary at the top. The summary updates as the thread grows. The benefit is for the second agent who picks up the thread on day three. Without summarisation they spend five minutes reading. With it, they spend 30 seconds. ### 3. Sentiment detection The system flags tickets where the customer is frustrated, abusive, or at risk of churning. The flag routes the ticket to a team lead or surfaces it on a dashboard. Good sentiment detection catches escalations before they become public Trustpilot reviews. Weak sentiment detection produces a wall of false positives that the team learns to ignore. ### 4. Auto-resolution For high-confidence matches against your knowledge base, the system replies to the customer directly. Good ones close 30 to 70 percent of repetitive tickets (refund-status checks, password resets, shipping queries) without an agent ever seeing the ticket. We covered the use cases and the failure modes in [How AI Auto-Resolution Works](/blog/ai-auto-resolution). ### 5. Intent and routing Inbound ticket gets classified by intent (billing, technical, sales, escalation) and routed to the right agent or queue. The benefit compounds with team size. At 8 agents the time saved is meaningful. At 2 agents it is rounding error. ## The 2026 commercial state of play Here is what the major vendors charge for these features as of Q2 2026, at list price. | Vendor | AI add-on cost | Tier gate | | ----------------------- | ----------------------------- | ---------------------------------- | | Zendesk | $50/agent/mo (Advanced AI) | Suite Professional or higher | | Freshdesk | $35/agent/mo (Freddy AI Pro) | Pro or higher | | Intercom | $0.99 per resolution + base | All tiers, usage-priced | | HubSpot Service Hub | included in Pro and above | Pro tier ($90/seat/mo) | | Salesforce Service Cloud | $50/agent/mo (Einstein AI) | Service Pro or higher | | KimonDesk | included at every tier | from Free upwards | The Intercom model is interesting because it is the only one that does not look like a flat add-on. It charges per resolved conversation. For a small team, the maths works out comparable to a $30 to $50 per agent flat fee. For a larger team it can run higher. The KimonDesk row is the architectural commitment. We do not charge for AI separately because we built the product around it. The additional cost to us per ticket of running inference is real but small, and we absorbed it into the flat tier price. ## The buyer's checklist If you are evaluating a helpdesk and the vendor calls itself AI-native, six questions cut through the marketing fog. **1. Is AI included in the cheapest paid tier?** If yes, the vendor is committed to AI as a default. If AI requires upgrading from $69/agent/mo to $115/agent/mo plus a separate $50/agent/mo add-on, the vendor sees AI as a price-discrimination lever, not a default. **2. Can I swap the underlying model?** Vendors that built AI in late often hard-coded one provider (typically OpenAI). Vendors that built AI in early designed model-swap capability from the start. The swap matters for two reasons: regulatory (some industries cannot use US-hosted models) and quality (Claude consistently outperforms GPT-4o on customer-support tone in our own testing). **3. Are AI features wired into the agent's primary workflow?** Drafts in the reply box, summary at the top of the thread, sentiment flags inline. Or are they tucked behind a side panel and a click? The wiring tells you whether AI was a foundation choice or a retrofit. **4. What happens when the AI is wrong?** Good AI integrations show their reasoning, let the agent edit before sending, and learn from the edit. Bad ones present the AI output as a fait accompli and offer "approve / reject" only. The latter does not improve over time. **5. How is AI usage priced beyond the seat licence?** Per resolution? Per draft? Per token? Or simply included? Per-token pricing is honest but unpredictable. Per-resolution pricing is predictable but creates strange incentives (you pay every time the system tries to close a ticket, success or failure, on some vendors). Included is the simplest. **6. Where does the inference happen?** EU, US, model provider's region of choice? For data residency this matters. For latency it matters less than vendors imply, given inference times are now sub-second on modern models. ## Where AI-native does not yet mean better I want to be careful not to overstate the case. AI-native is a commercial and architectural commitment, not a quality guarantee. There are bolt-on AI implementations that outperform AI-native ones because the bolt-on vendor invested heavily in model fine-tuning. There are AI-native vendors whose models are mediocre because they prioritised breadth over depth. The right way to evaluate AI quality is the trial. Pipe a week of your real tickets into the candidate helpdesk, configure the AI features, and watch your team use it. Three things to measure: - Draft acceptance rate: what fraction of AI drafts does the agent send with zero or minimal edits? - Auto-resolution accuracy: when the system closes a ticket, what fraction of customers come back with a follow-up indicating the resolution was wrong? - Sentiment flag precision: what fraction of "high-frustration" flags are actually high-frustration tickets? If the numbers come out comparable across two vendors, the AI-native one wins on commercial grounds (no add-on, no tier gate, no surprise renewal increase). If the bolt-on AI is materially better on your tickets, pay the premium and revisit at the next renewal. ## What KimonDesk does For full transparency about our own product: - Reply drafts, summarisation, sentiment, auto-resolution and intent routing are all included from the [Free tier](/pricing) upwards. - You can switch between Claude and GPT in the model picker. We default to Claude because it produces better support tone in British English. - AI usage is metered per organisation but not billed per request. The flat tier price covers it. - Inference happens in the EU by default; you can opt into US inference if you need lower latency for US-hosted customer bases. - The AI is wired into the agent's primary workflow. Drafts appear inline. Summaries sit above the thread. There is no "AI panel" to click into. We made these choices because the alternative (charging extra) felt like a regression to the model that the rest of the legacy helpdesk industry built. The whole point of starting fresh was to not repeat the same mistakes that made an 8-agent team's renewal so painful in the first place. If you want the full pricing breakdown against the legacy vendors, the [Zendesk TCO post](/blog/true-cost-zendesk) walks through the numbers. The [comparison hub](/compare) gives the side-by-side capability table. ## The closing word AI-native is not magic. It is a deliberate commercial and architectural choice that the vendor either made or did not make. When you read the term in marketing copy, ask the six questions above. The answers will tell you whether you are looking at a vendor who built for 2026 or a vendor who is hoping you will not notice the add-on line item on next year's renewal quote. For our own pricing, the [main pricing page](/pricing) shows everything that is included at every tier, with no hidden footnotes. ## References - Zendesk Advanced AI pricing page, Q2 2026. - Freshworks Freddy AI pricing page, Q2 2026. - Intercom AI pricing page (per resolution model), Q2 2026. - HubSpot Service Hub pricing page, Q2 2026. - Salesforce Service Cloud Einstein pricing page, Q2 2026. Read more: https://kimondesk.com/blog/ai-native-helpdesk ### Why Your Support Shouldn't Run on Gmail (And What to Switch To) A shared Gmail inbox stops working at three agents. You cannot measure first-reply time, enforce SLAs, avoid duplicate replies, or search threads reliably after a few months. The right time to move is when you hire your second support agent, not when you hire your fifth. ## The honest version If you are running customer support out of a shared Gmail account, you are not doing anything wrong. Almost every SME starts there. It is free, the team already knows the interface, and for the first year it works fine. The reason to move is not because shared inboxes are bad. It is because there is a specific point at which the maths stops working, and that point arrives faster than most teams expect. I want to be honest about when that point is, what fails first, and what you actually get when you move. The marketing copy from helpdesk vendors makes the switch sound dramatic. In practice it is a two-day setup and a week of parallel running. The decision to move is harder than the move itself. ## The point where it breaks In my experience the point is the third support hire. Two people in a shared inbox is workable with verbal coordination ("I'm taking the Acme thread, you do the rest"). Three people is the point where: - You can no longer hold the picture of who is doing what in your head. - Two agents start replying to the same customer at least once a week. - The senior agent spends a chunk of every day asking "did you reply to this one?" instead of replying to anything themselves. Some teams hit this wall at agent number two if the volume is high (50 plus tickets a day) or the channels are split (some email, some chat). Some teams stretch it to four if the volume is genuinely low. Three is the median. The other trigger is the moment a customer or a regulator asks for a number you do not have. "What's your median first-reply time this quarter?" or "Show me every interaction with this customer for the past 12 months." Gmail cannot answer either question without manual labour. ## Five specific failure modes Generic warnings about scale do not motivate a switch. Specific failures do. Here are the five that actually trigger the move. ### Failure 1: collision (two replies to the same customer) The customer sends in a question. Agent A starts typing a reply. Agent B, scanning the inbox a minute later, also starts typing a reply. Both hit send. The customer receives two emails, often with slightly different information. They lose confidence in the team. Internal post-mortems consume an hour. A helpdesk solves collision with [collision detection](/glossary/collision-detection) at the ticket level. Agent A opens the ticket, Agent B sees a banner saying Agent A is currently in this ticket. The detection happens in real time and is sticky enough that even brief idle times do not produce false negatives. ### Failure 2: no first-reply time measurement You promise customers a one-hour first reply during business hours. You believe you hit it most of the time. You have no actual data. When a customer escalates a slow reply, you cannot defend yourself or look at the pattern. A helpdesk records the timestamp of the first inbound message and the timestamp of the first outbound human reply. The difference is the [first-reply time](/glossary/frt). It rolls up to a dashboard, segmentable by agent, channel, and time of day. The number itself is more useful than the SLA enforcement: it lets you see the pattern and fix the cause (Friday afternoons are slow, the night shift is leaving threads unanswered, agent six is taking three times longer than agent four). ### Failure 3: SLA enforcement Without timing data, SLAs are aspirational rather than operational. You cannot promise a contractually binding response time to a customer because you have no system that knows when the response time started. A helpdesk lets you set SLA policies (e.g. "premium customers, one-hour first reply during business hours, six hours overnight") and surfaces the breaching ticket count in real time. Agents can see what is approaching breach and prioritise accordingly. Customer-success managers can have honest conversations about which segments are getting which level of service. ### Failure 4: search becomes useless Gmail's search is fine for personal inboxes. For a shared support inbox accumulating thousands of threads a year, search starts producing wrong-or-overwhelming results around the six-month mark. The thread you remember from August about the integration that broke is buried under three hundred near-matches. A helpdesk indexes tickets with structured metadata: requester, status, channel, tags, custom fields. Search becomes a filter operation, not a fuzzy text-match. The thread from August is two clicks away. ### Failure 5: no audit log A customer disputes what was promised in a thread three months ago. You can pull up the email thread, but you cannot prove what state the ticket was in, who edited what when, or what internal notes were attached. For B2B customers and any regulated industry, the absence of an audit log is a deal-breaker. A helpdesk stamps every action: status change, assignee change, internal note, message edit. The log is exportable and retained for the contractual data-retention period. ## The migration anatomy Migrating from shared Gmail to a helpdesk is the easiest helpdesk migration there is. There is no source ticket schema, no SLA policy import, no macro library to translate. It is a fresh start with one piece of integration: forwarding the support email address into the new helpdesk. The actual sequence: 1. Create the helpdesk account. Configure your team, your tier, your SLA policies (start with one: "first reply within two hours during business hours"). One day. 2. Build five to ten macro templates. Pull the most common five replies from your sent folder, paste them into the macro library, save them with names. One hour. 3. Set up the email forward. Most domain providers let you forward `support@yourdomain.com` to the unique inbound address the helpdesk gives you. New inbound emails now create tickets in the helpdesk. Existing threads stay in Gmail. One hour. 4. Run in parallel for one week. Tickets created from new emails go through the helpdesk. Old in-flight threads finish in Gmail. The team gets used to the new interface. 5. Cut over. Disable the shared Gmail account or move it to internal-only use. Archive the old threads. Two days of focused work plus one week of parallel running. The teams that struggle are the teams that try to migrate every historic thread. Do not. They are reference material, leave them in Gmail or export to a static archive. ## What a basic helpdesk gives you on day one The five failures above all get fixed by any half-decent helpdesk. The specific things you get on day one: - A unified inbox where every channel (email, chat, WhatsApp, Instagram DM) lands as a ticket with consistent metadata. - Real-time presence and collision detection so two agents do not double-reply. - Auto-acknowledgement to the customer ("we got your message, ticket #1234, expect a reply within two hours") configurable per channel. - First-reply time and resolution time dashboards, broken down by agent, queue, and channel. - Macros so the most common replies are one click away, with personalisation tokens that pull in the customer's first name and order number. - Internal notes that stay invisible to the customer but visible to teammates picking up the thread. - A search that actually works. - An audit log that records every action. For a 3-to-8 agent team, those eight things together change the working day. The senior agent stops being a coordinator and goes back to replying. The junior agents stop colliding. The leader gets numbers to defend the team's performance with. The customer experience gets quietly better, because the small failures that used to leak through (duplicate replies, lost threads, vague SLAs) stop happening. ## What a basic helpdesk does not give you To set expectations, day one of a helpdesk does not give you: - A perfectly tuned automation. Auto-resolution and intent routing both need configuration and a few weeks of supervised operation before they pay off. - A working knowledge base. The KB takes time to write. Most teams launch with the helpdesk and add the KB in month two or three. - Reduced ticket volume. The helpdesk routes and measures, it does not deflect. Volume reduction comes from auto-resolution + KB + better product, all of which take months. - A culture change. The tool changes the workflow; the team's habits around customer empathy, escalation, and follow-up still need leadership. The honest pitch for the move is "you go from chaos to baseline competence in a week, then you build the higher-tier capabilities over the next two quarters." ## Where KimonDesk fits This is a generic operations post; the same advice applies whoever you choose. For full disclosure on our own product: - The [Free tier](/pricing) covers a one-agent team with every channel, AI, SLA, and the dashboards. There is no usage cap on tickets at the Free tier. - The Pro tier (£49 a month flat, up to 5 agents) is the natural home for teams of three to five. - Migration from Gmail is two days plus one week of parallel running. We help you set it up; the time is yours. If you are at the point where collision and search are biting weekly, the move is overdue. If you are at agent number two and not yet feeling the pain, watch for the third hire and start the move within a month of their start date. ## References - Forrester, "The Total Economic Impact of Customer Service Software for Mid-Market Teams," 2025 edition. Median first-reply-time improvement of 38 percent post-helpdesk-adoption for teams of 3 to 15 agents. - Intercom, "Conversational Support Funnel" report, 2025. Self-serve and proactive deflection benchmarks for SME teams. - Internal KimonDesk migration playbook (V2, March 2026), available on request. Read more: https://kimondesk.com/blog/why-not-gmail [truncated — see https://kimondesk.com/sitemap.xml for the full page list]