When an SMB buyer asks an AI assistant which help desk to buy, a shortlist comes back. We asked 44 different phrasings of that question across 2 engines, 88 answers in total, and counted every product named. This is the result. Nobody paid to be on it, and no vendor list was written in advance — the 50 vendors below are simply everyone the engines named.
Free to read, raw data included. One thing here is for sale — a $199 one-off report for a single vendor.
“Answers” is how many of the 88 answers named that product.
Share is the same number as a percentage. The two right-hand columns split it by engine.
Every row has its own link — the # beside a vendor name gives you
something like helpdesk.connexion.me/#freshdesk.
| # | Vendor | Answers | Share | ChatGPT | Perplexity | |
|---|---|---|---|---|---|---|
| 1 | Freshdesk# | 76 | 86.4% | 40 | 36 | |
| 2 | Help Scout# | 73 | 83.0% | 39 | 34 | |
| 3 | Zendesk# | 61 | 69.3% | 39 | 22 | |
| 4 | Zoho Desk# | 47 | 53.4% | 24 | 23 | |
| 5 | Intercom# | 29 | 33.0% | 24 | 5 | |
| 6 | HubSpot Service Hub# | 26 | 29.5% | 16 | 10 | |
| 7 | Front# | 25 | 28.4% | 15 | 10 | |
| 8 | Gorgias# | 22 | 25.0% | 17 | 5 | |
| 9 | Jira Service Management# | 12 | 13.6% | 9 | 3 | |
| 10 | Tidio# | 12 | 13.6% | 3 | 9 | |
| 11 | Crisp# | 8 | 9.1% | 6 | 2 | |
| 12 | Hiver# | 6 | 6.8% | 0 | 6 | |
| 13 | LiveAgent# | 6 | 6.8% | 2 | 4 | |
| 14 | osTicket# | 6 | 6.8% | 3 | 3 | |
| 15 | Salesforce Service Cloud# | 5 | 5.7% | 3 | 2 | |
| 16 | FreeScout# | 4 | 4.5% | 1 | 3 | |
| 17 | Gladly# | 4 | 4.5% | 1 | 3 | |
| 18 | Zammad# | 4 | 4.5% | 3 | 1 | |
| 19 | Featurebase# | 3 | 3.4% | 0 | 3 | |
| 20 | Kustomer# | 3 | 3.4% | 2 | 1 | |
| 21 | monday service# | 3 | 3.4% | 0 | 3 | |
| 22 | Plain# | 3 | 3.4% | 0 | 3 | |
| 23 | Groove# | 2 | 2.3% | 1 | 1 | |
| 24 | HappyFox# | 2 | 2.3% | 0 | 2 | |
| 25 | HESK# | 2 | 2.3% | 1 | 1 | |
| 26 | LiveChat# | 2 | 2.3% | 2 | 0 | |
| 27 | Missive# | 2 | 2.3% | 0 | 2 | |
| 28 | Pylon# | 2 | 2.3% | 0 | 2 | |
| 29 | Comm100# | 1 | 1.1% | 0 | 1 | |
| 30 | Desk365# | 1 | 1.1% | 0 | 1 | |
| 31 | eDesk# | 1 | 1.1% | 0 | 1 | |
| 32 | Freshworks Customer Service Suite# | 1 | 1.1% | 1 | 0 | |
| 33 | Gmelius# | 1 | 1.1% | 0 | 1 | |
| 34 | Helpable# | 1 | 1.1% | 0 | 1 | |
| 35 | Helpable Pro# | 1 | 1.1% | 0 | 1 | |
| 36 | Helply# | 1 | 1.1% | 0 | 1 | |
| 37 | Kayako# | 1 | 1.1% | 1 | 0 | |
| 38 | Klara# | 1 | 1.1% | 1 | 0 | |
| 39 | Nextiva# | 1 | 1.1% | 0 | 1 | |
| 40 | OhMD# | 1 | 1.1% | 1 | 0 | |
| 41 | OneDesk# | 1 | 1.1% | 0 | 1 | |
| 42 | ProProfs Help Desk# | 1 | 1.1% | 0 | 1 | |
| 43 | Raiseaticket# | 1 | 1.1% | 0 | 1 | |
| 44 | Re:amaze# | 1 | 1.1% | 1 | 0 | |
| 45 | RingCentral# | 1 | 1.1% | 0 | 1 | |
| 46 | Salesforce Health Cloud + Service Cloud# | 1 | 1.1% | 1 | 0 | |
| 47 | Shopify Inbox# | 1 | 1.1% | 0 | 1 | |
| 48 | SupportBee# | 1 | 1.1% | 0 | 1 | |
| 49 | Tawk.to# | 1 | 1.1% | 0 | 1 | |
| 50 | Udesk# | 1 | 1.1% | 0 | 1 |
Everything above is free and stays free. The leaderboard, the method, all 88 raw answers and the machine-readable ranking are published in full at the bottom of this page. No vendor paid to be listed, no vendor can pay to move, and nothing here goes behind a login later.
What is for sale is the diagnosis behind one vendor's row — $199, one-off, per vendor. It answers the question the table raises and cannot settle: when the engines did not name you, who did they name, and where did that come from? It contains:
It is a one-off document, not a subscription, not a seat, not a dashboard. There is no account to create and nothing renews.
To buy it: reply to lana@connexion.me with your vendor name. I send back an invoice and the report follows. Deliberately no payment details on this page — invoicing is handled one to one, by reply.
Straight about what this is: the measurement is automated end to end, which is exactly why every number is recomputable from the raw files rather than asserted. If you think your row is wrong, reply and say so — I will send the answers it was counted from and re-run any phrasing you want to give me, free.
Every claim on this page is reproducible from the files linked at the bottom.
ChatGPT (gpt-5.5), Perplexity (sonar). Each was called through its public API with a near-empty system prompt —
“answer the user's question directly and concisely” — and nothing else. The engines were
not asked to produce a list, rank anything, or name vendors. Steering them toward a list
would have manufactured the thing being measured.
44 distinct phrasings of one purchase decision, written before any measurement was run. No vendor name appears in any of them.
Each of the 88 answers was read back and every named product extracted. The vendor list is discovered: a product appears here if and only if an engine named it. Counted are customer-facing help desks, support ticketing tools, shared-inbox support tools, and broader suites whose customer-service module was named. Excluded, even when named: CRM and sales tools, standalone live-chat and chatbot widgets with no ticketing, status-page and uptime tools, incident and on-call paging tools, internal IT service desks sold only for employee IT support, knowledge-base-only tools, survey and NPS tools, call-centre phone systems, project management tools and developer issue trackers, and companies named only as an example user.
Spelling variants of one product are folded together — “Zendesk”, “Zendesk Support” and “Zendesk Suite” are one row, and a trailing “by company” is dropped. An edition name is folded into its product when the plain name is at least three times as common, so “Freshdesk Omni” counts as Freshdesk and “Zoho Desk Free” counts as Zoho Desk. Two further pairs were merged by hand, and they are the only place a human judgement touches the list — it merges duplicates, it never adds or removes a vendor:
hubspot counted as hubspotservicehubhubspotservice counted as hubspotservicehubA product is counted once per answer, however many times that answer mentions it.
A leaderboard whose top ten reshuffles between runs would not be worth publishing, so this check was run before the page was, and the page ships only because it passed.
This measures what two answer engines said on 2 August 2026, in response to these questions, through their APIs. It is not a quality ranking, not a review, and not advice on which help desk to buy. A product can be excellent and rank low here — that gap is the point.
All 88 answers, verbatim, one JSON object per line: run A · run B. The per-answer extracted product names: run A · run B. Machine-readable ranking: ranking.json.