Six help desk vendors publish customer stories, and the numbers fall into four groups: response time, automation, satisfaction, and cost. Branch says Front cut its average email response from 5.5 hours to 1.4 hours. Pajar says Gorgias automated 45% of its tickets. Almost nobody says what else changed at the same moment.
This is independent analysis from public sources. Zendesk, Freshdesk, Intercom, Help Scout, Gorgias, and Front did not take part in it or approve it. Every result below is as published by the vendor, and SoftwareVital has not verified any of it.
How fast do support teams say they now reply?
Zendesk sells a service platform that its own site calls the complete service solution. Qualia, a digital real estate closing platform, is one of the customer stories on its site. Zendesk says Qualia cut first response times by 75 percent after it deployed Zendesk AI, Suite, and other tools in July 2024. The Qualia page hedges its own claim in the body, saying response times are “down by as much as 75 percent.” That “as much as” matters. It also says the company replaced Salesforce Service Cloud and several point tools in the same project, so the headline number belongs to a whole rebuild, and we saw no before and after figure in minutes.
Front builds shared inboxes for customer teams, and its response-time stories are the most specific we read. In the TEquipment story, a test and measurement distributor, Front says replies used to take 6 hours on average before the company set up load balancing and SLA rules, and it now advertises a 15-minute SLA. Notice that a service target is not the same as an average, and the page does not report the current average.
Front’s Branch story gives the cleanest comparison in this whole set. The insurer’s average email response time was 5.5 hours, and the page says it now averages 1.4 hours against an SLA benchmark of 2 hours, within six months of switching. Branch had come from Zendesk, so two products were swapped and nobody ran a controlled test.
Gorgias is a help desk aimed at ecommerce, and Pajar is a Canadian footwear and outerwear brand. The Pajar page says first response time fell from 15 days to 12 minutes, and resolution time from 29 days to just under 2 days. It also lists Zendesk as Pajar’s previous helpdesk. The 15-day starting point is striking, and we would want to know whether it reflects an ordinary week or a peak-season backlog, because the page says the team of five agents was stretched by winter and Black Friday.
Help Scout offers a shared inbox for support teams. In its Brain.fm story, the focus-music company’s chief operating officer reports a 95% happiness score, under an hour response time, and a 60% first email resolution over the last 60 days. That is a snapshot with no earlier figure, and it comes from a quote, not from a report.
What do the automation numbers measure?
Automation is where the vendors stop using the same words. Intercom positions its helpdesk around its AI agent, Fin. The page for Lightspeed says Fin is resolving between 45 and 65% of support volume across workspaces and is involved in 99% of conversations. Involvement and resolution are different things, and Intercom’s own tiles show both. Copilot, a separate feature, is credited with 31% more conversations closed daily by agents.
The Anthropic story is more modest, and more useful for that. Intercom says Fin reached a 50.8% resolution rate just over a month after launch, with 96% involvement and 1,700 hours saved in the first month. Its customer index shows higher numbers elsewhere, including an 80% automation rate for Kalshi and 71% Fin resolution at Vanta. We did not read those two stories in full, so treat them as headlines.
Gorgias counts automation as tickets fully resolved without a person. Pajar’s page says its 30-in-30 onboarding program aimed to resolve 30% of volume in 30 days and that Pajar reached 45%. The Cocorico story, a French apparel brand, reports 48% of tickets automated, more than 2,400 tickets closed through AI, and a 22-second average first response time, less than six months after launch. The page frames this as avoiding extra headcount. It does not say how many tickets the 48% is out of.
Does an AI agent really resolve half of support tickets?
The vendors say it can. Intercom reports 50.8% at Anthropic, Gorgias reports 48% at Cocorico, and Zendesk says Qualia’s daily ticket volume fell 30 percent. Those are three different measures: resolved by AI, automated, and total volume. A support lead comparing them side by side would be adding apples to oranges before the first meeting. Each team also chose which tickets the AI was allowed to touch. A 45% rate on a catalog of footwear questions tells you little about a software company whose tickets need account access. We found no page that reports a failed rollout, which is what you would expect from a marketing library.
A checklist for reading any support result
Before you repeat a vendor number to your finance team, we would work through these:
- Is the starting figure on the page, and does it cover the same period as the new one?
- Was the help desk the only change, or did the team also change channels, hours, staffing, or its knowledge base?
- Does “automated,” “resolved,” and “involved” mean the same thing on every page you are comparing?
- Is the number an average, a target, or a single good month?
- Did the customer previously use a competitor, and is the gain partly a fix to a badly configured old setup?
- Can you speak to that customer, or to a similar one, before signing?
Do the satisfaction numbers hold up?
Freshdesk is the help desk from Freshworks, and its stories lean on omnichannel setups and ticket routing. The Body and Fit story reports 83% CSAT, a 28% lower contact rate, and a 32% decrease in contact volume for a wellness retailer. It gives no earlier CSAT, so the 83% cannot be read as an improvement.
Branch is the only story here with a satisfaction before and after. Front says its CSAT rose from 85% to 90% in six months, and that Branch was already above the 81% it quotes for the insurance industry. We could not trace where that industry figure comes from, so we would treat it as Front’s framing. Omada Health, in a Zendesk story, lists 95% CSAT and an 85% first touch resolution rate across all channels after leaving Help Scout, again without a starting point. The page does say the move saved 2 minutes per ticket.
Help Scout’s Litmus story is the odd one out, and our favourite for that reason. The headline says Help Scout boosted retention by 26%. The body says something narrower: customers who contacted support within their first three months were retained at a 26% higher rate. That is a pattern in the data. People who ask for help may simply be the engaged ones, and the page does not claim otherwise. Litmus had been using Intercom and WordPress for its help center, which is the consolidation theme again.
What do these tools cost, according to the vendors?
Cost claims are the thinnest in the set. Front says Branch earned a 482% return on investment, driven by over $715,370 in productivity and cost savings, and saved 58% of its costs just by moving from Zendesk. No time window for the dollar figure appears on the page and no formula, so we cannot reproduce it.
The Hired story says the company saved almost $4,000 per month over the traditional license model, which depends on what license model you compare against. The same page lists an 86% first response SLA, and does not say how that compares with the period before. Zendesk’s Qualia page lists 20% cost savings in a summary tile and says nothing more about it in the text we read. The Commercial Bank of Dubai story takes the volume route to savings: monthly tickets fell from 2500 to 1800, which Freshworks ties to self-service and customer education, along with a headline claim of over 50% lower ticket resolution time.
Which help desk is best?
These stories cannot answer that, and we are not naming one. A footwear brand with five agents, a bank with a knowledge base of 500 guides, and an AI research lab are not running the same test. What the pages do show is how each vendor wants to be judged: Front on response speed, Gorgias and Intercom on automation, Freshdesk and Help Scout on satisfaction and visibility, Zendesk on a bit of everything. Pick the metric that your own leadership already reports, and ask for a trial against last quarter’s tickets.
How we chose these stories
SoftwareVital is a directory of software and AI tools, and vendors can pay us for advertising. We used only each vendor’s own site, read between one and three customer stories per vendor, and preferred recent ones with a support or productivity outcome. Dates are missing on several pages, so recency is a judgment call. We looked for outside confirmation of the figures and found none. Pages were last checked on 7 November 2026. If you work at one of these vendors and can share fuller numbers, we would be glad to update this article.