From AI Hallucinations to Zero Complaints: How Innovorder Deflected 67% of Support Tickets with Pluno
A conversation with
Aurélien Wel
About
Innovorder provides a complete SaaS suite to help restaurants digitize their operations: POS, kitchen display, care system, and AI-powered tools. Learn More
Industry
Restaurant Tech / SaaS
Company Size
51-200
Support Tool
Zendesk
Key Features
Deflection AI in Web chat & Whatsapp support
AI Copilot
AI Tagging & Field Filling
Innovorder is an 11-year-old French SaaS company helping restaurants digitize their entire operation, from point-of-sale and kitchen display systems to AI-powered tools for larger chains. With €12 million in turnover and 30% year-on-year growth, the company is scaling fast, now expanding into Spain, Italy, and its first major international contract with a cruise ship operator.
“We have a lot of repetitive questions, repetitive trainings, reminders of how to create a product, how to connect a printer, basic questions we see over and over.”
Aurélien Wel
Rapid growth created pressure in an unexpected place: customer support. Aurélien, who leads the 12-person support team, knew something had to change. They turned to AI, but not before a few painful false starts.
The Challenge: A Complex Product, Scattered Knowledge, and Repetitive Tickets
The support team handles around 3,000 tickets per month, with seasonal spikes in January and September. Roughly a third of those are hardware and software bugs, spread across more than 20 different product modules and multiple hardware brands. The remaining two-thirds are how-to questions and configuration requests.
Part of what makes the volume relentless is the nature of the restaurant industry itself. Staff turnover in hospitality is high, which means Innovorder's customers are always onboarding new employees who need to learn the system from scratch.
“We have a lot of repetitive questions, repetitive trainings, reminders of how to create a product, how to connect a printer, basic questions we see over and over.”
But the real challenge wasn't just volume, it was complexity.
Innovorder's suite spans 20 solutions with thousands of configuration options. The knowledge to answer any given ticket might live in Notion, in Zendesk, in a Jira issue, or buried deep in a previous support ticket.
“Agents were spending significant time just tracking down the right answer before they could even begin to write a response.”
Slow Responses and a Long Onboarding Curve
Before implementing an AI solution, response times on direct channels like chat and WhatsApp ranged from 20 minutes to over an hour.
New agents faced an even steeper problem: it took six to eight weeks before they had enough product knowledge to independently resolve tickets and write accurate, well-formed responses.
“The answers coming back weren't always complete. The time it took to find the right piece of documentation often meant agents would respond with partial information.”
A Market Full of Half-Solutions
Aurélien didn't go straight to Pluno. He and the team evaluated five or six AI solutions before finding the right fit. Each one covered only part of what they needed.
Some offered only a copilot, a tool to help agents draft responses, but had no auto-reply capability. Others had auto-reply but no copilot.
Some could only be trained on PDFs and static text files, requiring the team to manually copy their entire documentation into a new platform.
Others were website chatbot widgets that had no integration with Zendesk at all.
“There were a lot of small, narrow solutions, but none that answered all our issues. And the big one we found had a budget that was stratospheric. An American vendor quoted us $1 million a year. We're a startup, we needed something cost-effective and adaptive.”
On top of the feature gaps and cost, there was another problem that kept surfacing: hallucinations.
Several of the tools Aurélien tested produced confident but incorrect answers. Customers noticed. Complaints about the AI itself became a metric Aurélien started tracking as a signal of quality.
“If we have complaints about the AI, it's a bad AI. With every tool we tried before Pluno, customers could tell, and they complained.”
The Solution: One Tool That Connects Everything
Pluno first reached out to Aurélien on LinkedIn, initially proposing help with escalations to Jira. But after a few conversations, the scope expanded. What emerged was a full integration: AI auto-reply for customer-facing channels, an agent copilot, and a deep connection to the knowledge Innovorder already had across its tools.
Knowledge That Lives Across Many Places
One of Innovorder's core challenges was that their knowledge was distributed, and keeping a centralised, up-to-date knowledge base was an ongoing battle. Pluno solved this not by asking the team to rebuild their documentation, but by connecting to what already existed.
“Pluno can pull from Notion, from Jira, from past tickets, from internal documentation. It takes what it needs and mixes it together to propose a solution, and it also gives the agent a link to the source, so we can see if the documentation needs updating.”
The ability to train on past tickets was particularly valuable. Because Innovorder's agents solve problems in tickets every day, not in documentation, the institutional knowledge of the team was already captured in their ticket history. Pluno could surface that, even for edge cases that had never been formally documented.
“We don't always have time to write documentation. But 99% of the time we're giving solutions in tickets. Pluno finds those past tickets and uses them for new ones. We don't have to maintain a separate knowledge base the way we had to with previous tools.”
Auto-Reply That Customers Don't Complain About
The primary use case is AI auto-reply on chat and WhatsApp, the two channels where customers expect fast responses. Unlike every previous solution the team tried, Pluno's responses don't generate complaints. For Aurélien, this is the headline metric.
“With all the solutions we tried before, customers could see it was AI and they complained. With Pluno, we have not had a single complaint. That, for me, is the number one KPI.”
A Copilot That Agents Actually Use
Beyond auto-reply, the team uses Pluno as a copilot for handling complex tickets, particularly for diagnosing hardware issues and navigating multi-step software problems.
When a human agent needs to step in, Pluno provides context, suggests a diagnostic path based on similar past tickets, and hands off cleanly.
A recent feature addition has become part of Aurélien's daily workflow: a button that automatically fills in ticket fields, renames the ticket title, and generates a closing email when a ticket is resolved.
Aurélien requested it and the feature was live within a week.
Fast Setup, Fast Results
Onboarding with Pluno was straightforward from the start.
The setup took under an hour: connect Zendesk, configure some prompts and guidelines, and the system was producing auto-replies.
When Aurélien presented it to his CEO, one test question was all it took.
“My CEO tested it once, asked one question, and immediately saw that the AI was impressive. That was it.”
The responsiveness of the Pluno team during the build phase was also notable.
Requests for new features were turned around quickly, sometimes within days. and the early collaboration shaped the product into something that fit Innovorder's specific workflows.
The Results: Fewer Tickets, Faster Responses, Happier Teams
Innovorder has been using Pluno since September 2024. The measurable impact across the following six months has been significant and the team is still only using auto-reply on two channels.
67% of Tickets Resolved by AI
“We have a large number of tickets AI-resolved in one month on chat and WhatsApp alone. Email and phone channels are not yet on auto-reply. Once those are added, deflection is expected to rise further.”
Response Times: From an Hour to a Minute
“And when Pluno escalates to a human agent, it doesn't hand off a blank ticket. It provides the full conversation context, a summary, and any clarifying questions already asked, so the agent arrives ready to resolve, not re-read.”
+15 Points on CSAT in One Year
Customer satisfaction scores rose 15 points over the past year.
Aurélien is careful not to attribute all of that to Pluno alone, but credits faster and more complete responses as a meaningful contributor.
New Agents Productive in 1–3 Weeks Instead of 6–8
Perhaps the most unexpected benefit has been in onboarding. Three new agents were hired recently. Within one to three weeks, they were resolving tickets independently, a process that previously took six to eight weeks.
“Before, a new agent would wait six to eight weeks before really knowing the product and writing good answers without mistakes. Now, with Pluno alongside them, they're already resolving tickets in the first few weeks.”
Fewer Tickets Despite More Customers
Innovorder tracks a ratio that matters more to them than raw ticket volume: tickets per order processed (orders being a proxy for active customers using the platform).
25% more orders processed in 2025 vs 2024
5% fewer support tickets in 2025 vs 2024
More customers generating less support load is, as Aurélien puts it, the north star metric for a great support operation.
Pluno's auto-reply is making customers more self-sufficient, they get answers in seconds and move on without ever creating a ticket in Zendesk.
The Whole Company Uses It
Usage has spread beyond the support team. Other departments now use Pluno as an internal training tool, asking it questions about how to set up products, reconnect printers, or navigate the back office, the same kinds of questions that used to flood the support queue.
“People across the company are using Pluno like a training bot, asking it questions they'd otherwise call support for.”
Working with Pluno’s Team
When Aurélien first started working with Pluno, the collaboration was hands-on and fast-moving. Every integration request was actioned. Feature asks were turned around, sometimes overnight, sometimes within a week.
“From the very start it was a real pleasure. The team was always available to help us integrate everything. Every request I made was included. I asked for specific features and sometimes the next day, sometimes the week after, they were live. That was quite impressive.”
That responsiveness shaped the product in meaningful ways.
The initial scope was limited to Jira escalation. Over the course of early conversations, Aurélien shared his wider frustrations with copilot, auto-reply, and knowledge fragmentation, and the product expanded to address all of them.
A recent example: Aurélien asked for a button that would auto-fill ticket fields, rename the ticket title, and generate a closing email in a single click. It was in production within a week, and he now uses it every day.
“If it's a small thing and it's not on the roadmap, we get it very fast. If it's bigger, we see it in the following months. But the team always responds quickly and knows what has value.”
Today, the day-to-day contact is lighter because the tool simply runs.
But the foundation built during those early months means the team knows the product deeply enough to get the most out of it, and knows the Pluno team well enough to get fast answers when they need them.
“Test it. You'll see results on day one.”
When asked what he'd say to a support leader in a similar situation, Aurélien's answer was direct:
“You can set it up in under an hour. Connect Zendesk, add some prompts and guidelines, and you have auto-replies. If you test it for one day, you'll see results. The solution speaks for itself.”
He contrasts this with the bloated onboarding processes of other AI tools: weeks of configuration, separate knowledge base maintenance, inconsistent results. With Pluno, the results were immediate, and the quality was immediately evident.
The Real Metric: No Complaints
Aurélien's primary measure of AI quality isn't deflection rate or resolution speed. It's simpler than that: do customers complain about it?
Every previous tool failed that test. Pluno passed it.
For a company that will soon be serving international customers, including, for the first time, a major cruise ship operator, maintaining a high-quality, complaint-free AI experience isn't just a nice-to-have.
It's part of how Innovorder scales without proportionally scaling headcount.
“Given what Pluno shipped in just six months, I can't imagine what's next.”
Aurélien Wel