Every company that has ever emailed you a "How did we do?" survey is quietly hoping you answer. Most people don't. On a good day, maybe one in twenty clicks through and types something. Which means the mountain of opinion a business runs on - the tickets, the call transcripts, the one-star reviews, the throwaway line in a chat - mostly goes unread. Kapiche, a software company from Brisbane, decided that unread pile was the actual product.
The pitch is almost stubbornly simple. Stop treating the survey as the finish line. Feed the platform everything customers say, wherever they say it, and let an AI engine do in minutes what a research team would need weeks to attempt: find the recurring themes, gauge how people feel about each one, and attach a number to what it means for the business. The result is less a dashboard and more a translator - it turns thousands of scattered sentences into a short list of things worth fixing.
Kapiche was started in 2016 by Ryan Stuart and Kris Rogers, two software engineers who have been friends since high school and studied engineering together at the University of Queensland. Each spent roughly seven years writing software before they teamed up on the same problem. "Over 70% of the world's data isn't structured," Stuart has said, "so businesses really need to find a way to understand those opinions." That sentence is basically the whole company.
The 95% ProblemReading what everyone else throws away
Customer feedback has a survivorship problem. The people who fill out surveys are a self-selected sliver - often the delighted and the furious, rarely the quietly-drifting-away. Base decisions on them and you're steering by the loudest voices in a mostly empty room. Kapiche's core argument is that the interesting signal lives in the 95% who never answered but left traces everywhere else: the support chat, the app-store review, the cancellation call.
To do that at scale you need natural language processing that can handle messy, real-world text - misspellings, sarcasm, half-sentences - and cluster it into themes without a human pre-defining every category. That is the engineering the founders spent years building, and it is why Kapiche reads more like a research tool than a survey app. It doesn't just count keywords; it groups meaning, then measures how each theme moves metrics like NPS and CSAT.
Surveys are a rear-view mirror. Kapiche's bet is that the whole windscreen is worth reading.
The ProductsFrom raw chat to a short to-do list
The platform has grown from a text-analytics engine into a set of pieces that sit on top of it. AI Enrichments turn raw conversation into structured signals - sentiment, intent, root cause. AI Themes automatically surface what's trending across every channel and scores each theme by business impact. Ask Kapiche is the conversational layer that lets a marketer or product manager type a plain question and drill to the answer, no analyst required. And Agent QA grades support interactions automatically, turning quality assurance from a sampling exercise into full coverage.
every channel
structures it
by impact
costs the most
The through-line across all of it is impact. Plenty of tools will hand you a sentiment chart and leave you to guess what to do with it. Kapiche's design goal is to end on a ranked list: here is the theme, here is how customers feel about it, here is roughly what it is doing to your numbers. That last step is what turns a nice-to-have insights tool into something a CFO will sign off on.
The CustomersInsurers, publishers and a lift company
Kapiche is aimed at customer experience, insights, support and product teams at mid-market and enterprise organisations, with its heaviest concentration in Australia and the United States. The named list runs across industries: property-exchange network PEXA, motoring and insurance group RAC WA, health insurer nib, entertainment company Village Roadshow, academic publisher Research Square, Western Sydney University, mental-health research body SuperFriend, and lift-and-escalator giant Schindler, which serves some 400,000 customers across more than 60 countries.
The case studies tend to share a shape: a team drowning in feedback, a reporting process measured in days, and a switch to something measured in minutes. PEXA reportedly turned a monthly grind of three people over three days into a task that takes minutes. RAC WA estimated it clawed back more than 11,800 hours of manual analysis. Research Square watched its rolling NPS climb 22 points after it started acting on what the feedback actually said.
Village Roadshow went from analysing feedback in weeks to reading it in minutes.
The BusinessHow it makes money
Kapiche is a B2B software-as-a-service business. It sells annual subscriptions - offered in Professional and Enterprise tiers - to teams that own the customer relationship, with pricing quoted rather than published and scaling with data volume, users and integrations. There's a free trial to get in the door, and Google Cloud runs the infrastructure underneath. Third-party trackers have put the company's revenue in the neighbourhood of US$1.5 million a year with a team in the tens rather than the hundreds - a lean operation for the amount of data it chews through.
The money behind it is modest by Silicon Valley standards and pointed by Australian ones. An early seed round of A$1.75 million in 2019 was led by Main Sequence Ventures - the deep-tech fund with CSIRO roots - alongside the Queensland government's Business Development Fund and local backers including Brisbane Angels and River City Labs. Total disclosed funding sits under US$8 million. When they raised, the founders were not shy about the ambition: a $2 billion valuation "in the not-too-distant future."
The NeighbourhoodWhere it sits in the market
Feedback and text analytics has quietly become a crowded category. Kapiche shares the field with Thematic, Chattermill, Enterpret and the text-analytics modules bolted onto giants like Medallia, Qualtrics and Verint. The larger players sell breadth - whole experience-management suites. Kapiche's counter is focus: it does one job, reading unstructured feedback and pricing its impact, and it leans on being fast to set up and clear about the "so what."
It also has the underdog's usual advantage of being close to its users. This is a small, engineering-led team in Brisbane that describes its own reason for being as removing the uncertainty in understanding brand, customer, employee and product experiences. In a market full of platforms that produce more charts, Kapiche's pitch is almost the opposite - fewer charts, one answer. Whether that focus scales into the $2 billion dream is the open question. What is already clear is the premise it is built on: your customers have been telling you what's wrong for years, and most of it has gone unread.