Kurtis Cicalo, co-founder and CEO of Voiceflip, the Ottawa-based company behind Ardi, an AI support assistant used by about 40 multiple listing services (MLSs) in the U.S. and Canada, is attending the 4th International MLS Forum, November 16-17, 2026, at Raffles Istanbul in Istanbul, Türkiye. He is there to build connections in Europe and to learn how real estate works in markets outside North America.
“We’re always looking to expand into markets and trying to figure out what other countries are doing better than we are, and how we can adopt that,” Cicalo says. “So, yeah, here to learn.”
Founded in 2018, Voiceflip has spent about five years in organized real estate. Its product, Ardi, short for Artificial Reasoning and Decision Intelligence and represented by a blue owl, provides round-the-clock member support for MLSs, associations and brokerages. Clients include the Toronto Regional Real Estate Board (TRREB) and Canopy MLS.
The need comes from how agents work. “Real estate’s not a 9-to-5 job. Realtors generally work evenings, weekends and really strange hours,” Cicalo says. “But brokerages, MLSs and associations have a hard time staffing that, because the human cost to answer phones 24 hours a day is incredibly expensive.”
Ardi answers members’ questions through eight channels, including text message, phone, WhatsApp, Facebook Messenger, Slack, Microsoft Teams and a web widget, and in over 25 languages, even when an organization’s own materials exist only in English. If an agent wakes up at 1 a.m. realizing they forgot a form, Ardi can walk them through it.
Why Ardi Uses A Small Language Model
General-purpose AI models are trained to know about almost everything, from history to science, and Cicalo argues most of that is useless for real estate support. “It’s like taking a sledgehammer to crack a walnut open,” he says.
Ardi instead runs on a small language model trained only on each client’s own material. “A small language model is very contained in what it knows,” Cicalo says. “We only know what our clients give us.” When Ardi does not have an answer, it says so and sends the member to a person. It also does not answer legal questions or anything else a client chooses to leave out.
That simple rule is the point. “It sounds like a really dumb way of solving hallucinations, but it works well,” he says. “If you just admit you don’t know, everyone’s fine with that. It’s okay to not know something.”
He has watched organizations try the alternative. “People try to spin up a product similar to ours for their own MLS, and then they realize the answers it’s providing are more or less half hallucinations, or small truths but not quite the full truth,” Cicalo says. In the U.S. especially, that creates legal risk. “It’s the largest purchase someone’s ever going to make. You don’t want to screw that up at all. So you have to trust the model you’re using.”
Why Trusted Data Matters Across Borders
The Global Data Exchange (GDX) is a federated network, built on data standards from RESO, the Real Estate Standards Organization, that lets MLSs, associations, brokerages and technology providers share public listing data across borders. It is not a central database or a portal: each partner keeps ownership of its own listings and decides which partners can display them.
Cicalo’s first day in Istanbul showed him how different other markets are. He learned that real estate in China runs largely through social media rather than listing portals, and heard from a German attendee that buying a home there typically takes about nine months. His own last purchase in North America took four weeks.
The difference, he says, is trusted, shared data. “You only have one listing per home. There aren’t 10 different agents selling the same house. There are fewer scams because the data is trusted. That is huge,” Cicalo says. “I really love how a lot of these countries are trying to adopt the good, and maybe not the bad, from the North American markets.”
He sees a role for AI built on that kind of data in markets where consumers cannot always judge who to trust. A consumer-facing version of Ardi, he says, could let buyers check what their agent tells them. “If you can ask an AI tool that’s actually trained on real, accurate data, then you can at least trust that your agent is telling you the same thing,” he says. “You can fact-check your agent.”
Why Kurtis is Attending the 4th Annual MLS Forum in Istanbul
The feature Cicalo is most excited about is what he calls proactive support. Across its roughly 40 MLS clients, Ardi fields about 10,000 support questions a month, and Voiceflip now has two years of that data to analyze. The patterns let it predict what a member will ask next.
That changes what an MLS can do for its members. “People are always too embarrassed to ask the actual question, or express the actual problem they have,” Cicalo says. “They’ll kind of tiptoe around it, and we can find that actual problem right off the bat.” An MLS can then reach out first, offering a course or help before the member gets stuck.
For Cicalo, that is the shift support needs to make. “Instead of support just being reactive,” he says, “you can actually proactively go and help people that you know are going to need it at some point down the line.”
Cicalo is in Istanbul for the Forum, November 16-17. The Forum brings together MLS executives, brokers and technology providers from a few dozen countries. Tickets are available at mlsforum.org.
About The International MLS Forum
The International MLS Forum brings together MLS leaders, real estate organizations and technology providers from around the world to build more connected, transparent property marketplaces. The 4th edition, organized by RESO, ADRES and RealtyFeed, takes place November 16-17, 2026, at Raffles Istanbul.
This article is based on information provided by the expert source cited above. It is intended for general informational purposes only and does not constitute legal, financial, or real estate advice. Readers should conduct their own research and consult qualified professionals before making any real estate or financial decisions.
Disclosure: Individuals or companies mentioned may have a commercial relationship with KeyCrew.