Karsten Wenzlaff, Advisor
August 26th, 2025
September 21, 2026 | NCFA Market Activity | Artificial Intelligence And Data, Payments Infrastructure And Money Movement, Banking And Credit Infrastructure, Risk Compliance And Regtech

On September 21, 2026, Mastercard Canada, Toronto based Flybits and Rogers Bank announced a Mastercard Agent Pay purchase completed by an AI shopping assistant using a Rogers Red Mastercard. Mastercard describes it as Canada's first successful deployment connecting an AI commerce platform, card issuer and payment network.
The customer gave the Flybits powered assistant instructions and a spending limit. The agent found an eligible product within those conditions and completed the purchase using Mastercard Agent Pay. Flybits supplied the AI shopping experience and explainability around the selection, Rogers Bank supplied the card relationship, and Mastercard handled the payment through its network.
The AI went beyond recommending a product. It entered the payment process after the customer had already set the buying instructions and budget. Mastercard says the purchase was completed, although they didn't mention the merchant, transaction amount, exact payment credential or settlement details.
It's a Rogers Bank service that is still being tested and is not yet available to customers. Having said that, the companies describe a successful deployment and say further phases are planned. One completed transaction shows the workflow can function with a real issuer and payment network, but not yet if it can handle thousands of customers and merchants at scale including handling disputes.
Card networks already know how to authorize card payments. Agentic commerce adds a new control problem because the issuer and merchant need to know that the software had permission to make the purchase and stayed within the customer's instructions. Mastercard says the consumer's instructions were validated in this transaction and the agent operated within a predefined spending limit.
The agent wasn't given an open right to spend against the card. Mastercard hasn't disclosed whether the customer approved the exact product again before checkout, how long the authority remained active, whether permission could be revoked after the instruction was issued or how much detail the customer could inspect before payment. Those details become important once consumers allow software to make more of the buying decision for them.
Mastercard Agent Pay is built to keep the consumer, software agent and payment credential identifiable inside the transaction. That can give issuers and merchants a record of which customer granted authority, which agent acted and whether the payment stayed within the customer's conditions. Once software starts buying instead of only recommending, those records become part of the payment control.
Flybits handled the AI shopping experience. Its software interpreted the customer's request, selected a product and provided an explanation for that choice. If the purchase is disputed later, the bank may need to see what the customer asked for, why the agent chose that item and whether it stayed within the approved budget.
Mastercard also offers Agent Connect to link AI agents with merchants and payment providers. The Rogers Bank and Flybits transaction goes a step further by showing an AI agent using a Canadian issued card to complete a purchase.
Flybits also plans to use Mastercard Insight Tokens in later phases. These are designed to let an AI agent use approved customer information without giving it unrestricted access to raw data. They were not part of the September 21 transaction.
Existing card protections still apply to the payment underneath the AI experience. Rogers Bank and Mastercard continue to operate within fraud, dispute and network rules.
What happens when a customer genuinely authorizes software to spend but later argues that the agent misunderstood the request, chose the wrong product or bought something outside what the customer intended.
A customer might tell an agent to buy an eligible product for less than C$100 and receive a C$95 item that fits the budget but misses an important preference. The card credential may have been used legitimately and the amount may stay inside the approved ceiling, yet the customer can still dispute what the software decided to buy. That is very different from resolving a stolen card number or account takeover.
Today's announcement doesn't establish new liability rules for that situation. It also doesn't explain how responsibility would be divided among the consumer, Flybits, Rogers Bank, Mastercard and the merchant if valid payment credentials were used but the agent acted outside the customer's intended instructions. Earlier NCFA analysis of AI payment consent reached the same operating issue from the regulatory side: delegated software needs clear limits around what the user has actually authorized.
That makes transaction records commercially important. An issuer may need evidence showing who authenticated the request, which agent acted, what instructions were active, what spending ceiling applied and what the software ultimately purchased. Merchants need similar visibility to separate legitimate shopping agents from fraud bots and automated abuse. Mastercard Canada's Craig Reiff summarized the commercial constraint directly when he said “trust will determine how quickly these experiences scale.”
Mastercard isn't working on this problem alone. Visa is developing agent payment controls, while Stripe and Google are building infrastructure that lets software agents transact within defined permissions. Canada already has another live reference point. In July, Nuvei completed a Visa agentic commerce proof of concept using a tokenized Visa credential and live payment rails, with planned controls including Know Your Agent checks and agent risk scoring. That transaction involved a merchant AI agent, so it isn't the same consumer shopping use case as today's Rogers Bank purchase.
For Canadian banks and fintechs, the Rogers transaction brings agent authority into an operating card flow. A regulated issuer has now participated in a purchase where software interpreted customer instructions, selected a product and proceeded to payment within a spending ceiling. That gives banks something concrete to assess across fraud controls, audit records, customer consent and dispute handling.
Mastercard's August survey found that 41% of 2,000 Canadian adults were open to using AI assistants for online shopping. Among respondents who already felt positively about AI shopping assistants, 53% were open to AI assisted purchasing within preset preferences and budgets. That second figure applies to the favourable subgroup rather than all Canadians, and the research was commissioned by Mastercard.
The Rogers Bank transaction shows that an AI agent can select a product and complete a card purchase inside a customer defined spending limit. However, when an agent buys something the customer didn't expect, can the payment record prove what the customer actually authorized and who is responsible for the decision?
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
![]() | ![]() | ![]() |
|---|---|---|
![]() | ![]() | ![]() |
September 15, 2026 | NCFA Market Activity | Lending Consumer Credit And BNPL, Embedded Finance, Artificial Intelligence And Data

On September 15, 2026, Toronto-based Canadian proptech Zown updated its homebuying app with Rent Rewards alongside AI property search, affordability estimates, mortgage pre-approval and transaction services. Zown advertises up to 8% back on rent, giving it a reason to start working with consumers years before many will be ready to buy a home.
The 8% combines two potential rewards. Zown Money says Zown currently provides up to 4% cashback directly on rent, while an eligible credit card can add up to another 4% depending on the card's terms. At C$2,500 in monthly rent, Zown's 4% portion would equal C$100 a month or C$1,200 a year. If a renter also earned the full additional 4% through their card, the total could reach C$200 a month or C$2,400 a year before any card or payment-related costs.
The Canadian iPhone app, developed by Zown Realty Inc., also lets users upload a lease and proof of rent, search properties through an AI assistant called Zoro, view estimated affordability, request showings with licensed agents, seek mortgage pre-approval, submit offers and coordinate parts of closing. Zown Realty is an Ontario-registered real-estate brokerage. Mortgage rates, terms and qualification are provided through Vine Mortgage Group, and Zown says it isn't a direct mortgage lender.
Zown's model starts with a difficult Canadian problem. CMHC's 2026 Mortgage Consumer Survey found that recent buyers needed an average 4.4 years to save a down payment, while first-time buyers needed 4.7 years. Savings supplied the largest share of the down payment for 51% of first-time buyers. Another 23% of homebuyers received a financial gift, with a median gift of C$30,000. Rent Rewards give Zown a recurring reason to stay connected during those years.
Zown already has a more established incentive for buyers. Its Down Payment Boost returns up to 1.25% of a home's purchase price, capped at C$25,000, using part of the brokerage economics generated when a customer buys through Zown. On a C$1 million home, 1.25% equals C$12,500. Despite the product name, Zown's current guidance says the money arrives at closing and isn't counted as part of the mortgage down payment itself. Buyers can use it for closing and post-closing expenses.
Zown isn't alone in treating rent as financially useful activity. KOHO introduced rent cashback and credit reporting in Canada, while FrontLobby reports verified rent history to credit bureaus. Zown takes a different approach by connecting rent rewards with a later property purchase and the services surrounding it.
1. Rent is becoming a financial product. Canadian fintechs are attaching payments, rewards and credit reporting to one of the largest monthly household expenses. Toronto-based Chexy shows how quickly the category can scale. In March, the company raised C$14 million after starting with rent payments and said it had reached more than C$1 billion in annual payment volume and C$20 million in rewards value. Zown is pursuing a different end market, but rent serves the same commercial purpose of establishing a recurring financial interaction before other higher-value services are needed.
2. Housing costs are attracting more fintech models. Rent reporting, payment routing, rewards and short-term financing are competing for the same household expense. NCFA has tracked how fintech is entering rent and housing payments through companies including KOHO, Borrowell, Zenbase and Chexy. Zown adds another model by using rent rewards to encourage future homeownership and then connecting the renter to brokerage and mortgage services.
3. AI is getting closer to the financial decision. CMHC found that 16% of mortgage consumers who searched online used AI for mortgage information in 2026. HouseSigma says its Canadian platform has more than two million registered users and over five million monthly web visits, with AI used for valuation and market analysis. Zown's Zoro is competing in the same environment, focused on helping someone make a better property or financing decision. NCFA has identified the same commercial issue in decision intelligence across financial services.
It's a competitive field already. Wahi competes on digital brokerage and buyer cashback, Perch on digital mortgage readiness, FrontLobby on rent reporting, and HouseSigma on property search, valuation and market data. Zown's difference is the attempt to connect those stages much earlier, while the customer is still renting.
The economics improve if Rent Rewards keep customers engaged rather than simply subsidizing renters who eventually buy elsewhere. A traditional brokerage usually starts competing once someone begins seriously looking for a home. Zown can enter much earlier, stay connected through monthly rent, introduce affordability and mortgage tools, and eventually earn brokerage revenue if that renter buys through the platform.
That could lower the cost of finding future buyers and generate more revenue from each customer relationship. It can also become expensive if Zown funds rewards for several years without converting enough renters into completed transactions. The operating number worth watching should therefore be how many Rent Rewards users eventually become profitable Zown homebuyers?
Canadian mortgage companies, brokerages and property apps usually compete once someone is already thinking seriously about buying. If Zown can give renters enough financial value to earn their attention four or five years earlier, how much of the future homebuyer relationship can it own before the mortgage application even begins?
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
![]() | ![]() | ![]() |
|---|---|---|
![]() | ![]() | ![]() |