When Statistics Canada surveyed businesses this spring, 19.2 per cent said they had used artificial intelligence to produce goods or deliver services over the previous 12 months, triple the 6.1 per cent reported two years earlier. Among those users, virtual agents and chatbots were the third most common application, at 28.2 per cent, behind data analytics and text analytics.
For customers, the change shows up as faster answers at any hour, and as pages, prices and offers arranged around each person’s own history. Both are real improvements, and each carries a catch: someone still has to stand behind a machine’s answer, and an offer built for one person is hard to compare with anyone else’s.
Online casinos show the second catch clearly, because the bonus itself can now be tailored player by player. In Ontario and Alberta, provincial rules also limit where a bonus may be shown, which keeps a tailored offer between the operator and the player it was made for.
Answers in under two minutes, not 11
Klarna, the payments provider, put numbers on the first improvement in early 2024. A month after its AI assistant went live, the company said the assistant, powered by OpenAI, had held 2.3 million conversations, two-thirds of Klarna’s customer service chats. Errands that had taken customers 11 minutes to resolve were taking under two, repeat inquiries had fallen 25 per cent, and the assistant worked around the clock in more than 35 languages.
Klarna said the assistant scored as well as its human staff on customer satisfaction, while customers could still choose a live agent. By May 2025, chief executive Sebastian Siemiatkowski was telling Bloomberg that cost had been “a too predominant evaluation factor” and that the result was “lower quality”, and Fortune reported that Klarna was recruiting human agents again. Customers, he said, had to know that “there will be always a human if you want”.
A chatbot’s answer is the company’s answer
A Canadian case tested who answers for a chatbot’s mistakes. Jake Moffatt’s grandmother died on Remembrance Day 2022, and that day Moffatt used the support chatbot on Air Canada’s website. It said a traveller who had already travelled could submit a ticket for a reduced bereavement rate “within 90 days of the date your ticket was issued”, so Moffatt bought full-fare tickets to Toronto and back for more than C$1,600.
Air Canada later said bereavement rates did not apply to completed travel. Before British Columbia’s Civil Resolution Tribunal, it argued that it could not be held liable for information from its agents, servants or representatives, a chatbot included. In effect, wrote tribunal member Christopher Rivers, the airline was suggesting the chatbot was “a separate legal entity that is responsible for its own actions.” He called that “a remarkable submission.”
In a February 2024 small-claims decision, Rivers found the airline had not taken reasonable care to make its chatbot accurate and ordered it to pay Moffatt C$812. “It makes no difference whether the information comes from a static page or a chatbot,” he wrote, because the chatbot was just a part of the airline’s website. Moffatt had also kept a screenshot of the chatbot’s words and gave it to the tribunal, a habit worth copying.
Personal prices are harder to compare
Relevance is the second improvement. Canada’s Competition Bureau took up the pricing side of it in a June 2025 discussion paper, describing algorithms that “tailor prices to individuals or groups of individuals based on various characteristics.” The Bureau said this personalized pricing, also known as surveillance pricing, may let companies target each consumer’s willingness to pay as closely as possible.
The consultation drew 103 responses, 77 of them from members of the public, and the Bureau’s January 2026 summary gave both sides. Some respondents said personalized pricing lets businesses offer targeted discounts to people who might otherwise go without. Others warned about transparency, and the summary put their point this way: “When prices are personalized, it’s harder for consumers to compare offerings across competitors.” The discussion paper had noted that people often learn about good deals through friends and family, which personal prices may make harder.
The data behind all this matters to customers too. In Cisco’s 2022 survey of 2,600 adults in 12 countries, 81 per cent agreed that the way an organization treats personal data shows how it views and respects its customers.
Casinos tailor the bonus itself
At some online casino operators, personalization now reaches the offer itself. Evoke, the parent company of William Hill, 888 and Mr Green, set out the goal in its annual report for 2024: “the right offer with the right product at the right price and at the right time.” In that year’s fourth quarter it launched a customer engagement platform “powered by intelligent automation and AI”, which it said enabled “tailored products, communications and promotions that drive retention, loyalty, and higher player value.”
Its report for 2025, in a passage on AI, credited improved segmentation and optimisation tools with “more relevant customer interactions, stronger marketing return on investment and enhanced oversight of safer gambling interventions.” Two items on that list, relevance and safer-gambling oversight, work for the player. The marketing return works for the operator, and a tailored bonus is built to serve both sides.
The operator’s side of that bargain became a public argument on September 19, when The New York Times reported, citing former employees, that DraftKings had used a measure it called “elasticity” to estimate how much more a gambler might wager if offered the right promotions. DraftKings rejected the report and told the Times its promotions are directed toward customers who demonstrate “sustained, engaged use of our platform, not toward customers based on their losses.”
Where Ontario and Alberta let a bonus appear
Ontario has kept bonus advertising out of public view since its regulated online market launched in April 2022. Under Standard 2.05 of the AGCO’s internet gaming rules, advertising that mentions a bonus is barred with two exceptions: offers shown on the gaming site the operator runs itself, and marketing sent directly to a player who has actively consented. The regulator’s guidance says bonuses themselves are not banned, and it names targeted and algorithm-based ads among the public advertising the standard rules out.
Alberta’s regulator wrote a closely matching rule before the province’s open market launched on July 13. AGLC’s advertising section, dated June 18, holds the operators and suppliers it registers to the same two exceptions, an operator’s gaming site and direct marketing that follows express player consent, and it names “algorithmic based ads” in the ban.
An AI model working inside those rules has two routes to a player: the operator’s own site, or an email, text, phone call or social media message the player signed up for. The AGCO spelled out the first route before its market opened: “Once players choose to visit an operator’s gaming site or app, inducement, bonus and credit offers may be displayed.”
|
Rule |
Ontario |
Alberta |
|
Public ads that mention a bonus |
Banned, targeted and algorithm-based ads included (Standard 2.05) |
Banned, algorithmic ads included |
|
Where a bonus offer may appear |
The operator’s own gaming site, or direct marketing after active consent |
The operator’s own gaming site, or direct marketing after express consent |
|
Withdrawing consent to bonus messages |
Possible at any time (Standard 2.07) |
Possible at any time, through a method that is easy to find |
|
Calling an offer “free” |
Only if it is free (Standard 2.06) |
Only if it is free |
|
Minimum age for casino play |
19 |
18 |
Measuring a private offer against a public one
Picture a message in a casino account’s inbox: 30 spins, no deposit needed, chosen for you, with a 50x wagering requirement and three days to use them. It may be a fair offer. The model that shaped it has studied the player’s history, though, while the player has seen nobody else’s offer.
Whether 50x and three days add up to a generous deal is the part this article can’t settle, because offers change and differ by province. The player needs a reference point that anyone can check.
Outside Ontario, that reference point exists. BonusFinder Canada keeps a public page of the best no deposit bonus casino deals it is able to list for a reader’s province, and most listings give a wagering multiple in their details panel. Its own checklist calls 30x or lower ideal and names game restrictions, expiry times and maximum withdrawals as the other terms to weigh, so a private offer can be checked against open ones.
BonusFinder Canada says it cannot advertise no-deposit bonuses in Ontario, so the page gives a player in that province no public list to set a private offer against. Standard 2.06 does at least require a private offer’s material conditions to be visible when it is first presented, with every other condition kept within one click.
The same data can flag a player in trouble
The behavioural data that shapes an offer can also spot a player in trouble, and both provinces require operators to use it that way. Ontario’s Standard 2.10 calls for a mechanism that monitors player risk profiles and behaviour for signs of harm, and the AGCO’s guidance on it uses the same words as Alberta’s standards: each operator needs “an effective mechanism for monitoring player behaviour” that proactively identifies people who may be at risk of harm, and the mechanism should draw on all available data sources to assess risk player by player.
The AGCO’s guidance and an attachment to Alberta’s standards also tell operators to use technology to scale tailored interventions to more players, rather than relying only on one-on-one contact with the highest-risk few. Their warning signs include late-night play, cancelled withdrawals and financial or emotional distress voiced to customer service, and their sample interventions include personalized emails that raise a responsible-gambling concern.
Both provinces forbid marketing that entices potentially high-risk players, and both require measures that hold back marketing to players already known to be high-risk, so a personalization engine has to know whom to leave alone. In Alberta, a request to loosen a limit can come only from the player, and nothing changes until at least 24 hours have passed.
Self-exclusion covers each market as a whole. In Ontario, adults of 19 or over can use BetGuard, a portal iGaming Ontario opened this spring, to shut themselves out of all regulated igaming sites in one step, OLG’s included. The fixed terms are six months, one year and five years, a custom term is also possible, and operators are told to keep enrolled people out of their direct marketing.
In Alberta, every operator must link its site to AGLC’s central list of self-excluded players and strip them from its marketing as soon as it is notified.
ConnexOntario, the Ontario service, connects people with gambling, addiction and mental health support for free, 24 hours a day, and answers by phone at 1-866-531-2600 as well as by text and live chat. For Albertans, the number to keep is 1-866-332-2322. It belongs to Alberta’s 24/7 Addiction Helpline, and gambling problems are among the calls it handles. Both lines are there for the moment when limits, opt-outs and self-exclusion are not enough on their own.
