AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

BMO says it used IBM quantum computers, including an IBM Quantum Nighthawk r2 processor, to test a feature-selection method for machine learning developed with Democritus University of Thrace. The bank reports that the method matched conventional techniques while using roughly one-tenth to one-hundredth as many data samples, but it has not reported a deployed banking service or measured business results.

BMO says it demonstrated a machine-learning feature-selection method on IBM quantum computers, reporting that the approach matched conventional techniques using roughly one-tenth to one-hundredth as many data samples. Developed with Democritus University of Thrace, the work explores whether quantum computing can help financial firms identify useful information in complex datasets; it is a research demonstration, not a reported banking product or client service.

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

The method adapts the Proxy Quantum Games framework to rank features in a dataset. BMO describes each feature as a “player” in a quantum game: features receive higher scores when they move with the target outcome and lower scores when they provide little useful information or repeat other features. The resulting rankings indicate which inputs may be most relevant to a machine-learning model.

BMO said the method produced results of similar quality to conventional feature-selection techniques while requiring one to two orders of magnitude fewer samples. That is the bank’s reported comparison; the announcement does not provide the datasets, detailed evaluation metrics, sample counts, or experimental results needed to independently assess the comparison. The work used IBM quantum computers, including the IBM Quantum Nighthawk r2 processor, and BMO said IBM’s quantum research team reviewed it.

The bank presented possible uses in client modeling, risk assessment and detection, and more personalized banking. Those are potential applications, not outcomes demonstrated in the announcement. BMO also said it is pursuing responsible applications of quantum technologies to artificial intelligence and machine learning, while acknowledging that application development remains at an early stage.

At a glance
announcementWhen: Announced in a BMO press release; the r…
The developmentBMO announced a quantum-computing machine-learning demonstration that it says achieved comparable feature-selection quality with substantially fewer data samples than conventional methods.
Quantum computing · Briefing

IBM quantum tests for banking machine learning

BMO says a quantum feature-selection method matched conventional techniques with far fewer data samples. The work is a research demonstration, not a deployed banking service.

10–100× fewer
data samples reported
February 2025
BMO joined IBM Quantum Network
9th
among financial institutions in 2026 Quantum Index Finance

What BMO tested

The research was developed with Democritus University of Thrace and run on IBM quantum computers, including Nighthawk r2.

1
Rank data features

The method scores inputs by how they relate to the target outcome and whether they add useful information.

2
Compare methods

BMO reports similar quality to conventional feature selection with roughly one-hundredth to one-tenth as many samples.

3
Explore banking uses

Client modeling, risk assessment and personalized banking are possible applications, not demonstrated outcomes.

Research progress, not deployment

The announcement describes a possible computational approach; it does not report a banking product or measured business results.

Reported
  • Similar feature-selection quality
  • Fewer data samples required
  • Tests on IBM quantum hardware
Still unproven
  • Improved risk or client outcomes
  • Accuracy, cost or speed gains
  • Use in customer decisions

Evidence still needed

The release leaves key details open, limiting independent assessment and practical conclusions.

Benchmarks

Show the comparison

Datasets, sample counts, evaluation metrics and conventional methods were not provided.

Validation

Test broader use

Independent replication and results across more tasks would clarify how broadly the finding applies.

Banking trials

Measure real impact

Practical trials would need to show improvements on relevant banking data and defined tasks.

CaveatBMO has not announced a deployment schedule or customer-facing application.
The takeaway
BMO reports a promising quantum research result, but technical validation and real-world banking evidence are still needed.

Potential Uses in Banking

Feature selection affects which information a machine-learning system uses to produce a prediction. If a method can identify useful inputs with fewer data samples while maintaining comparable results, it could be relevant to organizations working with complex datasets. In banking, more informative models could, in principle, support client analysis, risk decisions and tailored services.

For now, the significance is chiefly a demonstration of a possible computational approach. BMO has not said that the method is being used to make lending, fraud, or customer-service decisions, nor has it reported improvements to accuracy, cost, speed, or client outcomes. The announcement therefore points to a research direction rather than evidence that quantum computing has already changed the bank’s operations.

The result also illustrates a distinction between using quantum hardware for a task and establishing business value from that use. The release characterizes the work as showing “quantum utility,” describing it as reliable work at a scale that exact traditional computing methods cannot practically reproduce. It does not provide an independent benchmark or explain in detail how that characterization was tested. Further technical disclosure and real-world trials would be needed to judge the method’s practical value.

Products Worth Considering

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

Bose QuietComfort Wireless Noise Cancelling Headphones - Twilight Blue Limited Edition

Bose QuietComfort Wireless Noise Cancelling Headphones – Twilight Blue Limited Edition

As an affiliate, we earn on qualifying purchases.

BMO’s Quantum Research Program

BMO joined the IBM Quantum Network in February 2025, according to the announcement, and says it has invested in quantum research, talent development and applications. The bank also cited partnerships with the Chicago Quantum Exchange and Quantum Industry Canada as part of its work on financial-services applications.

BMO says its Institute for Applied Artificial Intelligence & Quantum is an enterprise-wide center focused on responsible AI governance and advancing quantum capabilities. The release also points to the bank’s position in external rankings: ninth among financial institutions in the 2026 Quantum Index Finance ranking, and recognition by Evident AI for AI use-case outcomes and responsible AI processes. These rankings describe institutional readiness and AI practices; they do not independently verify the results of this quantum feature-selection experiment.

““While we remain in the early stages of application development, these projects show how quantum computing can help solve business problems in new ways.””

— Kristin Milchanowski, BMO’s Chief AI and Quantum Officer

Evidence Still Needed for Deployment

The announcement does not include a technical paper, independent replication, or detailed test results. It leaves unspecified the datasets and sample sizes used, the exact meaning of “similar quality,” the conventional methods used for comparison, and whether the reported sample reduction held across different tasks. Without those details, readers cannot determine how broadly the result applies or compare its performance with classical alternatives.

It is also unclear what BMO means by the work operating beyond what exact traditional computing methods can practically reproduce, including the scale and methods used to support that description. The bank has not announced a deployment schedule, a customer-facing application, or measured effects on risk assessments or client services. Those questions remain open; the stated banking benefits are potential applications.

Further Testing and Banking Trials

BMO said the project is part of its wider work on quantum and machine learning, but it did not give dates for publication, follow-up experiments, or deployment. The next evidence that could clarify the result would include a technical account of the method and its benchmarks, independent testing, and results across additional datasets and machine-learning tasks.

For banking use, further steps would need to show that the approach works on relevant data and improves a defined task under practical conditions. BMO has not said whether it will conduct such trials or when a use might reach customers. Until it provides those details, the announcement supports a claim of research progress, not a confirmed change to banking services.

Key Questions

What did BMO demonstrate?

BMO says it tested a quantum method for ranking machine-learning data features on IBM quantum computers. The bank reports similar result quality to conventional feature-selection techniques with roughly one-tenth to one-hundredth as many data samples.

Does this mean BMO is using quantum computers for customer decisions?

No such deployment was announced. BMO described client modeling, risk assessment and personalized services as possible future applications, while saying its application development remains at an early stage.

What does feature selection do?

Feature selection ranks or chooses the inputs a machine-learning model uses. BMO’s method treats dataset features as players in a quantum game and scores them based on their relationship to a target outcome and whether they provide useful, non-redundant information.

Has the reported performance been independently verified?

The announcement says IBM’s quantum research team reviewed the work, but it does not provide an independent replication or enough benchmark details to verify the reported comparison. The datasets, sample counts and evaluation metrics were not specified.

Source: thequantuminsider.com

You May Also Like

What Is Real And What Is Fear In Big Tech’s AI Sell-Off?

Big Tech shares fell as AI infrastructure spending surged. Cloud growth is strong, but investors have yet to see whether returns will justify the bill.

Top 14 AI-Driven Marketing Tools To Watch In 2026

Explore the 14 leading AI-powered marketing tools set to shape 2026, highlighting features, benefits, and strategic insights for businesses.

Can A Live Preview Feed Help Sell Paid Community Access?

A proposed live preview feed would show selected community activity on creator sales pages. Its effect on membership conversions remains untested.

Marketing in the AI Era: Personalization and Customer Insights

Harness the power of AI-driven personalization and insights to transform your marketing strategy, but discover why staying ahead is essential for success.