Karsten Wenzlaff, Advisor
August 26th, 2025
Mar 19, 2026 | NCFA Feature | AI Finance Policy

On Jan 27 2026, the UK Financial Conduct Authority launched the Mills Review into the long term impact of AI on retail financial services to examine how AI could reshape consumers, firms, markets, and regulation through to 2030. NCFA flagged the review earlier in NCFA Weekly Fintech Intelligence Jan 24-30, 2026.
On Mar 4 2026, Innovate Finance submitted its response to the Mills Review, setting out where the UK fintech industry believes deployment will stall unless policy and infrastructure move faster. The paper cites Bank of England and FCA data showing 75% of firms now use AI, up from 58% in 2022. The issue is no longer whether AI adoption will happen. It is what still blocks firms from using AI inside live financial workflows at scale.
AI is clearly evolving from chatbot assistance to execution at scale. The submission describes AI agentic systems that can act on behalf of users. One example is an AI bot that handles everything from comparing mortgage deals to submitting the application and coordinating with conveyancers under user permission.
The value is no longer only in the model itself (ie. speed, quality, cost, expertise), but rather the full operational chain from customer permission to data access to execution to payment. That's why industry is focused on a stacked layer of tech solutions from Open Finance and Digital ID to payment access and rulebook friction to ensure AI can fully complete financial tasks.
The same logic applies to industry concerns over gatekeepers. As AI agents begin to initiate and route transactions, control moves to the layer that connects the agent to the payment method and the financial product. If that layer becomes concentrated, a small number of providers can influence access, routing, and competition.
The response uses real commercial examples and market data, highlighting that AI in finance is already underwriting, trading, core banking, and compliance.
The next phase isn't whether or not firms can build AI tools. It is whether regulation and infrastructure will allow them to use those tools in broader customer and transaction flows.
The stronger points made is that AI in finance won't scale on model quality alone. It will however scale on the stack around the model. That means smart data, Open Finance, Digital ID, fraud data sharing, wallet infrastructure, and payment access. Without those layers, AI stays stuck in narrow support roles. With them, it can move into lending, advice, payments, and automated execution.
That is why the response is more useful than another generic values and ethics based AI policy statement. It identifies where deployment slows, where control could become concentrated, and what has to move together if the UK wants AI to scale significantly inside financial services.
Canada is also building its next AI strategy. The federal government launched an AI Strategy Task Force in September 2025 as part of a 30 day national sprint, and later said it heard from more than 11,000 Canadians and 28 task force members. The Canadian process is broad. It is focused on national AI leadership, trust, safety, adoption, and public interest.
That broad approach is already raising execution questions. NCFA covered this earlier in its analysis of Canada’s AI strategy and capital flight risk, which argued that deployment, investment, and commercialization need clearer direction.
The UK industry response to the Mills Review is more targeted. It focuses on what is blocking AI deployment inside financial services today. Open Finance, Digital ID, payment access, wallets, third party model assurance, and rulebook friction sit at the center of that response.
Canada is still discussing the national direction of AI while UK fintech industry is already laying out what has to change for AI to work inside live financial workflows. The lesson for Canada is straightforward. AI policy cannot move on its own. Open Finance, Digital ID, wallet policy, payments modernization, and data access frameworks need to move with it or adoption in regulated finance will stay limited.
There is also a market structure lesson. If agent led payments grow, whoever controls the interface between the agent, the wallet, and the payment rail can control distribution. Policymakers who want competition and innovation to hold need to keep that layer open.
The industry response to the Mills Review is not just a call for clearer AI rules. It argues that the next barrier sits outside the model. Data access, identity, payments, and regulatory clarity now decide whether AI in finance stays at the support layer or moves into execution. The firms and jurisdictions that solve those bottlenecks will have the advantage.
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
![]() | ![]() | ![]() |
|---|---|---|
![]() | ![]() | ![]() |
Mar 19, 2026 | NCFA Fintech Market Activity | Payments Operations And AI

On Mar 12 2026, Razorpay launched Agent Studio, a product it says lets businesses build and run AI agents across payment and post payment workflows. Razorpay says the system is built on Anthropic’s Claude Agent SDK.
Based in Bengaluru, India, Razorpay was founded in 2014 and has grown into one of India’s largest payment platforms. The company says it supports millions of businesses, reaches more than 300 million end consumers, and processes about $180 billion in annualized payment volume. Its stack spans payment acceptance, processing, disbursements, payouts, and business banking through RazorpayX.
Agent Studio targets the operational layer that sits around those flows. In its launch post, Razorpay lists agents for dispute response, subscription recovery, abandoned cart follow up through WhatsApp or email, settlement summaries delivered through messaging, and cashflow forecasting over a 3 to 7 day window. It also includes tools for cash on delivery orders, such as identifying and analyzing returns sent back to the seller.
These workflows cost merchants and payment teams money every day. Disputes create losses and take time to resolve. Failed recurring payments reduce revenue. Abandoned carts lower completed purchases. Cash on delivery returns add shipping and handling costs. Razorpay is building automation around these problem areas, not just the payment itself.
The company is also introducing a no code agent builder in beta and plans to open the system to third party agents. That points to a setup where merchants choose automation for specific tasks across the payment lifecycle, instead of relying only on bundled platform features.
This extends Razorpay’s recent work on AI driven payments, including its earlier work with NPCI on agentic payments to enable AI driven transaction flows. The focus now moves to what happens after a payment, where recovery, customer support, and problem cases still rely heavily on manual work.
Acceptance rates and pricing are still important for PSPs and merchant platforms. Operational efficiency is becoming part of the core value proposition, especially where it directly improves recovery and reduces cost.
If agents handle disputes, recovery, reporting, and cash on delivery losses, does competition start to favour the provider that cuts the most cost for merchants, not just the one with the lowest processing price?
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
![]() | ![]() | ![]() |
|---|---|---|
![]() | ![]() | ![]() |
Mar 17, 2026 | NCFA Market Insight | AI And Operating Model Reset

On Mar 16 2026, reports that Meta may cut 20% or more of its workforce alongside plans for up to $135 billion in AI related capital spending point to an emerging pattern in how large technology firms operate in the era of AI. Companies are increasing AI investment while reducing headcount and restructuring teams to improve efficiency. Multiple large firms now follow the same pattern.
The sequence is becoming consistent across large technology and fintech firms, and the playbook is as follows.
Increase AI and infrastructure spend.
Reduce or flatten headcount.
Reassign work to automation.
Improve margins and operating leverage.
Rinse and repeat as necessary
Markets reward this because cost discipline becomes visible.
Meta is not alone.
While expanding AI infrastructure, Amazon has cut several hundred roles across AWS and other units.
Google cut hundreds of roles across Assistant, hardware, and engineering teams as it reallocates resources toward AI priorities.
Microsoft cut about 3% of its workforce while continuing to increase AI investment across cloud and enterprise products.
Dominoes? AI spending rises while workforce structures tighten around a new cost model.
And last month, fintech showed the same logic.
Block cut more than 4000 roles in an AI led cost reset while repositioning around automation and efficiency. Investors now expect fintech firms to translate AI into margin improvement, not just new features.
Payments, lending, and financial operations are structured workflows, which makes them easier to redesign around automation.
Fewer headlines at scale but the playbook is apparent.
Shopify requires teams to justify hiring against AI capability. Teams must show AI cannot do the work before adding people.
The new operating model mirrors fewer hires, higher output per employee, and clearer cost control.
The immediate impact of AI is not revenue growth. It is cost structure.
AI increases capital spending while reducing labour intensity, which alters operating leverage across digital businesses.
In fintech, this shows up as:
* automated fraud detection replacing manual review
* automated underwriting reducing analyst workload
* AI support systems replacing large service teams
The cost per decision falls. That is the first advantage.
This pattern will likely not stop at just large tech firms.
Public fintech and payments companies are now under similar pressure to show that AI improves efficiency.
Firms with large cost bases and heavy reliance on manual processes are the most exposed. That includes payments platforms, neobanks, and global processors where margins depend on operational scale.
Companies such as PayPal, Adyen, and other listed payment firms face imminent change. Investors will look for evidence that AI reduces cost per transaction, improves fraud outcomes, and increases operating leverage.
The question is no longer whether these firms use AI. It is whether AI measurably improves their cost structure.
Meta contemplating mass job cuts at scale further highlight a clear cost reset that's defining how digital businesses operate. Amazon, Google, and Microsoft reinforce the pattern. Block shows how it applies in fintech. Shopify demonstrates how it's taking up mental space and decisioning in Canada.
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
![]() | ![]() | ![]() |
|---|---|---|
![]() | ![]() | ![]() |
Mar 17, 2026 | NCFA Fintech Market Activity | AI Infrastructure And Capital Markets

On Mar 16 2026, Nvidia CEO Jensen Huang delivered an opening keynote at GTC (GPU Technology Conference in San Jose, California) raising the company’s data center revenue outlook to $1 trillion for 2025 through 2027, up from a prior $500 billion estimate. The revision reflects accelerating demand for AI infrastructure as the industry moves from model training into continuous inference at scale.
This is not just a bigger number. It also points to where value will accumulate in AI and essentially who will pay for it.
The industry is moving past model training into continuous inference, so everything picks up from here. Growth isn't in experimentation alone, it's in scaling production.
Inference is not a one time cost. It runs every time a model answers a question, scores a transaction, flags fraud, or generates a recommendation. It scales with usage, not development.
Nvidia’s roadmap, including Blackwell and next generation systems, targets that reality.
For fintech, it directly relies on AI for its operating model.
AI driven underwriting, fraud detection, support agents, and personalization are no longer batch processes. They run live, at scale, on every user interaction.
That means:
* compute becomes a recurring cost tied to usage
* margins depend on inference efficiency, not just acquisition
* product design starts to include cost per decision, not just conversion
These operational costs represent a structural shift. It's not a one and done feature upgrade.
A $1 trillion infrastructure market doesn't stay fragmented.
It concentrates.
Compute, models, and distribution stack together. The firms that control inference capacity influence pricing, latency, and access.
Fintech doesn't compete with Nvidia directly. But it builds on top of the stack Nvidia helps define.
That creates a dependency layer many firms have not priced in yet.
The new forecast raises expectations across the market.
Capital now asks harder questions:
* where does AI revenue actually show up
* who captures margin versus who absorbs cost
* which use cases justify continuous compute spend
This pressure is already visible across public tech names. Strong AI narratives no longer carry valuations on their own. Unit economics are key yet again.
This flows straight into checkout, risk, and customer experience.
Payments firms, neobanks, and platforms are all moving toward:
* real time fraud scoring
* dynamic pricing and routing
* AI driven support and engagement
Those features increase conversion and reduce risk. But they also increase compute intensity per transaction.
That creates a new tradeoff: better decisions versus higher cost per interaction.
Firms that optimize that balance get it right for now. Others compress margins without realizing it.
AI is a capability And cost layer.
Fintechs and financial institutions should consider to:
* design products with inference cost in mind from day one
* choose infrastructure partners strategically, not just technically
* focus on high value decisions where AI materially improves outcomes
The risk of course is adding AI features that increase cost faster than revenue.
If AI infrastructure becomes a $1 trillion market, fintech’s advantage won't come from using AI first. It will come from using it efficiently.
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
![]() | ![]() | ![]() |
|---|---|---|
![]() | ![]() | ![]() |