For most of the digital commerce era, businesses have focused on designing experiences for human customers. Websites, apps and checkout journeys have been optimised to make it easier and faster for people to discover products and complete purchases. Artificial intelligence is beginning to change that assumption.
As AI tools become more capable, they are moving beyond recommendation and analysis towards autonomous execution. They can streamline purchasing by comparing products, monitoring prices, managing budgets, and initiating purchases on behalf of users.
The foundations for this shift are already in place. In 2024, a joint survey by the Bank of England and the Financial Conduct Authority found that 75% of UK financial services firms already use artificial intelligence, with a further 10% planning adoption within the next three years. As AI becomes more embedded across financial services and commerce, its role is expanding from analysis to decision-making and execution.
When AI becomes the customer
Early versions of AI-assisted purchasing are already emerging across consumer and enterprise environments. For example, platforms such as Stripe and Brex use AI to analyse transaction data, forecast cash flow, and automatically trigger payments, credit decisions, or supplier financing options in real time—reducing manual intervention in financial operations.
Many consumers already use services that monitor spending and recommend cheaper broadband, insurance or energy providers. The next step is allowing these systems to execute the switch automatically once user preferences are defined.
As these capabilities mature, merchants may increasingly find themselves selling to software agents acting on behalf of customers, rather than interacting directly with the customer themselves.
Designing commerce for machine decision-makers
This shift introduces a new challenge for digital commerce platforms. Human buyers are influenced by design, branding and marketing. AI systems evaluate different signals. They prioritise structured data such as product specifications, pricing transparency, availability and delivery terms.
For merchants, this means product catalogues and pricing data must increasingly be structured for machine consumption. Clear APIs, consistent data formats and reliable inventory information will become essential. In financial transactions, this mirrors the growing adoption of standards such as ISO 20022, which enable richer, more structured data exchange between systems. If an AI agent cannot easily interpret a product listing or compare it with alternatives, that product may simply be excluded from the decision process altogether.
In effect, merchants will need to optimise not only for search engines and human shoppers but also for AI agents performing automated product comparisons and purchasing decisions.
Implications for payments infrastructure
The rise of agent-driven commerce also has significant implications for the payment ecosystem. Most payment systems today assume a human actor at the point of transaction. Authentication mechanisms, behavioural analysis and fraud detection models are designed around patterns of individual user behaviour. AI systems behave differently.
An autonomous purchasing agent might analyse multiple merchants simultaneously, execute transactions at high speed or adjust purchasing behaviour dynamically in response to pricing changes. From the perspective of traditional fraud monitoring systems, this behaviour may resemble malicious bot activity rather than legitimate economic participation.
As a result, payments infrastructure must evolve to support authorised automated transactions. One key requirement will be stronger frameworks for delegated authority. Individuals and organisations must be able to define clear rules governing what AI systems are permitted to do.
These rules may include spending limits, merchant restrictions, time-based conditions or transaction thresholds. Instead of approving each transaction manually, users define the parameters within which automated systems can operate.
Identity systems will also need to adapt. Payment networks must be able to verify that a transaction has been initiated by an authorised AI agent that is acting on behalf of a verified individual or organisation. This will require a shift in fraud modelling such as detecting anomalous behaviour from approaches based on device fingerprinting, to verifying trusted digital identities through mechanisms such as verifiable credentials. In this model, legitimacy is established through cryptographic proof of identity and delegated authority, rather than inferred from behaviour alone.
Building trust in autonomous commerce
Trust remains the foundation of any payments ecosystem, particularly as automation increases. AI-driven transactions must operate within transparent governance frameworks that allow institutions to monitor behaviour, maintain clear audit trails and explain automated decisions where necessary.
Financial institutions and FinTech providers therefore face a dual challenge. They must support the speed and efficiency required for machine-driven commerce while ensuring that security, compliance and accountability remain intact.
This balance is particularly important in regulated markets such as the UK and Europe, where data governance and consumer protection remain central to financial innovation. It also raises questions around how the legal liability framework is distributed between the user, the agent provider and the merchant if an automated purchase goes wrong, and how frameworks such as Strong Customer Authentication will need to evolve, particularly in light of PSD3 and the EU AI Act, to support pre-authorised, agent-driven transactions.
Preparing for an agent-ready economy
Commerce has repeatedly adapted to new digital interfaces. The industry has moved from physical storefronts to websites, mobile apps and embedded financial services. Each shift has reshaped how value moves between buyers and sellers. AI agents represent the next interface layer.
Rather than simply assisting people during transactions, these systems may increasingly act on behalf of users within defined boundaries and manage parts of the purchasing process automatically.
For merchants, this means ensuring product data, pricing structures and digital platforms are accessible to automated decision systems. For FinTech providers and payment networks, it means building infrastructure capable of supporting secure machine-initiated transactions.
After decades of digital transformation in financial services, one lesson remains consistent. Innovation rarely happens all at once. Instead, it emerges gradually as technologies mature and new behaviours become normal. Agent-driven commerce is likely to follow that same path.
The organisations that recognise this early and begin designing systems that support both human and machine participants will help define how the next generation of payments operates.
Learn more at globant.com
- Artificial Intelligence in FinTech
- Digital Payments












































































































































