What is agentic commerce?

Commerce is entering a new phase. For the past two decades, digital commerce has been built around a familiar model: businesses present options on their own websites or apps, or via retail platforms. Customers visit those websites directly to browse, compare and transact. While the technology behind this model has evolved with the growth of e-commerce, the underlying dynamic has remained largely unchanged.

Today, however, the way customers shop is changing dramatically. In South Africa, 31-34% of the population are already active users of ChatGPT – for use cases ranging from research, to purchase decisions.

Driven by this rapidly rising use of AI, agentic commerce represents a structural shift away from this paradigm. Instead of customers actively navigating every step of a transaction, they are asking AI agents such as GhatGPT to act on their behalf – discovering options, evaluating trade-offs and going so far as to execute purchases and payments within predefined constraints.

For enterprise leaders, this is not a marginal innovation. It has the potential to reshape demand generation, customer experience, payment flows and their competitive advantage.

From traditional e-commerce to agentic commerce

Traditional e-commerce is fundamentally user-driven. The customer initiates intent, searches or browses a marketplace, evaluates products, selects a payment method and completes checkout. Even where automation exists, such as through recommendations, saved payment details and subscriptions, the customer remains the central operator of the journey. 

To win in the traditional e-commerce space meant having the ability to attract and retain those customers, primarily within an environment you control. This extends from marketing endeavors such as social media ads that drive customers to your site, to the browsing and search experience once on your site, and ultimately to the payment process.

Agentic commerce flips this relationship. In an agentic model, autonomous or semi-autonomous AI agents act on a user’s behalf. These agents are given goals, preferences and constraints, then independently decide how to fulfil them.

As pointed out by Salesforce, agentic commerce enables AI agents to “reason, decide and take action across the entire commerce journey” rather than simply responding to prompts or optimising isolated steps such as product recommendations or pricing .

The difference is not cosmetic: traditional e-commerce optimises interfaces, whereas agentic commerce optimises outcomes. To win in this space will require an entirely new playbook, driven not by an optimised user experience but an optimised agent interface.

What makes agentic commerce fundamentally different

There are three core distinctions that separate agentic commerce from earlier forms of digital commerce.

Agency shifts from the interface to the system
In traditional commerce, the customer makes every decision explicitly. In agentic commerce, decision-making is delegated. AI agents assess options, compare suppliers, monitor prices or availability and execute transactions when conditions are met.

Mastercard describes this as a move from “click-based commerce to intent-based commerce”, where the system acts once it understands the user’s objective and risk tolerance.

Journeys become continuous rather than transactional
Most e-commerce journeys are discrete: search, buy, repeat. Agentic commerce is persistent. Agents can manage replenishment, renegotiate contracts, switch suppliers or optimise delivery methods over time without requiring repeated user input.

McKinsey notes that this persistence enables agents to operate across the full lifecycle of demand, from discovery through fulfilment and post-purchase optimisation.

Decision-making becomes multi-dimensional
Rather than optimising for a single variable such as price or speed, agents can weigh trade-offs across cost, reliability, sustainability, delivery windows, loyalty benefits and risk and present these to the customer. This allows commerce to move closer to how enterprises actually make procurement decisions internally.

Why agentic commerce matters now

Agentic commerce is emerging at the intersection of several forces. AI capabilities have reached a point where systems can reason, plan and execute across complex workflows. At the same time, commerce environments have become fragmented, with more channels, more suppliers, more payment methods and more operational constraints.

Google Cloud frames agentic commerce as a response to this complexity, where AI agents reduce cognitive and operational load by coordinating decisions across systems in real time.

For consumers, this promises convenience. For enterprises, the implications are much larger.

The opportunity for enterprises

For CEOs and CFOs, agentic commerce represents an opportunity to rethink how demand is captured, how revenue is defended and how cost is controlled.

From conversion optimisation to intent fulfilment

Much of modern e-commerce investment has focused on improving conversion rates with faster checkout, better recommendations and fewer steps. Agentic commerce reframes the problem: If an agent already has permission to act, conversion becomes implicit rather than contested.

This has direct implications for marketing and loyalty. As Ad Age notes, brands that succeed in agentic environments will be those whose value propositions are legible to AI agents, not just emotionally resonant to humans .

More predictable demand and revenue

Persistent agents enable recurring and variable purchasing patterns to be managed automatically. This can stabilise demand, reduce churn and smooth cash flow, particularly in categories such as groceries, utilities, subscriptions and B2B procurement.

For finance leaders, this creates opportunities to improve forecasting accuracy and working capital efficiency.

Lower cost to serve

Agentic commerce reduces the need for constant re-engagement. Instead of repeatedly prompting customers to reorder, upgrade or renew, agents handle these decisions within agreed parameters.

McKinsey highlights that this shift could materially reduce customer acquisition and servicing costs, particularly in high-frequency or low-margin categories .

Why payments become more strategic in agentic commerce

Agentic commerce changes not just how decisions are made, but how and when money moves.

In a world where agents execute transactions autonomously, payment systems must support:

• Variable transaction amounts
• Pre-authorised spending limits
• Ongoing permissions that can be paused or revoked
• Real-time settlement and reconciliation

Traditional, one-off payment models are poorly suited to this environment. Instead, agentic commerce favours permissioned, flexible payment infrastructure that mirrors how agents operate. 

This is why bank-based payments, variable recurring payments and programmable authorisation models are gaining relevance. They allow agents to act within guardrails, rather than forcing every transaction through a rigid approval step.

Above all, payments need to be consistent and reliable; an agent may automatically search elsewhere if the transaction fails. For enterprises, payments infrastructure becomes a strategic enabler of agentic experiences rather than a downstream utility.

What leaders should be asking now

Agentic commerce is not an overnight switch. It will emerge incrementally across categories, channels and use cases. However, enterprise leaders should already be asking:

• Where in our customer journeys does delegation make sense?
• Which decisions could agents make better or faster than humans?
• Do our payment and settlement systems support variable, permissioned execution?
• Are we designing experiences for human users only, or for AI agents as well?

Those who engage early will be better positioned to shape how agentic commerce develops in their markets, rather than reacting once standards and expectations are set.

Looking ahead

Agentic commerce represents a shift from interaction-driven to outcome-driven commerce. For enterprises, it offers the potential to capture demand more efficiently, serve customers more consistently and operate with greater precision.

The winners will not be those who simply add AI features to existing journeys, but those who rethink commerce systems end to end — from intent and authorisation to payment, fulfilment and governance.

At Stitch, we’re already working toward a solution in this space that will enable our merchants to take advantage of this new trend and remain ahead of the competition. Watch this space.

‍

Frequently asked questions

What is agentic commerce?

Agentic commerce refers to a model where AI agents act on behalf of users to discover products, evaluate options and execute transactions within predefined goals and constraints.

How is agentic commerce different from traditional e-commerce?

Traditional e-commerce is user-driven and transactional. Agentic commerce is intent-driven, continuous and delegated, with AI agents making decisions rather than users navigating each step.

Why does agentic commerce matter for enterprises?

It can reduce acquisition costs, stabilise demand, improve forecasting and enable more efficient, automated purchasing experiences across high-frequency use cases.

What role do payments play in agentic commerce?

Payments must support variable amounts, ongoing permissions and programmable controls, enabling agents to transact autonomously within defined limits.

Is agentic commerce a near-term reality?

Agentic commerce will roll out gradually, but foundational capabilities are already emerging. Enterprises that prepare early will be better positioned as adoption accelerates.