Supar App & SDK
Written by Jacinthe Busson, Co-founder & CPO at Fastory, based on her research and observations on mobile ecosystems, UX and Super Apps.
The most sophisticated digital experiences often feel simple.
You ask.
The system understands.
You act.
The interface disappears.
This is usually presented as progress.
Less friction.
Less navigation.
Less cognitive effort.
But when artificial intelligence begins to anticipate needs, organise decisions and automate increasingly important tasks, simplicity creates a new product question:
When does seamless UX stop being only an interface problem and become a governance problem?
Alipay and the wider Ant ecosystem provide an interesting lens through which to explore that question.
Over the past few years, Ant Group and Ant International have increasingly integrated AI into payments, healthcare, merchant services, financial operations and Super App infrastructure.
The ambition is clear: move from digital services that simply respond to user commands towards systems capable of understanding context, orchestrating tasks and taking action.
That can make digital experiences dramatically more fluid.
It can also make the systems behind those experiences dramatically more influential.
AI changes what "frictionless" means
Traditional UX optimisation removes visible obstacles.
Fewer fields.
Fewer screens.
Fewer clicks.
Faster loading.
Clearer navigation.
AI introduces another possibility.
Instead of simplifying the steps, the system can begin to perform some of the steps itself.
A user may no longer need to search through multiple services.
An AI assistant can identify the relevant information.
A merchant may no longer need to manually configure every payment workflow.
An AI agent can recommend or automate parts of the process.
A healthcare user may no longer need to interpret a complex medical report alone.
An AI system can explain it.
The distance between intention and result becomes even shorter:
Intention → AI interpretation → Action
Sometimes the visible interface can almost disappear.
This is a fundamental shift.
The product is no longer simply helping the user navigate a system.
The product begins navigating the system on behalf of the user.
Alipay's evolution makes this particularly interesting
Alipay began as a payment service.
As explored in Inside Alipay: How a Super App Became an Everyday Operating System, it progressively expanded into a broader ecosystem of services, Mini Programs and shared infrastructure.
That ecosystem already has several characteristics that make AI particularly powerful:
identity;
payments;
merchant relationships;
service access;
large-scale transaction infrastructure;
embedded third-party experiences;
and repeated everyday interactions.
AI can operate on top of this infrastructure.
This creates something very different from adding a chatbot to a standalone application.
The AI can potentially become an orchestration layer across services.
And that is where the product implications become much larger.
From assistant to agent
There is an important distinction between an AI assistant and an AI agent.
An assistant primarily responds.
Ask a question.
Receive an answer.
An agent can go further.
It can interpret an objective, coordinate multiple steps and potentially execute actions within defined boundaries.
Ant International explicitly describes its current AI strategy around agentic AI.
In June 2025, it launched the Alipay+ GenAI Cockpit, an AI-as-a-Service platform designed to help fintech companies and Super Apps create AI-agentic financial services.
The platform is positioned around workflows such as:
payment orchestration;
customer onboarding;
compliance checks;
fraud detection;
dispute resolution;
and performance optimisation.
The difference is important.
AI is no longer simply producing information.
It is beginning to participate in operational workflows.
Orchestration is becoming a UX capability
In The Power of Instantaneity, we explored how good mobile UX increasingly depends on orchestration.
The user should not need to understand every system involved in completing an action.
AI extends this principle.
Consider a complex payment problem for a merchant.
Without orchestration, the merchant may need to understand:
payment methods;
technical integrations;
risk rules;
chargebacks;
regional requirements;
routing;
and compliance.
An AI system can potentially coordinate several of these layers.
Ant International's Antom Copilot, for example, has been expanded to support merchant workflows including payment integration, onboarding, payment-method recommendations, risk configuration and chargeback management.
The visible interaction becomes simpler because the system manages more complexity behind it.
This is the same UX principle we have seen throughout this series:
The more complexity the system can manage, the less complexity the user should have to manage.
But with AI, this principle has consequences.
Because the system is no longer only hiding complexity.
It may also be making decisions inside that complexity.
Healthcare makes the stakes more visible
The implications become even clearer in healthcare.
Ant Group has invested heavily in AI healthcare services.
Its current healthcare initiatives include the AI health application AQ, launched in 2025.
Ant Group says AQ provides capabilities such as:
medical consultation support;
health-report interpretation;
health-record management;
and other AI-powered healthcare functions.
By June 2025, Ant Group reported more than 70 million users for the service.
This is an important example because healthcare exposes the limits of the simplistic idea that "less friction is always better".
A user may genuinely benefit from having a medical report translated into understandable language.
But the interface needs to communicate:
what the AI knows;
what it does not know;
where the information comes from;
what level of certainty exists;
and when human expertise is required.
In healthcare, seamlessness without transparency can become dangerous.
The same principle applies elsewhere.
The higher the stakes, the more important it becomes to make the limits of automation visible.
Good UX sometimes requires visible uncertainty
Traditional interfaces often try to look confident.
Clear answer.
Clear action.
Clear result.
AI systems operate differently.
They can be probabilistic.
They may interpret incomplete information.
They can be wrong.
They may produce different answers to similar questions.
This means that a good AI interface sometimes needs to communicate uncertainty rather than hide it.
That can feel counterintuitive.
Product teams have spent decades removing ambiguity from interfaces.
AI occasionally requires adding it back deliberately.
For example:
"This may be..."
"Based on the available information..."
"Please confirm before continuing."
"Human review required."
"Confidence is limited."
These statements introduce friction.
But it is useful friction.
The objective should not be to make the AI appear perfectly intelligent.
It should be to make the system's level of authority understandable.
Convenience increases delegation
Every time a system becomes easier to use, users can delegate more to it.
At first, delegation may be minor.
Recommend a restaurant.
Summarise a document.
Suggest a payment method.
Then it can become more significant.
Organise a journey.
Analyse financial information.
Configure risk rules.
Interpret health data.
Initiate a payment.
The user's role changes.
Instead of selecting every individual action, they define an objective and allow the system to manage more of the path.
This changes the fundamental UX question.
Traditional UX asks:
Can the user complete the task?
Agentic UX increasingly asks:
How much of the task should the system complete for the user?
That is not merely a design decision.
It is a governance decision.
Who is responsible when the AI acts?
This becomes particularly important when AI participates in transactions.
Suppose an AI agent:
selects a payment method;
chooses a merchant;
recommends a financial product;
initiates a purchase;
or configures a risk rule.
Who is responsible for the decision?
The user?
The AI provider?
The Super App?
The merchant?
The payment processor?
The underlying model provider?
The answer depends on the architecture, jurisdiction and action.
But from a product perspective, the important lesson is that responsibility needs to be reflected in the interface.
Users need to understand:
what the system is doing;
what authority it has;
what requires approval;
what can be reversed;
and who controls the final decision.
If the UX removes those distinctions in the name of simplicity, it can create confusion rather than convenience.
Agentic commerce makes this concrete
This question is becoming increasingly real in payments.
Ant International has been developing what it describes as agentic payment infrastructure.
In these models, AI agents can participate directly in commerce workflows rather than merely recommending what a user might buy.
That changes the sequence.
Traditional digital commerce might look like:
User → Search → Select → Checkout → Pay
An agentic model could increasingly look like:
User → Defines objective → AI searches and evaluates → AI proposes or initiates action → Payment infrastructure completes transaction
The number of visible steps decreases.
But the number of decisions made by the system increases.
This creates an important paradox:
The simpler the interface becomes, the more complex the trust model underneath it needs to be.
Trust becomes infrastructure
This is why AI in Super Apps cannot be treated as a standalone feature.
A system capable of orchestrating payments, services and personal information requires trust mechanisms at infrastructure level.
Ant International's AI strategy explicitly emphasises security and compliance alongside automation.
Its 2025 sustainability reporting describes investments in:
anti-money-laundering systems;
fraud detection;
risk governance;
privacy-enhancing technologies;
and AI-powered security reviews.
These systems may not be visible to the user.
But they become essential when AI participates in operational decisions.
This is another recurring principle of mature digital ecosystems:
The smoother the visible experience becomes, the more governance infrastructure is required behind it.
AI can personalise the ecosystem
Another major implication of AI inside a Super App is discovery.
In Mini Programs: The Architecture Behind Super Apps, we explored the problem of organising potentially thousands of embedded services.
A Super App cannot simply display everything.
Users need help finding what is relevant.
Search helps.
Shortcuts help.
Usage history helps.
AI can go further.
It can interpret:
context;
past behaviour;
current intent;
location;
timing;
service availability;
and potentially other signals.
The platform can therefore evolve from:
Here are all the services available
to:
Here is the service you probably need now
This could dramatically reduce navigation.
But it also increases the platform's influence over discovery.
If an AI decides which merchant appears first, which Mini Program is suggested or which financial service is recommended, the ranking logic becomes economically significant.
The AI is not simply improving UX.
It is allocating attention.
Recommendation becomes governance
This point deserves attention.
When a user manually searches through services, the platform organises access.
When an AI proactively recommends one service over another, the platform begins influencing choice more directly.
That raises questions.
Why was this service recommended?
Was it objectively more relevant?
Was it sponsored?
Did the platform favour its own service?
Was the recommendation optimised for user value?
Conversion?
Revenue?
Retention?
These are not new questions.
Search engines, marketplaces and social platforms have dealt with ranking systems for years.
But AI can make the decision-making layer less visible.
The recommendation may appear conversational.
Natural.
Personal.
Helpful.
That makes transparency even more important.
Personalisation can create invisible interfaces
Traditional interfaces expose structure.
Menus.
Categories.
Tabs.
Search.
Lists.
Personalised AI can reduce the need for some of these elements.
Instead of navigating:
Transport → Train → City → Route
a user might simply ask:
"What's the fastest way to get there?"
The AI interprets the request and orchestrates the relevant services.
This is an attractive vision.
The interface becomes more human.
But structure still exists underneath.
Services still have rules.
Transactions still have consequences.
Permissions still matter.
The interface may become invisible.
The governance cannot.
The Super App could become intent-driven
This leads to a broader product evolution.
Traditional applications are largely navigation-driven.
The user needs to understand where functionality is located.
Open section.
Find feature.
Complete flow.
Super Apps made this model broader by bringing more services into the same ecosystem.
AI can introduce another model:
intent-driven interaction.
Instead of navigating to a feature, the user expresses what they want.
"I need to get home."
"Pay this bill."
"Explain this medical result."
"Find the cheapest option."
"Resolve this chargeback."
The platform translates intent into actions across services.
This could be one of the most important transformations in mobile UX since the rise of touch interfaces.
The product hierarchy changes from:
App → Feature → Screen → Action
towards:
Intent → Orchestration → Result
But intent is not permission
There is an important distinction.
A user expressing an intention does not automatically mean they have authorised every possible action required to fulfil it.
"I need to travel to London tomorrow"
is not necessarily permission to:
purchase a ticket;
select a premium fare;
share passport details;
book a hotel;
or charge a stored payment method.
Agentic systems therefore need clear boundaries between:
intent
and
authorisation.
This boundary is a UX problem.
What can the AI do automatically?
What requires confirmation?
Which permissions persist?
Which decisions can be reversed?
When does the user need to re-enter the loop?
Designing these boundaries will become one of the most important challenges of agentic mobile experiences.
The best AI UX may involve selective friction
For years, product teams have treated friction as something to remove.
AI makes the picture more nuanced.
Some friction remains undesirable.
Repeated data entry.
Unnecessary navigation.
Manual tasks a machine can perform reliably.
But other friction becomes essential.
Confirming a high-value payment.
Reviewing sensitive information.
Approving a financial decision.
Understanding why an AI recommended something.
Escalating a medical decision to a professional.
The objective is therefore not:
AI should make everything instant.
It is:
AI should automate low-value complexity while preserving human control where judgment matters.
That distinction may become one of the defining principles of responsible agentic UX.
Super Apps have an unusual advantage
AI can exist inside any application.
But Super Apps have a particular structural advantage.
They already connect multiple services.
They already contain shared identity.
They may already contain payment infrastructure.
They already host third-party experiences.
They already coordinate discovery.
They already operate as orchestration layers.
AI can therefore amplify an architecture that already exists.
A standalone application might use AI to improve one function.
A Super App can potentially use AI to coordinate multiple functions across an ecosystem.
That makes the impact much larger.
It also makes the governance challenge much larger.
AI turns ecosystems into decision systems
This may be the most important evolution.
A conventional digital ecosystem connects services.
An AI-powered ecosystem can begin to interpret, prioritise and coordinate those services on behalf of users.
The platform therefore moves through several stages:
Application
provides a function.
↓
Super App
connects functions.
↓
AI-powered Super App
helps decide which functions should be used, when and how.
At that point, the product is no longer simply an ecosystem.
It is becoming a decision system.
That is a much more powerful position.
Governance should not be added afterwards
AI governance is often treated as a legal or compliance topic.
Something handled after the product is designed.
That approach will become increasingly difficult.
If AI controls:
recommendations;
permissions;
transactions;
personalisation;
automated workflows;
and access to third-party services,
then governance decisions directly shape the user experience.
Transparency becomes UX.
Consent becomes UX.
Explainability becomes UX.
Control becomes UX.
Escalation becomes UX.
Reversibility becomes UX.
Governance is no longer separate from product design.
It becomes part of the interface architecture.
The lesson from Alipay is not "add AI"
The superficial lesson from Ant's current AI strategy would be:
Super Apps should add AI assistants.
That misses the point.
The more interesting lesson is that AI becomes most powerful when it operates on top of existing infrastructure.
Identity.
Payments.
APIs.
Mini Programs.
Merchant services.
Security.
Data permissions.
Operational workflows.
Without this architecture, AI remains relatively isolated.
With it, AI can become an orchestration layer.
The strategic question therefore changes.
Not:
Which AI feature should we add?
But:
What should AI be allowed to orchestrate inside our ecosystem?
That is a much harder question.
And probably a much more important one.
When seamless UX becomes governance
The evolution is easy to understand.
First, digital products helped users perform tasks.
Then mobile products reduced the steps required to perform them.
Super Apps connected multiple tasks inside the same ecosystem.
AI can now begin coordinating those tasks automatically.
At every stage, the visible experience becomes simpler.
But each stage also requires more infrastructure underneath.
More permissions.
More data.
More trust.
More rules.
More responsibility.
This leads to a paradox at the centre of modern product design:
The more invisible the interface becomes, the more visible its governance needs to be.
Seamless UX cannot mean invisible decisions.
AI should remove unnecessary complexity.
But users still need to understand when a system is recommending, deciding or acting on their behalf.
That may be one of the defining challenges of the next generation of Super Apps.
Not how much AI they can integrate.
But how much autonomy they can provide without making human control disappear.
Frequently Asked Questions
How is Alipay using artificial intelligence?
Alipay and the broader Ant ecosystem use AI across areas including healthcare, merchant services, payments, risk management and financial operations.
Ant Group and Ant International have also developed AI assistants and agentic systems designed to automate increasingly complex workflows.
What is agentic AI?
Agentic AI refers to AI systems capable of moving beyond answering questions to coordinating or executing tasks towards a defined objective.
Depending on their permissions, AI agents may interact with multiple tools, systems or workflows while completing a task.
What is the Alipay+ GenAI Cockpit?
The Alipay+ GenAI Cockpit is an AI-as-a-Service platform launched by Ant International in 2025.
It is designed to help fintech companies and Super Apps build agentic and AI-native financial services across workflows including payment orchestration, onboarding, compliance, fraud detection and dispute resolution.
What is Antom Copilot?
Antom Copilot is an AI agent developed by Ant International's merchant payment business.
Its capabilities include supporting payment integration, merchant onboarding, payment-method recommendations, risk-management configuration and chargeback resolution.
What is AQ?
AQ is an AI healthcare application launched by Ant Group in 2025.
Ant Group describes it as providing capabilities including consultation support, medical-report interpretation and health-record management.
The company reported more than 70 million AQ users by June 2025.
Why is AI particularly powerful inside a Super App?
Super Apps already connect multiple services through shared infrastructure such as identity, payments, APIs and embedded experiences.
AI can potentially operate across those services, helping users discover, coordinate and complete actions without manually navigating every individual feature.
How can AI reduce friction in mobile UX?
AI can reduce friction by interpreting user intent, automating repetitive tasks, simplifying complex information and orchestrating actions across multiple services.
This can shorten the distance between what the user wants and the result they receive.
What is the difference between an AI assistant and an AI agent?
An AI assistant primarily provides information or support in response to user requests.
An AI agent can potentially take additional steps, use tools and execute actions on behalf of the user within predefined permissions and constraints.
The distinction is increasingly important when AI interacts with payments or other high-stakes services.
What is agentic commerce?
Agentic commerce describes digital commerce in which AI agents participate in activities such as searching, selecting, negotiating or initiating transactions on behalf of users.
The model requires clear rules around authorisation, payments, identity and accountability.
Can AI agents make payments?
Technically, payment systems are increasingly being developed to support agent-initiated transactions.
However, the exact level of autonomy depends on the platform, user permissions, payment rules and regulatory environment.
Clear authorisation remains essential.
Why does AI governance matter in UX?
AI systems can influence recommendations, decisions and actions.
Users therefore need to understand what the AI is doing, what authority it has, how recommendations are generated and when human confirmation is required.
These governance questions directly affect the usability and trustworthiness of the interface.
What is useful friction in AI UX?
Useful friction is a deliberate pause or confirmation that helps protect the user before an important or potentially irreversible action.
Examples include confirming a significant payment, approving access to sensitive information or reviewing a high-stakes AI recommendation.
Can AI make Super Apps easier to navigate?
Yes.
Instead of requiring users to manually navigate large catalogues of services, AI can potentially interpret user intent and suggest the most relevant service or action.
However, this also gives the platform greater influence over discovery and recommendation, making transparency important.
Could AI replace navigation in mobile apps?
Not completely, but it may reduce reliance on traditional navigation.
Intent-driven interfaces can allow users to express what they want conversationally while AI coordinates the relevant services.
Traditional navigation will likely remain important for control, exploration and transparency.
What is the main governance challenge of AI-powered Super Apps?
The central challenge is balancing automation with human control.
As AI becomes capable of coordinating more services and taking more actions, platforms need clear boundaries around consent, authorisation, transparency, responsibility and reversibility.
Sources & Further Reading
Ant International
Ant International Pushes AI Strategy with AI Platform for Fintechs, June 2025.
Official announcement of the Alipay+ GenAI Cockpit and Ant International's strategy around agentic AI, AI-native financial services, security and automated fintech workflows.
https://idocs.alipay.com/intl-website/intl-website/en/ant-international-pushes-ai-strategy-with-ai-platform-for-fintechs
Ant International
Ant International Expands Merchant Payment AI Functions with Antom Copilot 2.0, July 2025.
Documentation of Antom Copilot's AI-supported capabilities across payment integration, onboarding, payment-method recommendations, risk management and chargebacks.
https://idocs.alipay.com/intl-website/intl-website/en/ant-international-expands-merchant-payment-ai-functions-with-antom-copilot-2
Ant Group
2024 Sustainability Report highlights AI-Powered Digital Inclusion, June 2025.
Official information on Ant Group's AI healthcare strategy, AQ and AI-supported merchant services.
https://www.antgroup.com/en/news-media/press-releases/1751248800000
Ant Group
Corporate History and AI Milestones.
Official timeline documenting Ant Group's development of healthcare AI, financial AI and other AI initiatives.
https://www.antgroup.com/
Ant Group
Leadership — Junjie Zhang, President of Ant Health.
Official overview of Ant Group's healthcare development, from Alipay Future Hospital to the launch of AQ in 2025.
https://www.antgroup.com/en/about/leadership/junjie-zhang
Ant International
2025 Sustainability Report — Democratising AI and Strengthening Trust, May 2026.
Current overview of Ant International's developments in agentic payments, FinAI, AI travel assistants, AML, fraud prevention and risk governance.
https://idocs.alipay.com/intl-website/intl-website/en/ant-international-highlights-democratising-ai-and-strengthening-trust-in-2025-sustainability-report
