We are entering an era where organizations can automate more work than ever before.
AI can analyze information, generate content, answer questions, make recommendations and increasingly take actions on our behalf. Agentic AI is taking this even further — moving AI from something we ask for information to something capable of completing transactions and executing work.
But just because we can automate something doesn’t necessarily mean we should.
That was one of the strongest messages from Julie Nestor, Executive Vice President, Marketing and Communications, Asia Pacific at Mastercard, during her presentation “Priceless in an Age of Distraction” at Something Digital 2026.
The conference described this emerging period as a “paradox era” — where technologies such as AI and quantum are advancing rapidly but are not yet fully mature, creating both an opportunity and a responsibility to keep humans at the center of how technology is designed and deployed.
For business leaders, this raises an increasingly important question:
Where should AI take over — and where should humans remain firmly in the process?
Knowing Isn’t the Same as Understanding
One of Mastercard’s first observations was deceptively simple:
Knowing is not the same as understanding.
Businesses now have extraordinary amounts of customer data. We know what people search for, what they click, what they buy, what they abandon and increasingly what they might do next.
But behavioral data doesn’t necessarily tell us why someone behaved that way.
Nestor used the example of an abandoned shopping cart. The typical response is more automation — retarget the customer, offer a discount, send another message.
But perhaps the real question isn’t:
What can we predict about this person?
It is:
What haven’t we understood about them yet?
Mastercard cited research showing that 81% of consumers ignore messages they perceive as irrelevant — despite years of increasingly sophisticated personalization.
AI can help organizations analyze more information, but empathy, curiosity and judgement still matter.
That distinction is going to become increasingly important as organizations redesign their processes around AI.
Smarter Technology Doesn’t Automatically Create Better Experiences
Another paradox Mastercard highlighted is that while technology has become smarter, people have become lonelier.
Digital connection and human connection are not necessarily the same thing.
Nestor argued that the opportunity for organizations isn’t simply to use AI to communicate with more people. Technology can instead operate behind the scenes to help create meaningful human connection.
Her example was wonderfully simple: rather than a brand pushing another promotional offer, AI could help identify people interested in participating in a local running group, cooking class or community event — using technology as the matching engine, while the actual value comes from humans connecting with humans.
It changes the role of technology.
Instead of asking:
How can AI replace this interaction?
We can ask:
How can AI make this interaction more valuable?
Sometimes Friction Is Valuable
Perhaps one of the most interesting ideas in the presentation was that sometimes friction is good.
Businesses have spent decades trying to remove friction.
Fewer clicks.
Faster transactions.
More automation.
More self-service.
Usually that makes sense.
But sometimes the thing we label as “friction” is actually part of the experience people value.
Nestor described visiting a premium restaurant where the human service experience had been replaced at the table by a QR-code ordering system.
Technically, it was efficient.
Experientially, it wasn’t.
What disappeared was the conversation with the waiter: What do you recommend? How is this cooked? What is everyone ordering?
The automation removed work — but it also removed value.
That is an important lesson for organizations adopting AI.
Efficiency cannot be the only design objective.
We also need to consider trust, experience, judgement, relationships and accountability.
Keep AI Backstage Where It Makes Sense
Mastercard’s Australian research presented another particularly interesting insight.
According to the presentation, 58% of Australians fear AI automation will prevent them from connecting with a human, while only 4% believe AI companies are worthy of their trust.
Yet Australians already interact with AI constantly through services such as Netflix and Spotify.
Nestor called this “silent AI.”
Her point was that consumers aren’t necessarily anti-technology.
They are often pro-authenticity.
Technology can work brilliantly backstage — predicting demand, reducing waste, recommending information, automating administration or identifying patterns.
But organizations should be transparent about where technology stops and humans begin.
Mastercard’s research also indicated that 74% of consumers still value in-person assistance in store and 66% value human assistance when making a purchase decision.
The challenge therefore isn’t choosing between humans and AI.
It is determining:
Which one belongs in which moment?
This Is an Operational Governance Question
This is where the conversation becomes particularly relevant to Way We Do.
As organizations introduce AI agents into their operations, we believe they need to move beyond simply asking:
“What can we automate?”
Instead, they need to understand the objective of the process, break the work into its component activities and determine the appropriate participant for each task.
That participant might be:
- A human – where judgement, empathy, accountability or relationships matter.
- An AI agent – where information needs to be retrieved, analyzed, classified, compared or prepared.
- Traditional automation – where deterministic rules can reliably execute the task and reduce the costs (no AI tokens required!).
And frequently, the answer will be a combination of all three.
Consider something as straightforward as verifying whether a supplier has adequate insurance.
The workflow might involve:
Retrieve certificate → Extract coverage → Compare against requirements → Assess exceptions → Decide compliance → Record evidence → Escalate if inadequate.
AI may be excellent at extracting information and performing the initial comparison.
But depending on the risk involved, a human may still need to make the final decision.
That is the emerging role of operational governance: establishing who — or what — is authorized to perform each part of the work, under what conditions, using what information, with what evidence and with what level of human oversight.
Trust Becomes Even More Important When AI Can Act
Mastercard’s presentation ended with perhaps the most important issue of all: trust.
Mastercard demonstrated Agent Pay, where an AI agent doesn’t simply provide information — it can complete a transaction on someone’s behalf.
Mastercard framed trust in this agentic environment around three questions:
- Intent – Did it understand what I actually wanted?
- Consent – Did I authorize it?
- Control – Do I retain agency throughout the process?
Those three questions extend far beyond payments.
They are equally relevant when an AI agent:
- approves a supplier;
- generates a customer response;
- changes a business record;
- assesses a compliance requirement;
- schedules work;
- recommends a decision; or
- takes an action within a business process.
The more authority we delegate to AI, the more important governance becomes.
Organizations will need to be able to demonstrate not simply what the AI did, but why it was authorized to do it, what rules governed its behavior, what information it relied upon and where human oversight occurred.
Designing Human-AI Workflows
The future of work isn’t likely to be humans OR AI.
It will be humans AND AI — operating together inside carefully designed processes.
And the organizations that succeed may not necessarily be those using the most AI.
They may be those that are clearest about where AI belongs.
Automate the administration. Use AI to retrieve, analyze and assist.
Keep humans involved where judgement, accountability, empathy and relationships matter.
And make the boundaries visible.
Because as Mastercard concluded, people rarely remember the technology itself. They remember the experience: the moment they felt understood, included or trusted.
That may ultimately be one of the most important principles for the age of AI:
The best technology doesn’t remove the human experience. It creates more room for it.
