close
close
BLOG

The Most Important Ingredient in AI Commerce Is Human Ambition

Tarun Chandrasekar
Tarun Chandrasekar
Chief Product Officer

What Microsoft Ignite’s focus on human ambition, intelligence and trust means for brands and retailers getting ready for Frontier Commerce

I spent this week at Microsoft Ignite surrounded, as you might expect, by a lot of AI.

Agents. Copilots. Models. Context. Reasoning. Infrastructure. More agents.

At some point, I began to wonder whether I was the only attendee without a digital twin.

But one of the ideas that stayed with me wasn’t a product announcement at all.

Satya Nadella framed this next era of technology around three deceptively simple ideas: human ambition, intelligence, and trust.

His point was that our job isn’t simply to build more capable AI. It is to turn technological progress into real leverage for people and organizations while keeping humans at the center, allowing organizations to own their intelligence and learning, and building trust into the system.

That resonated with me because, in a completely different corner of the technology world, we are wrestling with essentially the same questions at Syndigo.

Except ours involve a lot more shampoo bottles, cereal boxes, power tools and retailer attributes.

Human ambition

For years, enterprise software has mostly asked humans to adapt to software.

Open this application. Find this screen. Populate these fields. Export this file. Fix this rejection. Upload it again.

We have become remarkably skilled at turning highly capable people into very expensive swivel chairs.

AI gives us an opportunity to reverse that relationship.

That is how we’re thinking about agentic product experience management (PXM) at Syndigo.

The objective isn’t to remove people from product experience management. It is to remove as much unnecessary machinery around people as possible.

An agent can examine product information. It can find missing attributes. It can understand retailer requirements. It can recommend corrections. Increasingly, it can execute pieces of the workflow itself.

The human’s job moves upward: What are we trying to achieve? What is correct? What should we approve? What deserves judgment?

Our approach is simple: agents handle routine product-data work; people set the goals, approve the results, and remain accountable.

That feels less like replacing humans and more like finally giving them the promotion they were promised.

Intelligence

The second idea may matter even more.

The world has spent the last few years obsessing over which model is smartest.

Increasingly, I think that’s the wrong question.

What does the AI know that is uniquely true about my business?

Nadella described this as organizations building, retaining and continuously improving their own intelligence. Microsoft is expressing that through Microsoft IQ, bringing together organizational context so agents can reason over the knowledge that actually makes a company different.

In commerce, the equivalent is product intelligence.

  • What is this product?
  • What attributes describe it?
  • What claims are valid?
  • Which retailer will accept it?
  • What is missing?
  • Where is it available?
  • What do consumers say about it?
  • How should it appear on the digital shelf?
  • And increasingly: will an AI agent understand it well enough to recommend it?

That last question changes a lot.

For decades we’ve optimized product information for humans browsing shelves, websites and search results.

Now another consumer of that information has arrived: the machine.

At Syndigo, our network spans more than 15,000 brands and 3,500 retailers and distributors. Our emerging AI strategy is increasingly about turning that product and commerce knowledge into context that agents can understand and act upon.

The Syndigo Model Context Protocol (MCP) capability is one example. It exposes product data and actions as governed tools that compatible AI agents can use, so teams don’t need a custom integration for every assistant.

In the AI economy, context may become more valuable than the model consuming it.

Trust

And then comes the inconvenient part.

Giving AI an answer is relatively easy.

Giving AI permission to do something is where things get interesting.

At Ignite, Nadella argued that as intelligence becomes more powerful, we need to separate intelligence from authority.

That principle applies almost perfectly to commerce.

Imagine telling an AI agent: “Get these 10,000 products ready for all my retailers.”

The agent may be perfectly capable of determining that 1,742 records need changing.

That does not necessarily mean we should hand it the keys to the database and wish it luck.

This is why the less glamorous parts of AI suddenly become the most important: identity, permissions, auditability, deterministic rules, confidence thresholds, approvals and human escalation.

Today, Syndigo’s expanded MCP capabilities can let teams work with live product data and make governed updates through conversational AI, while proposed changes remain subject to human approval before they are written back. Because “the AI did it” is going to be a fairly unsatisfying explanation to a retailer, regulator, customer or CFO.

What we call Frontier Commerce

This leads to what we have started calling Frontier Commerce: commerce run on AI from end to end. AI that shops for consumers, and AI that runs the business behind every product. Both halves have to work.

Three simultaneous changes are creating this new era in commerce.

Commerce itself is becoming agentic. Consumers will increasingly discover, compare and buy products through AI assistants and agents.

Product experience management is becoming agentic. Agents will increasingly create, enrich, govern, validate, syndicate and optimize the information behind those products.

Enterprises themselves are becoming agentic. Humans, agents, applications and partners will increasingly work together across organizational boundaries rather than inside neatly separated software applications.

Agentic commerce will happen with or without brands and retailers. Frontier Commerce decides whether their products get chosen. That’s why Syndigo is investing in both: making product information ready for AI shopping, and Agentic PXM to run the work behind it. Plus, we’re evolving into an agentic enterprise ourselves.

But strip away all the terminology and the aspiration is surprisingly human:

  • Let people spend more time deciding what matters.
  • Let intelligence compound rather than disappear inside disconnected systems.
  • And let machines act, but only within boundaries we can understand and trust.

Which brings me back to Ignite.

AI will undoubtedly get smarter from here.

The models will improve. Agents will become more autonomous. Interfaces will disappear. There will almost certainly be several hundred new acronyms, and I accept that some of them may even be necessary.

But perhaps the ultimate measure of this era won’t be how much intelligence we manufacture.

It will be what humans are able to do with it.

The future isn’t humans versus AI.
It’s human ambition, amplified by intelligence, bounded by trust.

And if we get that right, that’s a future worth building.