Event: .NET Conf 2025 - AspNet Community Kathmandu at Sunway College
Date: 6 December 2025
Location: Sunway College, Kathmandu

I presented this topic as part of .NET Conf 2025, using it to explore a shift I have been noticing in how we build and use digital products.

Think about how people use tools like ChatGPT or Claude. You ask for a translation, and the same interface behaves like a translator. You ask it to create an image, and it behaves like an image editor. You ask for the meaning of a word, and it behaves like a dictionary.

Same interface. Different tools. The user simply expresses what they want.

That shift matters for product management and product delivery because product interfaces may no longer be designed only around screens, menus, and navigation. They may increasingly be designed around user intent.

From clicks to intent

Most software products still work through navigation.

A user opens a module, applies filters, moves to another screen, changes settings, checks the result, and then takes action. Each click is a small instruction to a different part of the system.

Now think about a healthcare administration tool.

A front-desk team needs to follow up with patients who missed appointments this week. Some patients speak English. Some speak Spanish. Some speak Nepali. Others may need another language.

In a traditional interface, the user might need to move through appointments, filters, messaging, translation settings, message templates, preview screens, and final sending.

But the user’s actual intent is much simpler:

Follow up with patients who missed appointments this week in their own language.

An AI agent could coordinate the appointment tracker, filters, messaging tool, and translation layer without forcing the user to manually move through every screen.

The old interface does not disappear

I do not think the multi-step interface disappears completely.

The future is more likely to be hybrid.

Agents may help users express outcomes. Forms, filters, tables, and review screens will still matter when users need discovery, verification, confidence, and control.

A user may be happy for an agent to prepare the follow-up messages. But they may still need to preview the patient list, check the wording, verify the language, and confirm before sending messages at scale.

That is where product design, product governance, and AI governance start to meet.

The question is not only:

Can the system do it?

It is also:

Should the system do it without review?

What changes for product teams

This changes what product teams need to think about.

Before, the main design question was often:

How do users navigate between tools?

Now, another question becomes important:

How do users express intent?

And after that:

When does the user need to see the steps before action is taken?

For Product Owners, product managers, designers, engineers, and governance teams, this means AI agents should not be treated as just another feature. They change the relationship between the user, the workflow, and the system.

A good product interface may need to support both modes:

• Guided navigation for users who are exploring or checking details.
• Intent-based commands for users who know what outcome they want.
• Review and approval points where risk, scale, or sensitivity is high.

This becomes especially important in healthcare, finance, research, education, and other data-heavy environments.

Product thinking still matters

Adding an AI agent to a product is not just a technical decision.

Product teams still need to decide what the agent can do, what tools it can coordinate, when the user must confirm, what should be logged, how errors are handled, and how people can understand what happened.

Without this thinking, an AI agent can become another confusing layer on top of an already complex product.

With the right product management approach, it can reduce friction while keeping the user in control.

The interface becomes a conversation with the system

The most interesting shift is this:

Before, users passed instructions through clicks and navigation.

Now, users may express outcomes, and the system coordinates the tools.

That does not mean every product becomes a chatbot. It means product interfaces may become more flexible. Some parts will remain structured. Some parts will become conversational. Some actions will be automated. Some will require review.

The best products will probably not choose between forms and agents. They will design the right balance between both.

For me, this is where product delivery, governance, and AI start to overlap.

The future product interface may not only be a screen users navigate. It may become a place where users express intent, review decisions, and stay in control of complex work.

---

Event page: .NET Conf 2025 - AspNet Community Kathmandu at Sunway College