The Agent AI Blog

“2025 to AI Is Like 1995 to the Internet”: What It Takes to Build Agents That Work

Written by Kyle James | Jan 14, 2026 2:55:16 PM

Sitting down with Jason Burke on the PROMPTED: Builder Stories podcast felt less like a standard interview and more like a guided tour through how thoughtful builders are approaching AI right now. Jason is the founder of All Stage, an InvestTech platform focused on improving early-stage investing and fundraising. But more than that, he is a lifelong product thinker who has spent years observing how people work, where friction shows up, and how technology can remove it.

That background makes his perspective on AI agents especially compelling. Jason is not chasing novelty. He is building tools that people actually use.

A Builder and an Investor

Jason describes his initial motivation for building AI agents "selfish." He's a product builder and an investor, and he wanted to understand what agents really were and what they could realistically do. As an investor, he needed to speak the language of AI fluently. As a builder, he wanted to get his hands dirty and learn through experimentation.

That dual role gave him a rare vantage point. He understands both the excitement and the skepticism surrounding AI. More importantly, he understands the operational realities that determine whether a tool becomes useful or just interesting.

His curiosity accelerated when he began exploring Agent.ai and experimenting with agent building firsthand. What started as exploration quickly became a serious part of how he thinks about product, workflows, and decision making.

What Makes an Agent an Agent

Jason has a clear and practical definition of an AI agent. At its core, an agent is a software capability that uses multiple tools and takes action to complete a task. That framing is important. It shifts the conversation away from hype and toward utility.

Agents aren't just chat interfaces. They're systems that can gather information, analyze it, connect to other tools, and produce outcomes that would otherwise require significant manual effort. That perspective shapes everything Jason builds.

Inside All Stage, agents support founders and investors by helping them analyze investment opportunities, understand markets, identify competitors, and anticipate diligence questions. These are tasks that traditionally take weeks of research and coordination. With agents, much of that work can happen in minutes.

The goal is not to replace human judgment. The goal is to reduce cognitive load and opportunity cost so people can spend more time thinking strategically and less time assembling information.

Why This Space Needed Innovation

One of the strongest themes from Jason’s story is the lack of innovation in early-stage investing workflows. Fundraising is still dominated by static pitch decks, outdated processes, and manual effort. Founders often describe it as a necessary evil.

Jason saw that pain firsthand. As both a founder and an investor, he experienced how much time and energy was wasted preparing materials, answering repetitive questions, and managing inconsistent information. That frustration became the catalyst for All Stage.

By layering AI agents on top of the platform, Jason was able to introduce new capabilities that were previously impractical to build. Agents could analyze data more comprehensively, surface insights humans might miss, and prepare founders for conversations before they happened.

The response from users reinforced the value of this approach. Entrepreneurs consistently highlighted time savings as the biggest win. In the early days of a company, time is the most valuable resource. Anything that gives it back is meaningful.

Trust, Risk, and the Human Loop

Not all agent actions carry the same risk. Some tasks are purely informational and low stakes. Others involve decisions or actions that require a higher level of confidence.

Understanding that spectrum is critical for builders. Today, many agents excel at analysis and preparation. Over time, as trust increases and systems improve, agents will take on more autonomous action. But that progression must be intentional.

Jason’s approach emphasizes humans in the loop, especially for high-impact decisions. Agents should augment thinking, not override it. That balance is what makes them sustainable.

The Future Belongs to Builders

Jason is optimistic about where this is all heading. He believes the technical barriers to building agents are falling rapidly. What once required deep engineering expertise can now be done by curious, motivated individuals.

This democratization changes who gets to build. It opens the door for product managers, operators, founders, and even kids to create tools that improve their lives and work. The limiting factor is no longer access to technology. It's imagination and willingness to experiment.

His advice is "do not wait." Don't assume the tools need to be perfect. Start building now, even if the first version is rough. The learning comes from doing.