The 2025 build list: AI ideas to reshape the physical economy

At Innovation Endeavors, we’re thesis-driven investors. That means we don’t just react to trends - we actively search for the white spaces: the unmet needs, messy workflows, and big technical unlocks where we believe a new company should exist.

Below are some of the questions we’re exploring at the intersection of AI and the physical economy. If you’re building in these spaces - or want to - let’s talk.

‍‍

1. Next-gen “prime” for wildfire 

There’s a growing ecosystem of point solutions for wildfire prediction, prevention, detection, suppression, and recovery. We believe there’s room for a platform layer that connects these efforts, especially for government use. Could AI sit on top to simulate, plan, or coordinate resources in real time?

2. AI x built environment

Real EstateReal estate runs on text - listings, inspections, contracts. Could LLMs streamline many of these manual workflows? 

Engineering-as-softwareStructural, civil, and MEP engineering are often outsourced. Could software automate routine design and documentation, reducing cost and time?

Interior DesignAttempts at consumer-facing design tools have fallen short. Is there an exceptional AI-enabled B2B platform for designers or prosumers - one that combines curated catalogs, budget management, visualizations, and cohesion?

Compliance & PermittingFrom environmental reviews to traffic studies to occupancy audits, compliance is a fragmented, manual world owned by consultants. How could LLMs streamline and bring these workflows in-house? 

3. AI as the new interface for Government (B2B2G)

Reporting & ComplianceWhether you’re labeling a consumer product or submitting a construction permit, government compliance is painful and underserved. What if AI could finally make it seamless?

Permitting & EntitlementsZoning, utility interconnection, and land use approvals slow down entire industries. Can AI help navigate, or even accelerate, the maze?

4. Vertical AI for real world industries

We’re continuing to invest in purpose-built AI for complex, technical, and real-world domains like manufacturing, construction, and energy. A few vectors we’re watching:

Visual & Sensor Intelligence

It seems like we’re at an inflection point in our ability to semantically parse image and sensor data. This is everything from “straightforward” security use cases on camera data to next-generation sensing capabilities like Elio.  What new capabilities become possible when machines can truly see?

Voice AgentsOrder-taking, field support, front-desk tasks - hands-free AI has huge utility. What if voice became the default interface for workers more broadly?

Trades EnablementHow can we use AI to build better tools for scheduling, inventory, compliance for trades like HVAC, electrical, and more? 

Infrastructure Management From roads to transmission lines, infrastructure has a long life. But the documentation is often brittle, unstructured, and siloed and assets are not assessed or maintained as frequently as we’d like. How can AI help us capture, understand, and maintain infrastructure better?

Back Office & AdminIt’s tough to convince customers you can do their “core” work better than they can. But, admin? That’s a different story. Companies are often eager to offload, streamline, or save on back-office tasks. Where can AI quietly take the wheel?