In our earlier blog, AI’s Impact on Supply Chain Investment Strategies, we talked about how the rapid pace of software development and easier access to domain expertise are forcing investors to rethink how they evaluate both new and existing opportunities.
That pace hasn’t slowed down—if anything, it’s accelerating. AI is evolving faster than any previous technology cycle. Innovations that used to take years are now emerging in months, reshaping what’s possible across industries.
Right now, there are three broad “flavors” of AI being discussed most often:
- Traditional AI and automation tools – rule-based systems that depend heavily on human input.
- Generative AI (GenAI) – creates new content in response to prompts, but still needs user guidance.
- Agentic AI – can plan and execute tasks on its own to reach a specific goal.
For this post, we’re focusing on Generative AI and its growing role in supply chain investing. Agentic AI—and especially the “superintelligence” version of it—is still early-stage, unproven, and not yet ready for broad use in the supply chain world.
What GenAI Brings to the Table
GenAI can create text, images, or code based on patterns it’s learned from huge amounts of data. The strongest models are built on what’s called foundation models, which are trained on massive datasets to recognize patterns that apply to many different tasks.
Customizing these models for specific use cases still takes some human expertise, but the process is now much faster and easier. APIs and prompt engineering have lowered the technical barrier significantly. The tougher part, ironically, is often domain expertise—knowing the supply chain inside and out.
That’s why we’re paying close attention to how GenAI is opening up sophisticated areas like network design, supply and demand planning, and distribution execution to a broader group of users. What once required deep algorithmic knowledge can now be done through more accessible, AI-powered interfaces.
We’ll break down how GenAI is transforming design, planning, and execution, and then look at four capability areas that are driving these changes.
How GenAI Is Changing Supply Chain Design, Planning, and Execution
1. Supply Chain Design: Moving from Static to Dynamic
Traditionally, supply chain design meant optimizing fixed networks—plants, warehouses, routes. Those models worked fine in stable conditions but could quickly fall apart when the world shifted (which, lately, it always does).
GenAI changes that. It enables dynamic, real-time design that can adapt to changes in demand, transport options, supplier performance, or even geopolitical disruptions on the fly.
We’re excited about startups building these types of AI tools—like our portfolio company Optilogic. Founders creating flexible, AI-driven network design platforms will help define the next generation of agile, decentralized, and resilient supply chains.
2. Planning: From Predictive to Prescriptive
Predictive AI—forecasting demand, inventory, or production—has been around for a while. What’s new is prescriptive AI, which doesn’t just predict what’s likely to happen but also recommends what to do about it.
Instead of just projecting demand, GenAI can now tell you how to respond—whether that means adjusting production schedules, restocking certain SKUs, or rerouting shipments.
Our portfolio company FirstShift is a good example: it’s using GenAI to help companies make smarter, faster planning decisions in real time. These kinds of prescriptive tools will drive big efficiency gains as supply chains get more complex.
3. Execution: From Automation to Autonomy
Execution is where GenAI’s potential really shines.
Right now, AI systems can automate complex workflows like logistics coordination or warehouse management. In the near future, they’ll move toward autonomy—making and executing decisions with minimal human input.
We’re already seeing it in action. Robotic systems use AI to make real-time order fulfillment decisions. GenAI tools like those from our portfolio company Drumkit.ai can reroute shipments dynamically based on traffic, weather, or political events.
Over time, these systems will go beyond simple automation to fully autonomous operations—cutting costs, improving reliability, and freeing up people for higher-level work.
Four GenAI Capabilities That Are Changing the Game
GenAI is showing up in dozens of ways, but we see four core capabilities driving the biggest impact:
- Sourcing and Managing Data
GenAI can discover, integrate, and clean up all kinds of data—traffic, weather, supplier metrics, inventory—helping teams make better, faster decisions. It can also spot obscure trends that humans might miss. - Process Automation
AI can simplify repetitive workflows and “self-heal” when things go off track. Think bots that create network scenarios, pull data from multiple systems, or chatbots that handle shipment bookings and customer inquiries automatically. - Rapid Query Formulation
Voice-enabled GenAI makes it easy to ask complex questions in plain language—like, “What’s my inventory for SKU XYZ across all channels?” That kind of quick insight helps companies react faster to shifts in demand. - Learning Algorithms
These systems don’t just predict—they learn. GenAI can remember past scenarios, refine its own models, and continuously improve forecasting accuracy. It’s still early, but this is where the long-term magic happens.
Where We’ve Invested So Far
We’ve been longtime believers in real-time supply chain visibility, especially data that lives in the public domain (not locked behind corporate firewalls). That’s why we invested in companies like Tive, Zekju, and Macropoint (acquired by Descartes).
We missed a few interesting plays—Altana and GenLogs, for example—mostly from being too cautious or just late. It happens.
We’ve also invested in process automation with companies like Drumkit and Rippey.ai (acquired by PayCargo). That space is getting crowded, though—hundreds of startups are competing for attention—so we’ve pulled back recently to wait for stronger differentiation.
Our planning bet, FirstShift, is moving fast into prescriptive AI and rapid query capabilities.
And we think learning algorithms are the next frontier. That space is still young but could be a major driver of value over the next decade. Our portfolio company Optilogic is experimenting with learning-based algorithms that outperform traditional fixed-network models.
How GenAI Is Changing the Way We Invest
The fundamentals haven’t changed—we still look at the same core things: the idea, the problem, the technology, the go-to-market, the team, and the business model.
What has changed is speed.
Building a viable supply chain tech solution used to take 18 months or more. With GenAI, that could drop to nine months—or less. That means founders need to have more of the team and infrastructure ready earlier. Go-to-market and customer support can’t wait until the product is done anymore.
Sales cycles (six to twelve months) will probably stay the same, maybe even lengthen a bit as large enterprises form AI screening task forces to sort through competing solutions. But the pace of product development and rollout will only keep accelerating.
We’re also asking new questions, like:
- Where can GenAI create the highest value for users of their tech?
- Have the founders built guardrails to prevent hallucinations and ensure consistency?
- Can the system deliver repeatable outputs—the same answer from the same input every time?
Enterprise customers demand reliability, and that’s non-negotiable.
We also want founders with deep industry experience—people who know the space cold, not those learning it as they go. Teams now need broader skills too: early GTM experience, implementation planning, and customer support readiness during development.
Looking Ahead: Agentic AI and What’s Next
Agentic AI—the kind that can plan and act on its own—will eventually find its place in supply chains. But we’re not there yet.
Supply chains are high-stakes environments. Small mistakes—like misallocating inventory or shipping the wrong order—can have big consequences. Until Agentic AI proves it can handle that complexity reliably, we’re keeping a “wait and see” approach.
Final Thoughts
The game is changing for both investors and founders.
AI will speed up everything—development, testing, deployment—but that doesn’t mean we should rush. The key will be balancing speed with sound judgment. Human insight still matters as much as ever.
For now, we’re focusing on GenAI—where the potential is massive and the applications are real.
Fun times ahead.