What are the key takeaways from this a16z Podcast episode?

Key insights include: **Trust-building ladder for autonomous agents:** No enterprise deploys fully autonomous negotiation agents on day one. LEO starts with human-in-the-loop oversight on small transactions, then progressively earns autonomy at $10K, $20K, and $100K negotiation thresholds. This staged approach simultaneously builds customer trust and feeds the agent with company-specific feedback that improves performance over time.; **The 80/20 exception problem:** Internal AI builds typically reach 70-80% task performance, but that threshold rarely translates to meaningful automation. At 70% accuracy, humans must still review 100% of outputs, potentially creating more work than before. The final 20% of performance requires vertical-specific data, integrations, and exception-handling logic that takes months to build correctly.; **Procurement's hidden work surface:** ERP systems like SAP only record final outcomes — a price of $8K for aluminum. The actual work happens across hundreds of emails, 20 spreadsheets, cost engineering models, and supplier negotiations that are completely invisible to the system of record. AI-native startups capture this unstructured context as a proprietary data asset incumbents cannot replicate.

What did Seema Amble and Vlad Keil discuss on a16z Podcast?

A16z partner Seema Amble and LEO co-founder Vlad Keil examine why AI-native startups can outcompete incumbents in enterprise procurement by owning end-to-end workflows across emails, contracts, supplier negotiations, and engineering data that no single system of record captures. Key topics include: **Trust-building ladder for autonomous agents:** No enterprise deploys fully autonomous negotiation agents on day one. LEO starts with human-in-the-loop oversight on small transactions, then progressively earns autonomy at $10K, $20K, and $100K negotiation thresholds. This staged approach simultaneously builds customer trust and feeds the agent with company-specific feedback that improves performance over time.; **The 80/20 exception problem:** Internal AI builds typically reach 70-80% task performance, but that threshold rarely translates to meaningful automation. At 70% accuracy, humans must still review 100% of outputs, potentially creating more work than before. The final 20% of performance requires vertical-specific data, integrations, and exception-handling logic that takes months to build correctly..

How long is this episode of a16z Podcast?

This episode is 59 minutes long. SignalCast provides an AI-generated summary so you can get the key insights in about 3 minutes.

a16z Podcast

Why AI Agents Can Beat the Incumbents

October 2, 2026

59 min episode · 2 min read

Seema Amble,Vlad Keil

Episode

59 min

Read time

2 min

Topics

Relationships, Startups, Fundraising & VC

AI-Generated Summary

Published Oct 2, 2026

Key Takeaways

✓Trust-building ladder for autonomous agents: No enterprise deploys fully autonomous negotiation agents on day one. LEO starts with human-in-the-loop oversight on small transactions, then progressively earns autonomy at $10K, $20K, and $100K negotiation thresholds. This staged approach simultaneously builds customer trust and feeds the agent with company-specific feedback that improves performance over time.
✓The 80/20 exception problem: Internal AI builds typically reach 70-80% task performance, but that threshold rarely translates to meaningful automation. At 70% accuracy, humans must still review 100% of outputs, potentially creating more work than before. The final 20% of performance requires vertical-specific data, integrations, and exception-handling logic that takes months to build correctly.
✓Procurement's hidden work surface: ERP systems like SAP only record final outcomes — a price of $8K for aluminum. The actual work happens across hundreds of emails, 20 spreadsheets, cost engineering models, and supplier negotiations that are completely invisible to the system of record. AI-native startups capture this unstructured context as a proprietary data asset incumbents cannot replicate.
✓1% procurement savings equals 10% revenue impact: A 1% reduction in procurement costs produces the same P&L improvement as a 10% increase in sales revenue. For manufacturers building aircraft, data centers, or robots, a single delayed part confirmation buried in 500 emails can cause hundreds of millions in project damage — making procurement monitoring a direct financial performance lever.
✓Multi-agent architecture requirement for end-to-end ownership: Single agents cannot complete complex procurement jobs. LEO deploys coordinated multi-agent systems where separate agents handle sourcing, RFQ drafting, supplier response parsing, price benchmarking, negotiation strategy, order confirmation, and shipment tracking — each sharing context and operating in sequence to replace workflows that previously required eight people across three departments.

What It Covers

A16z partner Seema Amble and LEO co-founder Vlad Keil examine why AI-native startups can outcompete incumbents in enterprise procurement by owning end-to-end workflows across emails, contracts, supplier negotiations, and engineering data that no single system of record captures.

Key Questions Answered

•Trust-building ladder for autonomous agents: No enterprise deploys fully autonomous negotiation agents on day one. LEO starts with human-in-the-loop oversight on small transactions, then progressively earns autonomy at $10K, $20K, and $100K negotiation thresholds. This staged approach simultaneously builds customer trust and feeds the agent with company-specific feedback that improves performance over time.
•The 80/20 exception problem: Internal AI builds typically reach 70-80% task performance, but that threshold rarely translates to meaningful automation. At 70% accuracy, humans must still review 100% of outputs, potentially creating more work than before. The final 20% of performance requires vertical-specific data, integrations, and exception-handling logic that takes months to build correctly.
•Procurement's hidden work surface: ERP systems like SAP only record final outcomes — a price of $8K for aluminum. The actual work happens across hundreds of emails, 20 spreadsheets, cost engineering models, and supplier negotiations that are completely invisible to the system of record. AI-native startups capture this unstructured context as a proprietary data asset incumbents cannot replicate.
•1% procurement savings equals 10% revenue impact: A 1% reduction in procurement costs produces the same P&L improvement as a 10% increase in sales revenue. For manufacturers building aircraft, data centers, or robots, a single delayed part confirmation buried in 500 emails can cause hundreds of millions in project damage — making procurement monitoring a direct financial performance lever.
•Multi-agent architecture requirement for end-to-end ownership: Single agents cannot complete complex procurement jobs. LEO deploys coordinated multi-agent systems where separate agents handle sourcing, RFQ drafting, supplier response parsing, price benchmarking, negotiation strategy, order confirmation, and shipment tracking — each sharing context and operating in sequence to replace workflows that previously required eight people across three departments.

Notable Moment

Vlad reveals that a product LEO built as one of its first commercial offerings three years ago — extracting quote data into SAP — is now assigned as an eight-hour engineering exercise for job applicants. What was once a flagship product became a hiring test, illustrating how rapidly AI capability baselines shift.

Know someone who'd find this useful?

Episode Transcript

If you want to build an aircraft, you need to procure thousands of suppliers. Someone sends a confirmation of like, hey, sorry, like, this part is going to arrive two weeks later. And if they missed this email, hundreds of millions of them. Procurement historically may have been more in a box, And now it's like, okay, it's touching legal, it's touching finance, it's touching a bunch of different software systems and people. The opportunity for the AI native startup is to say, we're gonna own that entire end to end arc. No company and no enterprise starts with fully autonomous negotiation agents from day one. Why? Because they don't trust us and they don't trust the technology from day one. And by having this human loop approach, we are feeding our agent with all the feedback and all the learnings, and then they suddenly trust us for, like, ten k negotiations, 20 k negotiations, 100 k negotiations. When you think about what a durable vertical AI company looks like, what are the qualities that you look for? It's really, really hard to forecast your moat going forward. If you look back at all the best businesses, at the early stages, they were... If the incumbent already owns the customer, the data, and the system of record, where does an AI native startup have an advantage? In this episode, Elena Berger sits down with a 16 z partner Seema Amble and LEO co founder and CEO Vlad Keil to answer that question through one of the most complex parts of the enterprise: procurement. They get into why so much of the actual work happens outside the system of record across emails, spreadsheets, contracts, engineering data, and conversations with suppliers. And Vlad explains how Leo is building agents that can coordinate that context and increasingly take on the job end to end. They also discuss how you earn enough trust to let an agent negotiate on your behalf. Why building an internal AI tool can be deceptively easy until you hit the exceptions, and what changes when agents eventually sit on both sides of a transaction. Welcome back to the a sixteen z podcast. I'm Elena Berger. And today, I'm joined by Seema Amble, a partner at a sixteen z, and Vlad Keil, cofounder and CEO of LEO, which builds AI agents for enterprise procurement. Seema, you recently wrote a piece called The Incumbents Are Coming. And then as a question facing almost every AI application company, if an established software vendor already has the customer and the data and the model can work across its tools, where does a startup have an advantage? And we have Vlad here, and Vlad can really help us understand where a startup has an advantage. So I think a good place to start is the cases for and against the incumbents. So if a company can connect a capable AI agent to the software it already uses, where does another application actually have a …

Get the full transcript (10,793 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all a16z Podcast transcripts →