a16z: AI Agents Could Power the Next Uber-Style Platform

iconTechFlow
Share
AI summary iconSummary
a16z Crypto highlights how AI and crypto news could fuel the next Uber-style platform, using AI agents to manage complex transactions. The firm explains how agents can interpret nuanced preferences and match them with suppliers, extending the Uber model to group dining and beyond. It also explores how crypto’s “intent” model can be applied to everyday commerce, with use cases in travel and procurement. New token listings may emerge as these platforms evolve.

Author: Guy Wuollet, a16z Crypto

Compiled by: Deep潮 TechFlow

DeepChain overview: Uber used a minimalist demand expression to connect riders with drivers, but in reality, many transactions involving multiple parties, dynamic budgets, and complex terms have long been difficult to digitize. AI agents are changing this, enabling systems to restructure bilateral matching networks around more granular, non-standard goods and services.

The a16z Crypto team, combining the established "intents" and order flow bidding models from the crypto space, breaks down the profit mechanisms and cold start pathways for agent-based business networks: In an era where software development is approaching zero cost, the machine networks that can accumulate履约信誉 and two-way traffic represent the true long-term moat.

Uber provides people with an incredibly simple way to express their need: to be picked up at point A and taken to point B. The network then automatically matches you with a driver willing to provide pickup and drop-off services at a price you accept. The more drivers there are, the easier it is for passengers to get a ride; the more passengers there are, the stronger the incentive for drivers to join the network.

I believe that AI agents will enable such two-sided networks to be built around an extremely broad range of goods and services. Agents can help humans express more complex preferences and find suppliers willing to precisely fulfill those preferences.

Some of these trading scenarios already exist today, but have been out of reach for most people due to excessively high prices or overly complex coordination; others have never been attempted simply because the cost of manual arrangement is too high. Agents will also give rise to entirely new types of trading we never even imagined—creating enormous opportunities for entrepreneurs.

Consider a simple, familiar everyday scenario: organizing a dinner with a group of friends. Each person has different availability, budget limits, and preferences for restaurant cuisines.

Today, it’s often necessary for a friend to act as the “coordinator” in a WeChat group: individually gathering everyone’s preferences, searching online for restaurants, and repeatedly going back to the group to renegotiate when the chosen restaurant doesn’t meet the criteria. Meanwhile, many restaurants have plenty of empty tables on Tuesday evenings but lack effective channels to reach groups whose plans haven’t yet been finalized.

If handled by your personal agent, it could quickly align with your friends’ agents to consolidate a clear request: find a six-person dinner in the Mission district, under $60 per person, between 5:30 PM and 8:00 PM, as long as it’s not sushi.

Subsequently, each restaurant's agent can respond with real-time bids based on available tables: one restaurant might offer a better discount for the early 5:30 PM slot, while another could proactively customize a family-style meal designed for six people. Your agent simply compares all offers against everyone’s preferences and automatically completes the reservation within your pre-approved limits.

Thus, restaurants with available tables at 6 p.m. can directly send targeted offers to those willing to dine earlier, before consumers have decided where to eat. Buyers have clear preferences and budgets, while sellers have idle seating and pricing flexibility—the only missing piece is the underlying network that enables these agents to discover and transact with each other intelligently.

Uber has demonstrated how networks can match buyers and sellers around relatively simple transportation needs, and agents are poised to extend this model to more complex consumption scenarios such as dining with intricate constraints. Most internet networks in the past were built for human behavioral patterns, but now we urgently need to build networks for machines.

Entrepreneurs can absolutely build dedicated matching hubs that allow individual agents to bring complex demands into the market, enabling suppliers to compete in fulfilling these demands.

How to build an agent trading network

In fact, the cryptocurrency space has long established a mature organizational paradigm around the concept of "intents": users simply declare their desired transaction outcome, and various service providers compete to deliver it.

Using UniswapX as an example, users sign an order specifying the tokens they wish to swap and the acceptable terms; market maker solvers then compete across multiple liquidity sources to fulfill the trade.

AI agents can extend this "intent model" to everyday business contexts. Large language models (LLMs) transform everyday natural language conversations into standardized requests that suppliers can execute, while agents scour the web for quotes and finalize transaction details. This low-barrier interaction has the potential to significantly reduce collaborative costs, making smaller transactions—like a dinner for six—more economical.

The same network effects that once drove Uber’s value will gradually accumulate in these more complex transaction scenarios. Each time a new restaurant joins, diners gain an additional precisely matched option—especially beneficial for small groups with strict dietary restrictions or time constraints.

As the number of groups requesting reservations increases, restaurants gain significantly greater opportunities for additional table-turnover revenue, prompting them to connect their own reservation systems to the network and offer more targeted pricing. Once participation in a single city reaches a critical threshold, this network can naturally expand to include more consumer categories.

With a cold start strategy, entrepreneurs can begin by focusing on a single city’s dining scene. The first-generation product must genuinely save users significant time or money, encouraging them to bring in friends; at the same time, it must enable partner restaurants to see new incremental orders they wouldn’t otherwise receive. Only when both sides benefit will they remain engaged and continue inviting new members.

Regarding monetization, I believe intelligent agents targeting end consumers can be fully open and free, with revenue generated through a model similar to "Payment for Order Flow." Specifically, users state what they want and the maximum amount they’re willing to pay, after which suppliers compete to fulfill the request.

If an agent quotes $300 to arrange a meal and the backend matches with a restaurant willing to accept $270, the platform can retain the difference or charge a routing commission for successfully directing the transaction to the restaurant. This matching revenue can then directly subsidize the consumer’s AI reasoning compute costs.

Of course, the platform must ensure transparency: users should clearly understand how their agents generate profits and the underlying logic behind how they filter and rank quotes. Ultimately, consumers' willingness to delegate spending decisions depends on their trust in this filtering mechanism.

Why do this type of business model have long-term barriers?

As artificial intelligence drastically reduces the marginal cost of software replication, having user-friendly software alone is no longer scarce. Instead, networks of individuals and businesses willing to persistently engage in transactions within a specific ecosystem are evolving into a more fundamental competitive moat. Looking back at the internet era, the greatest companies—Facebook, eBay, and Uber—all followed this principle: the more participants that join, the more valuable the network becomes for each individual.

The matching network for agent-based commerce can also accumulate this hard-to-disrupt first-mover advantage. Competitors may quickly replicate software with nearly identical features and interfaces, but they still must build from scratch a reliable, consistent network of buyers and sellers. Returning to the restaurant example, this means attracting a sufficient number of local restaurants willing to proactively quote prices, along with a large group of consumers ready to place orders at any time.

Ensuring that every matched trade can be reliably settled is a crucial part of building network infrastructure. Partner restaurants must be confident that the agent requesting payment has been authorized by the entire dining group; diners, in turn, need to be certain that once they confirm acceptance of the quote, their reservation will be securely locked in at the agreed price.

Therefore, the network must establish its own verification mechanism to approve agent permissions, record key booking terms, and smoothly arrange fund settlements. At the same time, in the event of履约 anomalies such as a group canceling a reservation at short notice or a restaurant failing to hold the reserved table as agreed, both parties should clearly understand the pre-agreed resolution process.

Entrepreneurs can build these trust mechanisms alongside the trading marketplace. As buyers and sellers begin to interact frequently on the network, the platform will naturally accumulate a reliable record that identifies who consistently fulfills their commitments, along with a mature dispute resolution process. A network that helps participants find high-quality offers and ensures extremely reliable fulfillment will naturally encourage everyone to return to it for their next transaction.

What other agent networks can founders build?

Restaurant reservation is just a simplified example I used to illustrate the concept. In reality, AI agents can also help build numerous distinctive vertical transaction networks, with the following directions being particularly worth exploring:

· Travel and Business Trips: A personal agent coordinates and aligns the schedules and budgets of multiple travelers, then negotiates bundled deals with multiple hotels and airlines to secure a trip that satisfies everyone. When the team remains flexible on departure dates or destinations, they can unlock hidden deals that traditional keyword searches cannot reach.

· Enterprise Procurement: An enterprise-specific procurement agent operates 24/7 to dynamically negotiate forward volume commitments and price lock-in periods with upstream suppliers. While comparing prices, it simultaneously evaluates delivery schedules and payment terms, helping buyers lock in total costs while providing sellers with predictable, stable order demand.

· Compute Resource Market: Agents can automatically match computational workloads with idle network compute power based on hardware specifications, data center geographic location, and time windows. Enterprises can even pre-purchase future compute capacity through this network to effectively hedge against the risk of soaring compute costs during periods of business expansion.

· Online dispute arbitration: The parties involved in a transaction can jointly agree at the time of signing that any dispute will be resolved by a shared arbitration forum, achieving far faster resolution than traditional judicial systems. Each additional pair of buyers and sellers adopting the same arbitration rules increases the forum’s appeal and credibility for future market participants.

· Physical device network: Agents can directly procure logistics services from a distributed drone delivery network or purchase real-time data from connected sensor devices. Each new device that joins directly expands the coverage radius of the physical network or adds a new service capability precisely needed by other participants in the ecosystem.

· Collaborative Health and Medical Research: Individual users can securely contribute their health data to a shared pool and participate in crowdsourced research projects. The larger the base of data contributors, the more robust the large-scale clinical trials that can be supported, enabling solutions to medical challenges that were previously impossible to address with small sample sizes.

In addition, rich entrepreneurial opportunities in the digital space also emerge around “genuine human relationships.” As machine intelligence becomes increasingly inexpensive, even nearly free, the real world still urgently needs effective ways to verify authentic identities, endorse professional qualifications, and build deep trust. As more work moves away from traditional employment institutions, networks of professional reputation credentials, membership clubs, and similar organizations will become increasingly valuable.

Conclusion

True business opportunities often lie in transactions people deeply want to make, but today, the time and cost barriers to manually coordinating and matching these transactions are simply too high. The largest future agent networks will almost certainly emerge from entirely new exchange scenarios we cannot yet imagine.

Start by entering a specific and niche vertical market, first ensure that both sides of demand are fulfilled, then use tangible incremental value to incentivize participants on each side to continuously bring in others. If you're building an agent network, I look forward to connecting and discussing with you.

Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.