Authored by Kiwi Yao, researcher @OKX Ventures
When combined, smart accounts and intent create a powerful tool that allows you to manage your digital assets more efficiently. The ERC-4337 standard enables third parties to process order flows for you. This makes it simpler to carry out tasks while also improving scalability. The alliance between smart accounts and intent paves the way for more advanced interactions in the future.
What's the connection between AA and intents?
While intents and account abstraction (AA) are connected, they serve distinct roles in the ecosystem. There isn't a fixed binding relationship between the two, but a complementary synergy.
Intent is putting the user experience first, swiftly breaking down user needs and translating them into one or more user operations (user ops). It finds the best operational route to achieve a specific goal.
As history shows us, intents have addressed more straightforward needs. However, as technology advances, we're poised to tackle complex challenges with numerous conditions, multifaceted steps, and a range of execution environments. In such advanced scenarios, AI agents might even come into play.
To understand this further, here are some examples:
Consider Telegram bots as an intent innovation. Even though the backend still employs externally owned account (EOA) wallets, the user experience remains unaffected.
Imagine a scenario where a user wishes to purchase 1,000 USD worth of Ethereum. Here, the solver determines the parameters, such as the chain selection, the decentralized exchange (DEX), and slippage tolerance. The derived user op then interacts with an EOA, multi-party computation, or AA wallet at the backend.
Projects like 1inch's DEX aggregator are pioneering the intentional narrative. Here, users merely state their input and slippage tolerance, leaving it to the contract to find the optimal operation. Similarly, Flashbots auction search, with its transaction order preference, epitomizes the searcher intent. As the digital world evolves, we can anticipate myriad forms of intent emerging.
On the flip side, AA is all about backend operations. It streamlines the execution of user commands, providing a seamless and efficient experience.
Classifying intent
Understanding the different types of intent is crucial:
Conditional intent: Triggered by specific conditions, similar to an "if statement". An example from the financial realm would be the stop-loss feature.
Continuous intent: Denotes repetitive operations that occur based on recurring scenarios or time intervals, such as the principle behind monthly dollar-cost averaging.
Multi-step intent: This cascades from one intent to another, behaving much like a state machine. For instance, if someone intends to purchase Ethereum and has conditions about price thresholds, transfers, and arbitrage, this would fall under this category.
Intent graph: Multiple intentions come together to create a complicated system that's seen when users seek to purchase or sell tokens based on certain conditions, such as changes in governance, block mining, or market prices.
Account abstraction makes intents more efficient
The concept of intent results from technological advancements and user experience. By minimizing the need for operational signatures, intent streamlines user interactions across Web3. With the evolution of intent, we develop new infrastructures, languages, and approaches to problem solving, making it pivotal in shaping future applications.
Key opportunities include:
Intermediary intent pool or intent public chain
This involves innovative message formats to communicate with smart contracts, replacing traditional mempool systems for transaction propagation and discovery.
Diversified solver implementations
Short-term: Integration of mature super smart contracts.
Long-term: Usage of high-freedom AIs, albeit with implementation challenges. This will inaugurate an off-chain preprocessing followed by on-chain processing paradigm, beneficial for protocols like Uniswap X. The ZK coprocessor Axiom stands as a testament to the potential of privacy-solving.
Language evolution
Current players like Juvix and Essential are pioneers in this area. Their success hinges on building user trust and navigating the challenges ahead.
Elevating wallets
As the foundational element of the intent layer, wallets are primed to evolve into unified entry points. By integrating with diverse protocols, they can tap into significant user traffic.
Intent pools
Intent pools are like pipelines. They take user intentions and turn them into transactions. There are two main ways to do this: MEV boost auctions or on-chain transactions. A number of intent pools exist today:
Unlicensed intent pool: Enables propagation across system nodes, granting executors unrestricted access. This pool does pose some risks, such as DoS threats and MEV issues.
Licensed intent pool: Offers a resilient defense against DoS attacks. However, its inherent trust requirement can conflict with the blockchain's decentralized ethos.
Hybrid intent pool: Marries licensed propagation with permissionless execution. Order flow auctions exemplify this approach.
Further explorations:
Intent layer: Platforms like Anoma and Flashbots SUAVE are ushering in dedicated layers to address MEV risks.
Diversified solver implementations: Strategies include integrating contracts with smart contracts, off-chain preprocessing paradigms, and making sure of privacy in operations.
AI automation: The future sees two primary avenues: AI as an integrated wallet feature or AI directly linked with user op. Competitors in this area could range from Siri and Google Assist to ChatGPT clients.
Novel languages for intent expression: Efforts are underway to develop languages that enable private intent expression. Notable projects include Juvix, which offers a platform for crafting privacy-centric decentralized apps, and Essential, which is pioneering a domain-specific language (DSL) for intent expression.
Wallets and opportunities for entry
Intent's backend can seamlessly integrate with conventional EOA and MPC wallets. For instance, Telegram bots employ intent and still use EOA wallets without compromising on the user experience. With users' security concerns front of mind, future strategies might use AA to better manage risk.
Smart contract wallets and SDKs
While the market for components like bundler, paymaster, and providers like Wallet Factory is dense, intent introduces avenues for innovation.
Features and modularization
Most smart contract wallets already include features like social recovery. However, integrating intent can take this further. Examples of modularized features are found in wallets like Argent, Candide, and Gnosis Safe, which support additions beyond the EIP-4337 interface.
Social recovery options
Web2 service providers: Argent offers recovery based on trusted contacts.
User devices: Including browser and mobile storage.
Email: UniPass uses emails for recovery while protecting user data privacy with zero-knowledge proofs.
Multi-Sig: BLS Wallet and Argent allow recovery via multiple user-specified EOA addresses.
MPC: Web3Auth's service divides user keys for enhanced security, keeping the complete key hidden.
Transactional applications
Integrating with decentralized finance (DeFi) protocols, supporting automated trading offers a spectrum of user benefits:
Uniswap X, for example, provides opportunities for order arbitrage.
Protocols can eliminate repetitive authorizations. A use case being dollar-cost averaging in DeFi protocols where investments occur automatically without manual intervention.
Cross-chain applications
These play a pivotal role in enhancing user experience across chains:
Cross-chain DeFi protocols: Orbiter, a decentralized bridge, offers low fees and swift multi-currency transfers.
Multi-chain DApp management protocols: Platforms like DAppOS function as cross-chain SDK protocols. By partnering with DEXs like Perpetual Protocol, they uplift the DeFi cross-chain experience.
ZK class application opportunities
The ECR-4337 standard offers strong encryption for user intents. This means zkProof can work better and potentially, a zkProof marketplace can emerge.
Encryption requirements
User ops can undergo threshold decryption or distributed key generation before transmission, providing decryption only after block confirmation by the bundler pool.
This interplay of zk, user ops, and gas verification propels the growth of zkproof and zkOracle, drawing parallels with platforms like the nil foundation.
Privacy payments process
The ECR-4337 mechanism allows users to execute privacy payments, safeguarding account confidentiality. Here's how it works:
Users place funds into a privacy-protected rollup, targeting withdrawal to a new, unfunded CREATE2 address.
They produce mixed withdrawal info, directing funds to a MixerPaymaster address, not their own.
A user op with this info is created with UserOp.paymaster set as MixerPaymaster and sent to the ERC-4337 mempool.
Bundlers pick this user op, verify its authenticity, and check MixerPaymaster's willingness to cover the gas fee.
Once validated, bundlers integrate the user op into the chain, directing it to the entry point. This step results in deploying an account to the user's new address.
The MixerPaymaster does the withdrawal, taking out the operation fee and sending the rest to the user's new address. The new address now has an account deployed. This procedure uses the ERC-4337 mempool without going through relayers.
Risks and thresholds
The architecture of intention mechanisms brings about critical trust, centralization, and transparency concerns. Remember, doing your own research is important to help keep your digital assets safe.
Centralization and trust
The cornerstone of intention mechanisms is trust. High trust expectations often lead to steep entry barriers for new entrants. Combined with unsupervised execution, it can result in block production being controlled by a few dominant entities, promoting centralization.
For instance, in the order flow auction space, established entities like Flashbots and CoWswap command significant market share, making it challenging for newcomers to gain traction without extensive resources.
Transparency issues
Intention mechanisms often allow users to relinquish some control over their on-chain assets. With many mempools being private, users might need clarification about how their intentions are executed. It's imperative for intention applications to make sure their systems are designed with caution, keeping potential threats at bay.
The final word
By seamlessly merging smart accounts and intent, we can usher in an era where digital asset management is efficient and sophisticated. The ERC-4337 standard entrusts third parties with order flows, streamlining operations and amplifying scalability. While intent and AA contribute uniquely, they exemplify how technology, with a user-centric approach, can simplify complex tasks and even integrate advanced tools like AI in future activities.
