Introducing AI Managed Accounts using Salt & Nillion
by Tamlyn Rudolph, Cofounder Salt
Bespoke wealth management in traditional finance is typically conducted through managed accounts. A managed account is an investment account owned by an entity but managed by a 3rd party such as a professional wealth management firm, a robo-advisor and soon, AI wealth agents.
Managed accounts differ from mutual funds or exchange-traded funds (ETFs) in that investors directly own the securities within their portfolios. This ownership structure provides greater transparency and control over the specific assets held, enabling high levels of customization rather than sending assets to a pooled vehicle that is managed according to a preset strategy or index.
Managed accounts are the early infrastructure that allows robo-agents to interact directly with our portfolios on our terms. As AI agents evolve to further tailor our wealth management to our individual financial goals and risk tolerances, they will use an evolved version of this infrastructure.
As such, we’ve been exploring how managed accounts for AI can be simply and usefully engineered within Web3.
Introducing Managed Accounts for Web3
Salt’s “programmable ownership” platform enables entities to set up segregated accounts and delegate the advisory and execution aspects of their wealth management needs without relinquishing control over their assets.
Wealth managers (including robo or AI) are delegated micro-key shares and can instruct and execute subject to pre configured policies. These policies provide the important guard-rails within which the AI or other wealth managers must operate.
This setup ensures that while the wealth manager can execute complex strategies, the ultimate control and ownership remain with the ownership entity.
Salt is enabling a marketplace where wealth managers (digital or human) can compete for entities to delegate the bespoke management of their portfolios to them. Naturally for wealth managers, the ability to scale their algorithms across a broad customer base without significant additional operational overhead is incredibly attractive.
The Importance of guard-rails
Implementing guard-rails is crucial to ensuring the safety and integrity of portfolios, particularly when delegating execution to a 3rd party. For example, the owner of the managed account may approve specific yield protocols, limit amounts or even daily rates of capital deployment.
Guard-rails are made up of a suite of rules which can be general across a portfolio or highly account and context specific. By engineering them in open-source code, the rules or policies that govern transactions and interactions are easily auditable. Despite being open-source, the application of these guard-rails is able to remain proprietary to maintain the uniqueness and competitive edge of each entity’s strategy.
Privacy-Preserving Policies with the Nillion Network
To facilitate privacy-preserving policies, we have been piloting the Nillion network. The Nillion network provides a solution for implementing open logic with private usage.
Policy logic is defined by Salt (and its community) and can be made accessible in an open library for anyone to peruse. However during implementation specific inputs to the policy logic (eg: spend limits, trusted counterparties etc) are set by, and private to, the Salt customer. The Salt platform can request a policy check at any time and receives only the result. There is no leakage by the Nillion Network of policy inputs.
Our pilot project has demonstrated the effectiveness of this approach, showcasing how the Nillion network can successfully implement privacy-preserving policies within the Salt framework.
A Simple Example
Consider Alice, the CEO of a family office who invested early in Bitcoin and has $30M $BTC and $ETH on a ledger. She wants to start actively managing her digital asset portfolio. She knows Bob who is active in crypto and she wants to engage his assistance. He doesn’t have a separate fund or pooled vehicle set up to manage her funds.
Alice and Bob co-create a plan for the management of the digital assets, including staking and lending as well as swapping some of the $ETH and $BTC to other assets.
For execution they agree a set of parameters that Bob can work within to financially engineer the portfolio. These include:
- Approved DeFi and staking services, including setting parameters
- Whitelisting asset transfers to Alice’s internal accounts
- Approved bridge services
- Approved EVM chains to hold assets on
- Restricting the approved asset list to the top 20 assets
- Requiring approval by Alice for any transfer over $5M in value
Alice uses Salt to set up an account that, by virtue of cryptography, works across all EVM chains. She delegates a minority key share to Bob with the above policy delegations set. Her robo-signing squad is always on, monitoring and blocking any transactions that breach the above policies.
If Bob resigns from managing Alice’s portfolio she can simply remove him from the account with one action and easily replace him with another portfolio manager.
Conclusion
The paradigm shift towards highly tailored financial asset management seems inevitable as AI converges with democratized financial infrastructure. At Salt we are building an open access platform to enable minority account delegations and granular safety controls to act as guard-rails for AI managed accounts in Web3.
The result is a powerful combination — democratized access to advanced financial strategies while preserving self-sovereignty.
For further information on our pilot project with Nillion, please check out the recent X space with Lukas from Nillion.
