Advisor technology usually fails for one of two reasons: it was designed by people who have never sat across from a client, or it hides its logic behind a model nobody can explain. We build the opposite of that.
These are not values on a wall. Each one has killed a feature that would otherwise have shipped.
Every workflow, every question, every sleeve definition comes from someone who has done this work for more than twenty years with real client capital at stake. That matters in ways that are hard to appreciate from the outside. The wording of a discovery question determines whether a client tells you about the second mortgage or not. The order of the screens determines whether a first meeting feels like an interrogation or a conversation. The default assumptions determine whether an advisor trusts the output enough to put it in front of a household. None of those decisions can be made well by a product manager reading a market research deck, and none of them survive contact with a real meeting unless someone who runs real meetings authored them. We write the logic first, use it in practice, and only then turn it into software.
Advisors do not operate in a vacuum, and a tool that ignores that reality is a tool that never gets used. Everything we build is designed inside the constraints advisors actually live under: Reg BI, FINRA advertising and communications rules, supervisory review, and firm-level marketing approval. That shapes the product at the language level, not just the disclosure level. We avoid the phrasing that gets material rejected — objective, predictive, guaranteed, and any claim implying an affiliation or scoring standard we don't own. Client-facing output is framed as planning indicators, resilience zones, and next actions rather than assurances. Advisor-facing output keeps the judgment with the advisor. The goal is straightforward: a tool an advisor can bring into a supervised practice without a six-month approval fight.
The same inputs produce the same outputs, every time, for every user. This sounds unambitious until you have tried to explain to a client why the software gave a different answer on Tuesday than it did on Monday. Determinism is what makes a tool defensible. It means an advisor can reconstruct exactly why a portfolio came out the way it did, months later, in front of a compliance reviewer or a skeptical prospect. It means the fifteen allocation configurations are authored, reviewed, and versioned rather than generated on demand by an optimizer chasing a covariance matrix. It means stress scenarios apply defined shocks rather than sampling a distribution. Where there is a rule, the rule is visible and the advisor controls it. Nothing important happens inside a black box.
We are named LinkPath AI and we are deliberately conservative about where AI touches the workflow. Language models are excellent at compression, summarization, drafting, and turning messy human input into structured notes — all of which consume enormous amounts of advisor time and none of which carry fiduciary weight on their own. So that is where we use them: synthesizing discovery conversations, drafting coaching content, cleaning up follow-up documentation, automating the administrative layer. What we do not do is hand portfolio construction, risk assessment, or client recommendations to a probabilistic model. Those outputs need to be reproducible and explainable, which is exactly what a generative model cannot promise. AI compresses advisor time. It does not replace advisor judgment.
The traditional grid — large value, small growth, international developed — describes what a holding is. It says nothing about what the holding is supposed to do for the household. We organize around the second question.
A sleeve is a named block of exposure with a defined job in the portfolio. Because each sleeve has an assignment, sizing conversations become concrete: how much thematic conviction does this household want, how much ballast does the plan require, how much of the income need is being met by which wrapper. Overlap becomes visible instead of hiding inside a fund lineup, and an advisor can explain any weight in the portfolio by naming the job it is doing.
Nine sleeves group into five architectural roles.
AI Super Cycle, Global Defense & Innovation, and Demographic Aging. Structural, multi-year capital cycles held with intent and sized against the core rather than layered on top of it.
Global Equity Core. The low-cost diversified equity backbone that anchors every profile and gives the thematic sleeves an honest benchmark to be measured against.
Macro Hedge. The sleeve whose job is to behave differently when equities break, sized to matter in a drawdown without bleeding the portfolio in normal markets.
Real estate, infrastructure, commodities, and energy. Carries inflation sensitivity and diversifies away from purely financial cash flows.
Equity Income, Core Taxable Income, and Core Tax Free Income. Three sleeves because cash flow needs and tax wrappers are not interchangeable — registration drives which one is used.
Most advisor software starts at the portfolio. That is the wrong end of the problem, and it is why so many recommendations feel generic to the household receiving them.
A household with no emergency reserve, an unfunded education goal, and stale beneficiary designations does not have a portfolio problem. Leading with an allocation in that situation produces a technically defensible recommendation that misses everything that actually matters. Structured discovery first means the recommendation arrives with context attached — the advisor knows which gaps are urgent, which are cosmetic, and which the client is emotionally ready to address.
It also changes the client experience. Discovery makes the household feel understood before anything is proposed, which is why RegiMint output feeds the Allocation Tool directly rather than the other way around.
Clear boundaries are part of the product. These are commitments, not caveats.
LinkPath AI LLC is a financial technology company. We build software for licensed professionals. Nothing we publish or output is investment, tax, or legal advice, and nothing on this site is an offer or solicitation for advisory services.
We hold no custody, execute no trades, and do not manage money for anyone through this company. The advisor's own firm and regulatory framework govern all of that.
Stress scenarios apply defined shocks to a stated allocation. They are illustrations of mechanical behavior under assumptions, not forecasts, and no output should be read as a projection of future results.
No proprietary black-box score, no unexplained recommendation engine. If a tool produces a number, an advisor can trace exactly how that number was produced and adjust the inputs that drove it.
Generative models assist with synthesis, drafting, and administrative work. They do not select allocations, assign risk profiles, or generate client recommendations.
Pilot data is used to improve the tools in anonymized form only. Household information entered by an advisor is not a product we monetize.