Intelligent Advising Systems: A New Category for Consumer AI
People make the decisions that matter most under the worst conditions for making them well.
Why nothing that exists solves it
Why now
The category: Intelligent Advising Systems
Characteristics of an IAS
Domains: regulated, unregulated and everywhere in between
Career navigation as the first vertical
What this means
Career transitions, financial planning, healthcare decisions, education choices and elder care are among the decisions that most directly shape a person's future: their economic security, their wellbeing, the opportunities available to them and their families. These decisions share a set of conditions that make them genuinely difficult. Information is fragmented and asymmetric, often controlled by the institutions on the other side of the decision. Expertise that could help is expensive and hard to access. The decisions themselves unfold over months or years, with consequences that are delayed, hard to attribute and often irreversible. And in every case, consumers spend significant money and time seeking support, turning to an array of disconnected tools and services, none of which solves the problem comprehensively.
These are also the conditions under which human judgment is least reliable. Decision science has documented this extensively: when outcomes are distant and uncertain, people discount the future more steeply than any rational analysis would justify. When the path not taken is invisible, learning from the decision becomes structurally harder. When the stakes involve identity, who a person understands themselves to be and who they want to become, loss aversion and status quo bias intensify. When options multiply and information grows more complex, people default to inaction or familiar heuristics rather than engaging more carefully. Without active support designed to account for these patterns, even thoughtful, motivated people make decisions that do not serve them well.
The cost of this is enormous and unevenly distributed. Institutions absorb poor individual decisions as statistics: attrition rates, default rates, employee turnover, readmission numbers. For the individual and the family, a poorly supported career transition, a mismanaged financial decision or an avoidable healthcare crisis is not a data point. It is a setback that compounds across years, limiting what comes next in ways that are difficult to reverse. The scale of this problem is not abstract. Millions of people, across every one of these domains, are spending real money and significant time trying to manage decisions they are not equipped to manage alone. They deserve better tools, better methodology and better support than what currently exists.
People facing consequential decisions turn to whatever is available: a patchwork of services and tools, each addressing a fragment of the problem, none addressing its full shape.
Information tools, from search engines and comparison sites to financial calculators and job boards, surface data without context, interpretation or any connection to the person's specific situation. A person can spend hours researching and end up with a collection of facts they have no structured way to evaluate, organize or revisit. The information was useful in the moment and disconnected from everything that follows.
Expert guidance, when accessible, provides what information tools cannot: experienced judgment, interpretation and accountability. But access is constrained by cost and availability. A fiduciary financial advisor, a skilled career strategist, a specialist physician: these professionals can change outcomes, but most people cannot access them regularly, and even those who do receive guidance in isolated episodes. A meeting produces insight that lives in notes or memory, disconnected from the decisions that happen between appointments and from the next professional interaction months later.
Courses and workshops teach relevant knowledge but do so generically, unconnected to the participant's specific circumstances, timeline or constraints. The knowledge does not persist in any structured form the person can act on or return to. When the program ends, they are on their own to apply what they learned, without a system to hold their work or build on it.
Coaching and group programs provide accountability and peer support, both of which matter. But the quality and methodology vary enormously across providers. Most lack structured, evidence-based approaches to the specific domain. Few offer analytical tools or persistent information systems. The result can be genuine motivation and community without the means to channel them effectively: commitment to action without a reliable structure for determining what the right action is.
General-purpose AI tools appear to offer the personalized support these decisions require, but they cannot actually provide it. They lack knowledge of the person's situation, constraints and history. They generate confident responses without showing their reasoning. They produce no durable record the person can build on. And the person using them has no reliable way to evaluate the quality of what they receive, which is the core capability they were looking for in the first place.
What connects all of these is a shared gap: no solution provides a persistent, personalized record of the person's research, analysis, decisions and evolving context that grows and compounds across time. This matters because consequential decisions are not one-time events. Careers change and change again. Health conditions evolve. Financial circumstances shift. Education planning spans years. Elder care escalates in complexity. Each of these domains requires revisiting prior information, updating prior analysis and making new decisions in familiar territory. When nothing persists, the person does not just lose time rebuilding context. They lose the opportunity to build on what they have already learned and gathered, to develop stronger skills with each engagement and to carry forward a structured foundation that makes each subsequent decision more manageable than the last.
The domains where people need structured support are becoming harder to manage alone. Labor markets are more volatile, with the capabilities required for a given role shifting faster than the language used to describe them. Healthcare decision-making has grown more complex across multiple dimensions: treatment options, insurance structures, provider coordination, cost management and the sustained burden of managing ongoing conditions. Financial decisions involve more options, more complexity and more institutional bias in the guidance available to consumers, where advice on evaluating and selecting between products is both costly and compromised by the incentive structures of the firms providing it. Education paths have multiplied while the cost and consequences of choosing among them have increased. In each of these areas, the individual bears more of the decision burden, the information environment is more complex and the stakes of navigating it poorly are higher.
At the same time, the technology to deliver a genuinely different kind of support now exists at consumer scale. AI has reached a capability threshold where it can provide cognitive augmentation to support complex decisions: expanding the information a person can access and synthesize, recognizing patterns across large and messy data sets, generating alternative interpretations and scenarios, modeling constraints and tradeoffs and helping construct the narratives and communications a person needs. These capabilities extend what a person can do with their own thinking. They do not replace the thinking itself. Alongside AI, the broader ecosystem of consumer technology tools has matured: customizable workspace platforms, relational databases, tracking and collaboration tools are now cheaply available, highly configurable and usable by non-technical people. Building a persistent, personalized environment for managing complex decisions no longer requires custom software development. It requires thoughtful design applied to widely available tools.
The pain of managing these decisions without adequate support is not new. Families and individuals have been bearing this burden for decades, and institutions have been absorbing the downstream effects as costs they externalize: turnover, attrition, default rates, avoidable healthcare spending. What is new is that the combination of domain pressure and technology maturity makes a real response possible for the first time. The IAS category is emerging because the need has become acute enough and the tools capable enough that the gap between what people face and what they can access has become both visible and addressable.
An Intelligent Advising System integrates three components that existing products and services almost never combine.
The first is technology tools, including AI. The AI capabilities within this layer perform cognitive augmentation: information expansion and research synthesis, pattern recognition across complex data, hypothesis and scenario generation, constraint and tradeoff modeling and narrative construction including drafting, reframing and translation across audiences. Alongside AI, this layer includes the information management infrastructure the person needs for a complex, ongoing process: organized workspaces, relational databases, tracking systems, timeline management and environments where information connects rather than fragments across sessions. The AI works transparently throughout: the person sees what is happening, directs the focus and evaluates the output.
The second is evidence-based domain knowledge. In every area where an IAS operates, there is a body of research, practitioner expertise and best practices that experienced professionals already know. In some areas this takes the form of competency frameworks or structured assessment instruments. In others it is accumulated practitioner knowledge, evidence-based protocols or decision approaches refined through years of professional experience and research. The person working through a consequential decision does not start from zero. An IAS brings organized, specific knowledge to bear on their situation: giving them a structure for their thinking, a vocabulary for their circumstances and defined steps that make an overwhelming process manageable.
The third is human guidance and accountability. This is facilitation, not advice. The humans in an IAS are not telling people what to decide or exercising expert judgment on their behalf. They are supporting people in learning to use the frameworks and tools, helping them develop their own analytical capability, providing accountability for follow-through and offering the relational support that technology cannot. AI can expand information, model scenarios and surface patterns, but it cannot exercise normative judgment about what matters most to a specific person. It cannot read the contextual dynamics of a situation: timing, power relationships, lived circumstances. It cannot provide relational accountability, the kind built on trust and reciprocity that sustains effort through difficult stretches. Human facilitation does what technology does not: it meets the person where they are and supports them in building the capability to advocate for themselves.
These three components, integrated, define the category. Individually, each one exists in the market in some form. What does not exist is the integration: a product that combines persistent technology and tools, structured domain knowledge and human facilitation into a single system designed around the person's ongoing decision process. The integration is required for the category, because the problems people face require all three working together.
Design characteristics are emerging to shape how an Intelligent Advising System should work. These represent principles taking form as the category's first products are built and tested, distinguishing IAS products from adjacent offerings.
Persistent, compounding engagement. The person's research, analysis, decisions and context accumulate in a system they own and return to. Because these areas of life inherently require revisiting, each return builds on what came before: information already gathered, skills already developed, patterns already recognized. The person who returns for a second career transition or the family managing an evolving care situation works from a structured foundation that grew from prior engagement. This transforms episodic support into a continuous resource, one the person can manage independently between active decision periods and that grows more useful with each engagement.
Transparent AI. The person sees how AI is being used, directs its application and evaluates the output. This prevents the dependency that characterizes most consumer AI interactions and builds a transferable capability: the person learns to work with AI effectively, to evaluate what it produces, recognize its limitations and integrate its output with their own reasoning. An IAS treats AI as a cognitive tool the person wields, not a decision-maker they defer to.
Selective social collaboration. Consequential decisions rarely happen in isolation, and the people involved need different access to different information at different stages. A family coordinating elder care involves multiple members who share some information and keep other parts private. A student's education planning involves parents and counselors who should engage with some elements but not others. Career navigation produces specific artifacts a person might share with a mentor without opening the entire workspace. An IAS supports this kind of controlled, selective sharing, and consumer collaboration tools now make this feasible without imposing the complexity of enterprise-style permission systems.
Data privacy and consumer ownership. An IAS collects deeply personal information: goals, values, constraints, decision patterns, financial circumstances, health considerations, family dynamics. This richness is what makes the system useful, and it is what makes consumer ownership and control non-negotiable. People must own their data, control access to it and be able to take it with them. Over time, the accumulated record in an IAS could become the foundation for a personalized agent that serves the person across decisions and years. That possibility depends on the person owning the data from the beginning. The business model of an IAS follows from this principle: a product funded by monetizing the person's data or attention is structurally incompatible with the trust the category requires.
The IAS opportunity spans areas with very different regulatory profiles, and it is worth being precise about this.
Some areas involve essentially no regulated professional activity. Career navigation is the clearest example: no licensing requirements, no compliance constraints beyond ordinary data security practices. The person's research, analysis, decision-making and outreach are entirely their own. This makes career navigation a clean environment for proving the IAS model and is one reason it was the first vertical developed.
Other areas involve meaningful personal data considerations without regulated professional practice. Post-secondary education planning is a multi-year process involving academic, financial and logistical decisions that begin early in high school and involve coordination among students, parents and school advisors. The planning itself is not regulated, but the data involved includes minors' academic and personal information, which introduces privacy responsibilities that an IAS must manage carefully.
Still other areas sit alongside heavily regulated professional activity, and the IAS opportunity is in the considerable space that surrounds the regulated transaction itself. Managing a chronic health condition involves ongoing decisions about treatment, medication and provider coordination. The medical decisions require licensed professionals. But the sustained work of tracking symptoms and medications, maintaining provider and insurance records, preparing for appointments, coordinating across specialists, researching options and managing the logistical weight of ongoing care is not medical practice. It is complex, evolving decision management that involves sensitive health information requiring appropriate protections. An IAS works alongside clinicians, not in place of them, and the structured record it maintains often makes clinical interactions more productive because the person arrives organized rather than reconstructing months of context.
Seen from the consumer's perspective, the amount of activity and information management that falls outside regulated or protected areas is striking, even in domains that include regulated components. All of these areas show conditions that an IAS could address. Not all will have equally attractive business models or equally simple paths to market. But the structural pattern is consistent: wherever individuals face complex, consequential, ongoing decisions with inadequate support, the components that define an IAS are what the situation calls for.
Career navigation is the first IAS vertical. Third Road developed the IAS category thesis and built its first offering, Career Strong, as an Intelligent Career System for career navigation. The category thesis extends to any area where the same structural conditions appear.
The IAS category has emerged because the need is great, increasing and now addressable. Individuals and families bear the weight of decisions that institutions treat as distributed costs: employee turnover, student attrition, avoidable healthcare spending, preventable financial hardship. For the person living through a poorly supported career transition, an unmanaged health condition or a financial decision made without adequate information, the cost is not distributed. It is specific, personal and compounding.
Intelligent Advising Systems are centered on the idea that consumers deserve advocacy: specifically, the technology, methodology and support to advocate effectively for themselves and their families. The category exists because the tools to deliver on that idea finally match the scale of the need.