Timothy Clifford - Plan Assist | 1 Million Cups Clermont - January 28, 2026 скачать в хорошем качестве

Timothy Clifford - Plan Assist | 1 Million Cups Clermont - January 28, 2026

Timothy Clifford, a certified financial planner, introduces PlanAssist—an AI-enabled “virtual conference room” designed to reduce financial decision noise and improve client–advisor alignment. He outlines the challenge of information overload and why clarity comes from structure, then explains common AI limitations (hallucinations, “yes-man” behavior, bias, and prompt sensitivity). The learning focus is how a guarded, principles-based prompt and shared group chat could help clients bring better questions to their advisor, support day-to-day decision-making, and address communication gaps while navigating privacy and compliance constraints. 0:05 – Speaker introduction and opening framing: information overload vs what to ignore 0:25 – Background as a financial planner and the recurring client question about “right” money decisions 1:20 – Introducing PlanAssist: AI tool to connect advisor and client AI usage in one shared space 2:08 – Research example on “more information” leading to worse decisions 2:59 – Medicare marketing example to illustrate event-driven information flooding 4:26 – Defining the problem as persistent “noise” from people, media, and the internet 5:04 – Transition to AI as a new layer of the information challenge; request for audience feedback 5:53 – “Clarity comes from structure” and why AI can feel structured even when it misleads 6:31 – Four AI limitations overview: hallucinations, “sickle fancy,” bias, and prompt dependence 7:12 – Hallucinations explained: confident wrong answers 7:23 – “Sickle fancy” explained: AI as a yes-man over long chats 7:58 – Bias explained, especially with paid tools retaining user history 8:48 – Prompting explained (AIM: actor, input, mission) and why phrasing changes outputs 9:34 – Advisor–client mismatch problem: clients ask AI separately and return with different contexts 10:12 – Goal: get advisor and AI “on the same page” to limit information and improve relevance 10:59 – Building a long, guardrailed prompt and applying first principles (liquidity, safety, growth) 11:48 – Health analogy to explain principles vs actionable “how” 12:40 – Using the PlanAssist principles (“have a plan, be diversified, seek counsel”) inside the prompt 14:07 – Product concept: group chat + dashboard as a virtual conference room with advisor visibility 15:17 – Walkthrough of client experience: ask a question, view answers, request review, dashboard focus 15:58 – Facilitated Q&A begins: presenter asks the room for candid feedback on the idea 16:22 – Audience feedback: confusion about what is being built; suggestion to clarify the solution 17:51 – Question on current product stage and expected business model 18:23 – Presenter response: initial model is to offer to existing clients due to compliance constraints 18:51 – Clarification: software mirrors ChatGPT group chat but with built-in prompts to simplify use 20:07 – Additional problem statement: clients avoid “bothering” the advisor on smaller decisions 20:41 – Question: tracking participation and visibility into client activity 20:52 – Presenter response: sign-in, daily summaries; not anonymous due to advisor-client context 21:25 – Question: whether clients have been asked; response that interviewed clients “love it” 22:40 – Audience warning: some clients may replace advisors with ChatGPT; risk to consider 23:17 – Presenter response: tool targets clients who want both an advisor and AI, not AI-only users 23:54 – Question: difference from using the same prompt directly in ChatGPT; “collaboration feature” 24:16 – Presenter response: copying collaboration; adds basic client context and preloaded prompts 25:01 – Audience suggestion: personalize by client profile/portfolio; group chat may scale poorly 26:07 – Follow-on agreement: individualized, portfolio-specific responses are where value increases 26:57 – Concern raised: confidentiality and putting sensitive financial data into AI systems 27:19 – Presenter response: balancing information depth; compliance is a major hurdle 27:42 – Question: why rely on AI vs traditional advisor relationship; time check by host 28:18 – Audience perspective: password friction and preference for human contact over another login 28:52 – Audience support: AI as a tool to extend professional research capability and responsiveness 29:44 – Moderator asks for the single biggest validation goal/problem to test today 29:57 – Presenter: validation is whether a small pilot group wants a “virtual conference room” concept 30:23 – Quick show-of-hands validation; about five people indicate interest 31:01 – Audience comment: AI reinforces ego; importance of validating outputs with professionals 32:31 – Audience suggestion: alternate market is other financial planners as users, not clients 33:43 – Closing: community help question answered by the show-of-hands; wrap-up and thanks