📣 Issue #3: OpenAI Releases First Open-Weight Models Since GPT-2

Plus: GPT-5 Arrives, Tool Spotlight: FINNY, Turn money-in-motion into clients, and more…

👀 Sneak Peek – What’s Inside This Issue

Here’s what we’re unpacking this week to help you work smarter with AI:

OpenAI Goes Open-Weight – gpt-oss-120b & 20b arrive under Apache 2.0, giving firms full control for compliant, private AI deployments.
GPT-5 Arrives – Smarter routing, stronger reasoning, and fewer hallucinations. A new baseline for advisors’ day-to-day workflows.
Tool Spotlight: FINNY – AI-native prospecting that scores, prioritizes, and automates outreach — from LinkedIn signals to handwritten notes.
AI for Your RIA – Stop leaving growth to chance. How data triggers like career moves and property sales reveal high-value prospects in real time.
How I AI – Fixing advisor marketing misalignment. Use AI to align your messaging with your most profitable clients.
AI Essentials – Deep dive into prompts. Why clarity matters, proven prompt patterns, and where prompting is headed in 2025.
How to AI – Lead segmentation & prioritization. Turn a raw list of names into a ranked, conversion-ready prospect pipeline.

Let’s dive in.👇

🚦 AI Pulse: What’s New?

🗝️ OpenAI Releases First Open-Weight Models Since GPT-2

OpenAI has released gpt-oss-120b and gpt-oss-20b, its first open-weight reasoning models since GPT-2 in 2019. Both are available under the Apache 2.0 license, allowing free commercial use, redistribution, and fine-tuning.

Why this matters: Unlike ChatGPT’s closed models, these can run locally or in private cloud environments — giving firms full control over data privacy, security, and compliance. They’re optimized for reasoning-heavy work like financial analysis, coding, and research, with performance on par with some of OpenAI’s own proprietary models.

For advisors, that means:

Running AI in compliant, on-premise environments without sending sensitive client data to external APIs
Fine-tuning for firm-specific workflows like portfolio reviews, market commentary, and client communications
Deploying models in VPCs or edge devices for secure, real-time insights
Lowering operational costs with efficient inference on modest hardware

One example: An RIA could fine-tune gpt-oss-20b on anonymized client history to automatically draft personalized quarterly review summaries — all hosted securely inside the firm’s infrastructure.

Bottom line: OpenAI’s move into open weights puts it in direct competition with Meta, Mistral, and DeepSeek. For financial firms balancing AI capabilities with control, this is a pivotal opportunity.

🚀 OpenAI Launches GPT-5: Smarter, Unified ChatGPT Core

OpenAI has released GPT-5, now the default model in ChatGPT, the API, and Microsoft Copilot. It replaces GPT-4 and prior models with a unified system that dynamically routes between “fast mode” for speed and “thinking mode” for deeper reasoning.

Why this matters: GPT-5 delivers stronger reasoning, coding, and multimodal capabilities, with fewer hallucinations and faster responses. It’s designed for complex workflows where accuracy and adaptability matter.

Early results: Advisors can use GPT-5 to:

Draft more accurate client reports with less editing
Switch seamlessly between quick tasks and in-depth analysis
Build interactive dashboards and presentations from natural-language prompts

Caution: The rollout was bumpy. Early users flagged errors in spelling, math, and geography, along with a less “human” personality compared to GPT-4o. OpenAI has reinstated older models for fallback and is refining GPT-5’s behavior.

Bottom line: GPT-5 is more evolution than revolution, but it sets a new baseline for AI assistants. For advisors, it means more reliable, versatile support — provided firms manage routing, compliance, and quality controls carefully.

👉 OpenAI announcement | Launch video

🔍 Tool Spotlight: FINNY for Smarter Prospecting

What it is:

FINNY is an AI-native prospecting platform that takes the guesswork out of lead generation. It combines a vast proprietary database with intelligent scoring and automated outreach to deliver personalized, high-converting prospects—without manual searching or tedious cold outreach. (finnyai.com)

Why advisors love it:

✅ High-intent matching – FINNY calculates a unique "F-Score" for each advisor–prospect pairing, predicting conversion likelihood rather than relying on generic lead scoring.

✅ AI-powered autopilot – After defining your ideal client profile, FINNY identifies, filters, and prioritizes prospects, then automates outreach and even schedules meetings for you.

✅ Multi-channel engagement – Beyond email, FINNY supports AI-generated voicemails, LinkedIn actions, and even handwritten-style messaging to engage prospects where they prefer.

✅ Proven traction – Since launching in 2024, FINNY has secured millions in funding, onboarded dozens of firms, and is growing rapidly with strong early adoption.

When it makes sense to use FINNY:

✔️ If you specialize in business-owner clients and want to target prospects who recently raised capital or had a liquidity event.

✔️ If you focus on corporate executives and want to identify those receiving new stock option grants or promotions.

✔️ If you work with pre-retirees and want a smarter way to find households in your geography that fit retirement readiness criteria.

✔️ If you’re expanding into niche wealth segments (tech founders, physicians, young professionals with equity comp) and want targeted prospecting at scale.

✔️ If you’d like to reduce manual prep—researching, qualifying, and scheduling—so you can spend more time in actual client conversations.

How it compares:

Traditional lead lists – Static, often outdated, and require heavy manual research. FINNY is dynamic, refreshed, and intent-driven.
LinkedIn Sales Navigator – Useful for networking, but requires hours of manual searching and outreach. FINNY automates both identification and engagement.
Generic CRMs – Great for organizing data, but they don’t find or qualify new leads. FINNY fills the top of the funnel with high-quality prospects before they enter your CRM.

Bottom line: FINNY combines the best of a prospect database, lead scoring engine, and outreach automation tool—purpose-built for financial advisors.

Key takeaway:

FINNY turns prospecting into a streamlined, data-driven process—helping you engage the right people at the right time, with less effort and higher conversion.

👉 Explore FINNY and request a demo

📌 AI for Your RIA

From Chance to Signals: Smarter Prospecting with AI

For years, advisors have relied on referrals, introductions, and a bit of luck to find new clients. But chance isn’t a growth strategy. AI is shifting prospecting from waiting and hoping to watching and knowing.

By scanning for data triggers that reveal financial needs in real time, AI helps advisors identify opportunities before prospects even raise their hand.

This is already happening:

Career changes on LinkedIn often mean rollovers, stock grants, or equity diversification needs.
Property transactions surface liquidity, mortgage, and insurance planning opportunities.
Life milestones — children, marriages, divorces — signal new financial priorities.
AI platforms like TIFIN Wealth and WealthEngine merge public records with behavioral data to highlight the most relevant prospects.
FINNY, which we cover in more detail later in this issue, combines life-event data with predictive scoring and automated outreach to help advisors prioritize and engage the right prospects at scale.

Key takeaway:

Referrals will always matter, but serendipity is no longer a strategy. With AI-driven signals, advisors can build a pipeline that’s timely, targeted, and far more predictable.

👉 See how money-in-motion creates high-value leads

💡 How I AI

Fixing Financial Advisor Marketing Misalignment with AI

Jonathan Michael (Founder, AdvisorX AI) recently spotlighted a hidden problem in advisor marketing:

👉 Your website says one thing about who you serve… but your most profitable, growth-driving clients are often someone else.

This misalignment between messaging and ideal client focus (ICP) is why so many advisor websites sound the same — and why growth stalls.

Traditionally, fixing this meant hiring consultants to define your niche and client persona. Today, AI makes it faster, cheaper, and more precise.

Why This Matters

RIAs with a documented Ideal Client Persona (ICP), value proposition, and referral plan saw 67% more organic growth and 3–4× more referral asset flow (Schwab 2024–25 RIA Benchmarking Study).
Taking just 60 minutes to clarify your ICP can unlock measurable growth in the next quarter.

How Jonathan Uses AI to Fix It

AI can now run a step-by-step ICP workflow that helps advisors:

Scrape your website + Form ADV to infer your current ICP.
Compare it with your true “favorite” and profitable clients.
Diagnose misalignments in tone, imagery, and positioning.
Recommend fixes to website language, visuals, and referral strategy.
Generate detailed ICP personas with demographics, values, fears, and goals.
Identify the best lead generation channels for that ICP.
Craft a homepage-ready value proposition.
Build a growth plan backed by Schwab’s benchmark data.

Instead of weeks with a consultant, this can now be done in a single ChatGPT workflow.

📝 Try this prompt:

"You are a strategic growth advisor trained on Schwab’s RIA Benchmarking Study and financial advisor marketing best practices. Analyze my firm’s website and Form ADV, compare it with my ideal client inputs, and recommend fixes to align my public positioning with the clients I most love and profitably serve. Provide ICP personas, lead gen channels, and a homepage-ready value proposition."

👉 The insight: Marketing doesn’t fail because advisors don’t try hard enough — it fails because they aren’t aligned with their true ICP. AI can now close that gap in hours, not months.

**👉 Read the full post on LinkedIn

📚 AI Essentials

Deep Dive – Prompts: Your Input to the AI

In our last issue, we explored the engines powering AI. But the results you get from even the most advanced model depend on something simple and often overlooked: your prompt. Think of prompts as the questions you ask or the instructions you give. Clear prompts = better outputs.

🔍 What Is a Prompt?

A prompt is the input you give an AI — a question, a task, or a piece of context. Just like good search queries yield better Google results, good prompts lead to more accurate, relevant, and actionable AI responses.

Short prompts: “Summarize this article.” → often vague, surface-level results.
Clear prompts: “Summarize this 1,000-word article into three bullet points that a financial advisor could use in a client newsletter.” → precise, useful results.

🧠 Why Prompts Matter

Guidance, not guessing: AI doesn’t read your mind; it responds to what you type.
Context is king: The more background you provide, the more tailored the answer.
Clarity beats creativity: Being specific about format, style, or audience improves outcomes.
Iteration is normal: Think of prompting as a conversation — refine and re-ask until it clicks.

🛠️ Prompt Patterns That Work

Here are common structures that consistently improve results:

Role Prompting:“Act as a compliance officer. Explain the risks of this investment strategy.”→ Anchors the AI in a persona with relevant expertise.
Instruction + Constraints:“Write a 200-word client email, professional tone, 3rd-grade readability.”→ Balances clarity with guardrails.
Examples for Style:“Rewrite this paragraph in the same tone as the Wall Street Journal excerpt below.”→ Gives the AI a target to mimic.
Step-by-Step Guidance:“List 5 strategies for portfolio diversification. Then explain pros/cons of each.”→ Encourages structured, logical responses.

🚀 Prompt Engineering in 2025

The term prompt engineering is evolving — it’s less about magic formulas, more about human-AI collaboration. Some trends:

Multi-modal prompts: Combine text + images + data files for richer input.
Prompt libraries: Shared repositories of “battle-tested” prompts for specific industries.
Agents and workflows: Prompts that trigger actions, not just text responses.
Personalized prompting: AI that remembers your preferred style and context.

🔮 What’s Next?

Prompts will soon feel less like “commands” and more like conversations with an expert partner. Expect:

Natural language coaching → AI suggests better ways to phrase your own prompt.
Context carryover → models that remember prior prompts across sessions.
Voice-first prompting → talking to AI as naturally as you would a colleague.

🛠️ How to AI

Lead Segmentation & Prioritization

Advisors often sit on long lists of prospects without knowing where to start. AI can help you cut through the noise by segmenting and scoring leads so you focus on the right conversations first.

Try prompts like:

👉 Segmentation by conversion likelihood

“Given this list of prospects [paste data], segment them into groups by likelihood to convert based on career stage, life events, and financial complexity.”

👉 Scoring by urgency of financial needs

“Create a scoring framework to rank prospects based on urgency of financial needs. Include criteria such as stock option exercises, liquidity events, retirement timelines, or upcoming college expenses.”

📊 Before vs. After AI

Before (Raw List)

John, 42 — Sr. Manager at Amazon, recently promoted, has RSUs
Priya, 50 — Entrepreneur, just sold business
David, 38 — Engineer, buying first home
Maria, 47 — VP at Microsoft, two kids entering college soon

After (AI-Segmented & Scored)

High Priority: Priya (liquidity event → urgent planning), Maria (college expenses → near-term funding needs)
Medium Priority: John (equity comp planning soon)
Lower Priority: David (longer runway, simpler planning today)

💡 Why it matters: Instead of treating every lead the same, you let AI surface who needs you most, right now—helping you move faster, personalize outreach, and ultimately convert more high-value clients.

📌 Key Takeaways from this Issue

✅ OpenAI resets the baseline: With GPT-5 and new open-weight models, firms now get stronger reasoning, faster workflows, and the option to run AI securely on-prem for compliance and cost control.

✅ Prospecting shifts from chance to signals: Tools like FINNY and AI-triggered events (career changes, liquidity, milestones) make pipelines more targeted and predictable.

✅ AI fixes marketing misalignment: Automated ICP workflows help advisors align websites, messaging, and strategy with their most profitable clients.

✅ Better prompts, better outcomes: Clear, structured prompts turn raw AI outputs into tailored insights, helping advisors score, segment, and prioritize high-value leads.

👉 Forward this to a colleague who wants to stay ahead.

👉 Reply to share your biggest AI question — we might cover it next!

That’s a wrap for this week

Got thoughts or suggestions? I’d love your feedback to keep making Advisor Copilot better for you.

Cheers,

Neil Pradhan

Founder & Editor | Advisor Copilot