Week 6 Deep Research Prompt :: The Financing Readiness Investigation
-
Metadata
Topic: New vs. Certified Pre-Owned: Let AI Make the Case — The new-vs-CPO decision framework using Deep Research methodology
Week: Week 2 of 7 ("AI at the Dealership: 7 Weeks of Prompts That Could Save You Thousands")
Series: AI at the Dealership
Content Type: Deep Research methodology + prompt breakdown + follow-up prompts
Platform Compatibility: ChatGPT (GPT-4 with web search), Claude 3.5 Sonnet, Google Gemini (with Google Search)
Prerequisite: Week 1 — "Should I Buy a Car Right Now?" (recommended to have confirmed budget and TCO analysis from Week 1)
Tags: Deep Research, new-vs-CPO, vehicle financing, depreciation analysis, dealer tactics, CPO programs, total cost of ownership, financial decision-making
Categories: Car Buying, Financial Planning, AI Research Methodology
Difficulty Levels of Related Posts: Beginner (Week 2 ChatGPT variation), Intermediate (Week 2 Claude variation), Advanced (Week 2 Gemini variation)
Reading Time: 15-20 minutes to read this post; 8-15 minutes to run the prompt; 30-45 minutes total to work through the output and build your decision framework
SEO Title (under 60 characters): Deep Research: New vs. CPO — AI-Powered Analysis
SEO Description (150-160 characters): Use Deep Research mode to investigate the new-vs-CPO decision across eight research threads. Get a structured, investment-grade analysis with financing, depreciation, and dealer tactics covered.
Publication Date: April 13, 2026
Last Updated: April 13, 2026
Week 3 Deep Research: The Financing Readiness Investigation
Auto financing is where the dealership makes its real money and where the uninformed buyer loses theirs. The visible negotiation — sticker price, trade-in value, monthly payment — is a deliberate distraction from the invisible one: the spread between the buy rate the lender offered and the sell rate the dealer wrote into your contract, the $1,800 in extra interest hidden inside a 2% APR markup on a $35,000 loan, the $4,000 in negative equity quietly rolled into an 84-month note, the trade-in tax shield worth $1,200 in Texas but zero in California. These numbers do not reveal themselves on the showroom floor. They require investigation before you walk in. This week's Deep Research prompt is the investigation — an eight-thread inquiry into the financing landscape that arms you with credit-tier rate benchmarks, lender arbitrage math, negative-equity workout paths, and regulatory protections you can invoke when the F&I office leans on you.
Why Deep Research?
Deep Research mode is different from a standard chat conversation. Instead of asking an AI a question and getting a quick answer, Deep Research lets you ask an AI to investigate a topic by searching across multiple sources, synthesizing patterns, identifying conflicts, and building a structured analysis from the ground up. It's the difference between "What APR should I get?" (which produces a confident but generic answer) and "Investigate the current APR landscape by credit tier, quantify the dealer reserve markup, compare five lender categories for my situation, and build a pre-shopping action plan" (which produces an investment-grade financing intelligence brief with source citations).
This week, Deep Research matters because auto financing is an information-asymmetry game. The dealer knows their buy rate; you do not. The dealer knows the dealer reserve markup they're adding; you do not. The dealer knows which state tax treatments favor a trade-in and which punish one; you do not. They know the Money Factor on the lease they're pushing converts to an APR of 7.2%, not the "low monthly payment" they're showing you. Only systematic, multi-source research — comparing current Fed data on consumer credit, Experian's state-of-the-auto-finance-market reports, CFPB enforcement actions, individual lender rate sheets, and state DMV tax treatment — gives you symmetric information before you walk onto the lot. That's what this prompt is built to produce.
The Deep Research Prompt
Prompt Breakdown — How AI Reads the Deep Research Prompt
The Deep Research prompt above is dense by design. Every section does specific work, and understanding how the AI parses each block lets you adapt the architecture to any high-stakes decision where information asymmetry favors the seller.
"I need you to conduct comprehensive research on auto financing readiness for my specific situation and produce a structured, investment-grade intelligence brief..." — The opening sentence sets three critical expectations at once. It anchors to a specific person (not generic advice), it specifies the rigor level ("investment-grade intelligence brief" tells the AI to cite sources, build frameworks, and defend findings), and it implies the output is for use, not consumption. Saying "intelligence brief" rather than "summary" or "overview" flips the AI from Wikipedia mode into research-analyst mode.
Transferable principle: Begin research prompts by anchoring to your specific situation and defining the rigor level of the output. "Intelligence brief" produces evidence-based, cited, methodical work. "Overview" produces surface-level summaries.
"CONTEXT — MY SITUATION: [Budget ceiling, down payment, loan term, credit score, history length, monthly debt, gross income, state, vehicle category, trade-in details, self-employment status, banking relationships, prior experience]" — The context block provides all the variables the AI needs to personalize findings, delivered up front. This prevents the AI from asking follow-up questions (which wastes turns in a Deep Research cycle) and prevents it from making wrong assumptions that contaminate downstream calculations. Note the specificity: "exact, not estimate" on the credit score, "payoff balance" on trade-in, "list credit unions, banks, online lenders" on existing relationships. Each parameter is a lever the AI can pull in its analysis.
Transferable principle: For research prompts, provide every relevant variable upfront in a structured list with specific-precision cues ("exact," "balance," "list"). Don't make the AI play 20 questions. Precise inputs enable precise analysis.
"RESEARCH MISSION: Investigate and synthesize findings across eight core research threads... For each thread, search across multiple authoritative sources, compile findings, quantify the financial impact on my specific situation, note conflicts or uncertainties, and flag assumptions." — This segment defines the research architecture. It tells the AI three things the AI would not otherwise do: (a) search multiple sources rather than settling for one, (b) quantify financial impact on the user's specific numbers rather than generic examples, and (c) flag conflicts and uncertainties rather than smoothing them over. The phrase "note conflicts or uncertainties" is the antidote to the AI's natural bias toward confident-sounding synthesis.
Transferable principle: Define the research architecture in the mission statement. Explicitly demand multi-source synthesis, quantified application to your numbers, and transparency about conflicts. Otherwise the AI will produce a confident-sounding average of its training data.
"RESEARCH THREAD 1 — MY CREDIT SCORE vs. THE SCORE THE DEALER WILL PULL: Which credit score version do auto lenders typically pull? How does FICO Auto Score differ from standard FICO 8? What free resources show me my actual Auto Score?..." — Each of the eight threads is structured identically: a one-sentence thread title identifying the research domain, followed by five to seven specific questions that define the sub-dimensions to investigate, followed by a named source list. This three-layer structure (title → questions → sources) is the architectural pattern that forces comprehensiveness. Without the questions, the AI might stop at surface level; without the sources, the AI might fabricate citations; without the title framing, the AI might drift into adjacent topics.
Transferable principle: Structure each research thread as title + sub-questions + named sources. Each layer disciplines a different AI failure mode: stopping early, hallucinating, or drifting. Use all three for maximum research quality.
"RESEARCH THREAD 3 — DEALER RESERVE ECONOMICS: What does a 2% markup actually cost on my specific loan? (Calculate: $35,000 at 60 months, 6% buy rate vs. 8% sell rate = $2,018 extra interest)" — Notice the parenthetical inside Thread 3: the prompt shows the AI what a quantified finding should look like. This is "example-driven specification" — when you need the AI to produce a specific output format, embedding a worked example inside the prompt is more reliable than describing the format abstractly. Here, the AI sees "calculate on my parameters the way this example calculated on these parameters" and produces work of matching rigor.
Transferable principle: Embed worked examples inside research threads when you need a specific output format. "Calculate X (example: Y parameters yield Z result)" produces more reliable outputs than "Calculate X in detail."
"ANALYSIS & DELIVERABLES: After researching these eight threads, synthesize findings into four deliverables: Executive Summary, Research Findings by Thread, Lender Arbitrage Matrix, Pre-Shopping Action Plan." — The deliverables section transforms research into action. Deliverable 3 (Arbitrage Matrix) is a table template the AI must fill in precisely — you provided the columns, so the AI cannot improvise a different format. Deliverable 4 (Pre-Shopping Action Plan) is a week-by-week calendar, so the AI must turn abstract findings into a concrete sequence of steps. This is the difference between a research paper and an operations plan.
Transferable principle: Specify deliverables with structure templates (pre-defined columns, pre-defined timelines, pre-defined headers). Templates force the AI to convert insights into operationalized outputs rather than narrative summaries.
"CONSTRAINTS FOR THIS RESEARCH: Search across multiple sources; if sources conflict, note the conflict... Every financial claim must be attributed to a specific, named source... Do not assume the dealer's quoted rate is fair — benchmark every dealer number against pre-approval buy rates..." — The constraints section is where you preempt the AI's natural failure modes. The attribution requirement preempts hallucinated citations. The conflict-noting requirement preempts false synthesis. The "do not assume the dealer's quoted rate is fair" sentence is domain-specific intellectual honesty: it tells the AI to treat dealer-provided numbers skeptically rather than credulously. Every constraint is a patch over a specific way the AI would otherwise disappoint you.
Transferable principle: Write the constraints section as a list of "if the AI does X bad thing, this constraint blocks it." Name the failure modes explicitly: hallucinated citations, glossed conflicts, credulous reception of counterparty data. Each becomes a constraint.
"TONE & STRUCTURE: ...Do not hedge with 'it depends' — state the conditional explicitly and resolve it based on my stated parameters." — The final section disciplines the presentation. The anti-hedging instruction is especially important for financing research: the AI's default behavior on ambiguous financial questions is to say "it depends" and list factors. That's useless to a decision-maker. The prompt instructs the AI to resolve conditionals based on the provided parameters: "If you are in Texas with a trade-in, then [specific answer]; if you are in California, then [specific answer]." Specificity over hedging.
Transferable principle: Explicitly prohibit hedging phrases ("it depends," "both have pros and cons") and require conditional resolution against your stated parameters. The AI's default is to hedge; the instruction to resolve forces it into useful specificity.
What to Expect from Deep Research
Output Length: Expect 10,000-18,000 words of output, depending on the depth of research available and the specificity of your inputs. The Executive Summary alone will be 400-600 words. Each of the eight research threads will produce 800-1,500 words with sub-findings, source attribution, and implications. The Lender Arbitrage Matrix is a full-page table. The Pre-Shopping Action Plan is a detailed two-week calendar with 15-25 specific actions.
Completion Time: Deep Research on ChatGPT (with o3 Deep Research), Claude (with web search/research mode), or Gemini (with Deep Research) typically takes 5-12 minutes to execute for a prompt of this scope. The research phase runs invisibly — the AI searches across Experian reports, Federal Reserve data, CFPB filings, state DMV sites, and lender rate sheets — and then synthesizes the findings. You don't see the searching; you see the final structured brief.
Structure: Output is organized by deliverable, with clear h2/h3 headers, source citations inline or in footnotes, and explicit "So what does this mean?" implications at each thread's close. The Arbitrage Matrix will be a scannable table. The Action Plan will be a week-by-week checklist you can print and execute against.
Quality Signals: High-quality Deep Research output includes specific numbers tied to sources (not round figures), conflicting data points with explanations (e.g., "Experian reports Prime tier median at 6.27%; Bankrate reports 6.8%; conflict likely due to different source pool dates"), assumptions stated explicitly, and clear links between findings and action items. If the output is heavy on generic advice, light on source citations, or hedges every claim with "it depends," the rigor is insufficient — ask the AI to redo the weak threads with named sources.
Key Research Questions the Prompt Answers
1. What APR should I actually expect at my credit tier, and how much extra would I pay at the tier below mine? Research Threads 1 and 2 resolve the credit-score-to-rate translation and quantify the cost of being in a lower tier. This is the foundation of every downstream decision.
2. How much is the dealer quietly adding to my interest rate as their profit? Research Thread 3 exposes the dealer reserve markup — up to 2.5 percentage points above the buy rate — and calculates the dollar impact on your specific loan. This is the single most expensive hidden fee in auto buying, and most buyers never see it.
3. Which lender gives me the lowest rate given my credit profile and existing relationships? Research Thread 4 maps the five lender categories (credit union, bank, online, captive, dealer indirect) to your situation, producing the Lender Arbitrage Matrix that shows exact expected APRs and total interest for each path.
4. If I have negative equity, is it cheaper to pay it in cash, roll it into the new loan, or delay my purchase? Research Thread 5 models all three workout paths with concrete math on your current loan balance and vehicle value. This decision alone can cost or save $3,000-$10,000 over the life of a new loan.
5. Am I better off trading in to the dealer or selling privately, given my state's tax treatment? Research Thread 6 applies your state's specific trade-in tax rules — full credit in Texas, zero in California, $10,000 cap in Illinois — to your specific numbers. The answer flips by state.
6. If the dealer pushes me toward a lease, what APR equivalent am I actually paying? Research Thread 7 converts Money Factor to APR (multiply by 2,400) and compares total-cost-of-ownership across lease vs. finance paths. Leases can be great or terrible depending on specifics; the math is the only arbiter.
7. What federal and state regulations protect me, and how do I invoke them if the dealer crosses a line? Research Thread 8 covers Truth in Lending Act disclosure requirements, FTC CARS Rule anti-junk-fee protections, CFPB enforcement trends, and state-level auto lending laws. These are your legal shield if a dealer behaves badly.
8. If I end up taking dealer financing for a rebate, when and how do I refinance? Research Thread 8 also covers the refinance window — typically 30-60 days after purchase — and the conditions under which refinancing mathematically works (credit improvement, rate market movement, no prepayment penalty).
Platform-Specific Tips for Accessing Deep Research
ChatGPT (GPT-5 Deep Research, or GPT-4 Turbo with web search): ChatGPT Plus and Pro include Deep Research mode, which is specifically designed for prompts like this one. Select "Deep Research" from the model picker (or tools menu) before pasting the prompt. GPT-5 Deep Research will search across Experian, Federal Reserve, CFPB, individual lender sites, and state DMV pages, and typically returns output in 5-10 minutes with inline citations. Expect strong results across all eight threads. If you're on a standard tier without Deep Research, GPT-4 Turbo with web search enabled will produce a compressed version of the same output with somewhat fewer citations.
Claude (Claude Opus 4.6 with Research mode): Claude's Research feature (available on Pro and Max tiers) is Anthropic's equivalent to Deep Research. Enable it before pasting the prompt. Claude excels at structured synthesis — it will produce the clearest Arbitrage Matrix and the most rigorous assumption documentation — and will explicitly flag data gaps rather than paper over them. Claude's research output is typically more conservative on rate benchmarks (it tends to cite ranges rather than point estimates), which is an advantage for financial decisions. Output time: 4-8 minutes. Without Research mode, Claude's training-data cutoff means rate data may be dated; supplement with ChatGPT or Gemini for current-rate threads.
Gemini (Gemini 2.5 Pro with Deep Research): Gemini's Deep Research is natively integrated with Google Search and produces the most systematic source coverage of the three platforms. Click "Deep Research" before pasting the prompt. Gemini will produce a plan preview (which threads it will investigate, in what order) before executing — review the plan and confirm before it runs. Expect strong performance on rate benchmarks (Thread 2), state tax treatment (Thread 6), and regulatory research (Thread 8), where live search access is decisive. Output time: 7-15 minutes because Gemini searches more exhaustively. Output tends to be more journalistic and reader-friendly than Claude's, which is an advantage for comprehension but occasionally less analytically rigorous.
Pro Tip — Multi-Platform Workflow: For maximum rigor on a high-stakes financing decision, run the prompt on two platforms sequentially. Start with Gemini or ChatGPT Deep Research for the data-heavy threads (current APRs, state tax law, CFPB enforcement actions). Then feed that research into Claude with the instruction: "Take these research findings, apply them to my specific parameters, and produce the Arbitrage Matrix and Pre-Shopping Action Plan with maximum analytical rigor. Flag every assumption." Claude will catch gaps the other platforms missed and build cleaner structured output. Total time: 20-30 minutes. Value delivered: institutional-grade financing intelligence before you ever call a dealer.
How This Connects to the Weekly Posts
This Deep Research prompt is the investigation layer of Week 3. The three platform-specific posts (ChatGPT, Claude, Gemini) teach you three different prompt variations for financing-readiness at Beginner, Intermediate, and Advanced difficulty — those prompts produce tactical, week-of-shopping outputs like pre-approval checklists, multi-lender comparison tables, and F&I negotiation scripts. The Deep Research prompt on this page goes deeper: it's designed for buyers who need the source-cited, benchmark-level data that backstops their tactical decisions. Where the weekly posts give you the playbook, this prompt gives you the scouting report.
Week 3 builds on Week 1's confirmed budget (from the "Should I Buy a Car Right Now?" prompts) and Week 2's new-vs-CPO decision (from "New vs. CPO: Let AI Make the Case"). If you completed those, you have a vehicle budget and a new-or-used choice. This week's Deep Research takes those decisions and builds the financing architecture that turns them into a contract you can defend line-by-line against the F&I office. The cross-platform comparison post for Week 3 shows which of the three AI platforms produced the strongest tactical output (Gemini won at 87.5, with Claude as runner-up at 85.5); if you're choosing where to run this Deep Research prompt, those rankings are a useful proxy for where the general Week 3 outputs were strongest.
Adaptability Tips: Using This Prompt for Other Decisions
1. Mortgage Pre-Approval Strategy (Housing): Replace auto-specific research threads with mortgage equivalents. Thread 1 becomes FICO mortgage scores (FICO 2, 4, 5) vs. standard FICO — a 20-point gap is common. Thread 2 becomes current mortgage APRs by tier and loan program (conventional, FHA, VA, jumbo). Thread 3 becomes mortgage broker compensation structures and yield spread premium. Thread 4 becomes the five lender categories for mortgages (credit union, bank, online, mortgage broker, wholesale). Thread 5 becomes PMI and down-payment mathematics. Thread 6 becomes state-specific closing costs and transfer taxes. Thread 7 becomes fixed vs. adjustable-rate mathematics. Thread 8 becomes CFPB mortgage regulations and refinance timing. The eight-thread architecture transfers directly; only the specific sub-questions change.
2. Commercial Equipment Financing (Business Capital): Adapt the prompt for a capital equipment purchase like a commercial truck, restaurant oven, or medical device. Replace credit-tier research with commercial credit evaluation (business credit score, personal guarantor analysis, DSCR). Replace OEM captive with equipment finance companies (CIT, Wells Fargo Equipment Finance, manufacturer financing arms). Add a Section 179 tax deduction thread and a bonus depreciation thread. Add a lease-vs-finance thread specifically evaluating capital vs. operating lease treatment. The regulatory thread shifts from CFPB to UCC filings and state commercial lending laws. The Arbitrage Matrix still applies; it just compares equipment finance companies instead of auto lenders.
3. Private School Tuition Financing (Education): For families financing private school or college through loans, replace auto threads with education-specific ones. Thread 1 becomes parent PLUS vs. private student loans vs. personal loans vs. 529 plan withdrawals. Thread 2 becomes current education loan APRs by type. Thread 3 becomes origination fees and capitalized interest mechanics. Thread 4 becomes the five lender categories (federal, private, credit union, school-arranged, tuition-payment-plan). Add a threads for financial aid optimization and tuition tax deductions. The output is a financing strategy for the school year, parallel to the auto pre-shopping action plan.
4. Medical Procedure Financing (Healthcare): For elective procedures, dental work, or out-of-network specialists, adapt the prompt to medical financing. Replace credit-tier research with CareCredit / medical-credit-card underwriting (they pull FICO but with different overlays). Add a thread on HSA/FSA eligibility for the specific procedure. Add a thread on negotiating cash-pay rates with providers (often 30-50% below billed charges). The regulatory thread shifts to the No Surprises Act, state medical debt protections, and CFPB medical debt research. The Arbitrage Matrix compares medical credit cards, personal loans, HSA/FSA payment, and provider-direct financing.
Follow-Up Prompts
Follow-Up 1 — "Build the Dealer F&I Negotiation Script": Once you have the Deep Research output and the Arbitrage Matrix, ask: "Using the rate benchmarks and lender arbitrage findings from the Deep Research, generate a word-for-word negotiation script for the dealership F&I office. Cover three scenarios: (a) the dealer offers an APR higher than my pre-approval, (b) the dealer offers an APR matching my pre-approval, (c) the dealer offers an APR lower than my pre-approval. For scenario (c), include a specific audit script to confirm the low rate is not contingent on mandatory backend products like extended warranties or GAP insurance." This turns research into battle-ready dialogue.
Follow-Up 2 — "Stress-Test the Recommendation": Ask: "The Deep Research recommends [specific lender path]. Now stress-test that recommendation by changing one variable at a time: (a) my credit score drops 40 points between pre-approval and contract signing, (b) the Federal Reserve raises rates 0.75 points during my shopping window, (c) my trade-in's market value drops $2,000 because of a new manufacturer model launch, (d) the dealer refuses to accept my outside financing. For each scenario, show how the recommendation changes and what backup plan is optimal." This produces decision resilience against real-world noise.
Follow-Up 3 — "Build the TILA Contract Review Checklist": Ask: "Build a specific 15-point Truth in Lending Act contract review checklist I will execute at the F&I desk before signing. For each item, state (a) the TILA box or contract field to inspect, (b) what the correct value should be given my pre-approved terms, (c) what to say if the field does not match, and (d) the federal regulation I cite if the dealer pushes back. Include checks for amount financed, APR, finance charge, total of payments, prepayment penalty clauses, capitalized fees, force-placed GAP insurance, and arbitration clauses." This produces the legal-defense overlay for your contract signing.
Metadata
Topic: Getting Your Money Right Before You Shop — Financing Readiness via Deep Research
Week: Week 3 of 7 ("AI at the Dealership: 7 Weeks of Prompts That Could Save You Thousands")
Series: AI at the Dealership
Content Type: Deep Research methodology + prompt breakdown + follow-up prompts
Platform Compatibility: ChatGPT Deep Research (GPT-5 or GPT-4 Turbo with web search), Claude Research mode (Opus 4.6 / Sonnet 4.6), Google Gemini Deep Research (Gemini 2.5 Pro)
Prerequisite: Week 1 ("Should I Buy a Car Right Now?") for confirmed budget; Week 2 ("New vs. CPO") for vehicle category decision. Recommended to complete both before running this prompt.
Tags: Deep Research, auto financing, pre-approval strategy, dealer reserve, lender arbitrage, credit tier APR, negative equity, trade-in tax shield, Money Factor, Truth in Lending Act, CFPB, F&I negotiation
Categories: Car Buying, Financial Planning, AI Research Methodology, Consumer Protection
Difficulty Levels of Related Posts: Beginner (Week 3 ChatGPT variation), Intermediate (Week 3 Claude variation), Advanced (Week 3 Gemini variation); this Deep Research post sits above the Advanced tier for research-intensive buyers.
Reading Time: 18-22 minutes to read this post; 8-15 minutes to run the prompt; 45-60 minutes total to work through the output and execute the Pre-Shopping Action Plan
SEO Title (under 60 characters): Deep Research: Auto Financing Readiness — AI Intelligence Brief
SEO Description (150-160 characters): Use Deep Research to investigate auto financing across eight research threads — credit tiers, dealer reserve, lender arbitrage, negative equity, tax shields, TILA protections.
Publication Date: April 20, 2026
Last Updated: April 20, 2026
