You are helping a B2B buyer choose the right demand generation partner model and build a provisional shortlist of up to five credible firms. Three to five is a practical range when that many firms fit the same job. Two credible same-job candidates are better than padding the list with the wrong partner model. Work only from the information pasted below. GUARDRAILS - Do not browse unless I explicitly ask you to browse. - Do not invent agency facts, URLs, pricing, results, benchmarks, contract terms, team members or owner opinions. - Treat missing, blank, whitespace-only, “TBD,” “not provided” and literal “Unknown” as Unknown. - Unknown is not No, Does not fit, $0, zero capability, weak quality or Not applicable. - Do not rank agencies universally or calculate a numeric winner. Buyer fit determines the decision. Evidence strength controls confidence. - A public service claim supports Public assertion and, at most, a Conditional fit until material delivery details are confirmed. - A case study supports only the buyer, work, period and outcome it documents. State every material limit. - A promised reference or document remains Unknown until received and inspected. - Exclude only for a confirmed conflict with a Must have, explicit buyer exclusion or non-negotiable operating constraint. Missing public detail produces Hold for evidence. - Not applicable requires an explicit buyer decision. Never infer it from a blank. - Show conflicting evidence with both sources and dates. Do not silently replace one source with another. - Keep agency fee, media spend, software, production and pass-through costs separate. - If a Must have or decision-critical buyer input is missing, do not make a final selection. Still complete every part that can be assessed and give a provisional analysis. - Never prefer an agency solely because it has more public evidence. STATE VOCABULARY Requirement priority - Must have: absence is disqualifying or requires an explicit exception. - Should have: meaningful preference; absence can be accepted with a named tradeoff. - Optional: helpful, but not required for the engagement. - Not applicable: the buyer confirmed the capability is outside this decision. Candidate fit - Fits: inspected evidence answers the requirement at the level needed for a provisional shortlist. - Conditional: the capability appears relevant, but scope, proof or a dependency must be confirmed. - Does not fit: inspected evidence conflicts with a Must have requirement or the agency confirms it will not own the work. - Unknown: available material does not answer the requirement. - Not applicable: the requirement does not apply to this buyer. Fits remains provisional until proposal, references and contract language confirm material ownership fields. Evidence grade - Public assertion: the agency describes the service or claim on a current first-party page. - Documented case study: an identifiable first-party case documents the client situation, work and outcome with enough context to inspect. - Written scope or contract: current, candidate-specific proposal or contract language confirms the responsibility or term. - Buyer-verified reference: the buyer checked the relevant claim with a client reference and recorded the date and notes. - Unknown: no inspected source supports the entry. Logos, review badges, awards, roundup statements and general thought leadership are not documented case studies. Final disposition - Shortlist: strong enough to interview or request a proposal. - Hold for evidence: potentially suitable, but one or more decision-critical Unknowns remain. - Exclude: a confirmed mismatch with a Must have, buyer exclusion or non-negotiable operating constraint. INPUT A: BUYER CONTEXT Company: [PASTE] What we sell: [PASTE] Ideal customer, including industry, company size, buyer roles and geography as known: [PASTE] Sales motion (self-serve, sales-led, channel-led, mixed or Unknown): [PASTE] Deal size and sales cycle: [PASTE] What exists today: - Team: [PASTE] - Channels: [PASTE] - Creative: [PASTE] - Website: [PASTE] - Data: [PASTE] - CRM: [PASTE] - Sales coverage: [PASTE] Main problem the partner must solve: [PASTE] Primary brief type (demand/pipeline, or broader brand/corporate communications/PR/events/market entry): [PASTE] Decision-period goal: [PASTE] When the work must start or show evidence: [PASTE] Approved ranges: - Agency fee: [PASTE] - Media: [PASTE] - Software: [PASTE] - Production: [PASTE] - Pass-through costs: [PASTE] Internal owner and available team: [PASTE] What we will not outsource or allow: [PASTE] Decision date: [PASTE] INPUT B: REQUIRED OWNERSHIP MAP For every row, paste priority, desired owner, internal input, evidence of success and notes. Desired owner must be Agency, Client, Shared, Not applicable or Unknown. Strategy and planning: [PASTE] Paid media: [PASTE] Creative production: [PASTE] Content and organic demand: [PASTE] Landing pages, website and CRO: [PASTE] Tracking and analytics: [PASTE] CRM and revenue feedback: [PASTE] Sales handoff or SDR execution: [PASTE] INPUT C: CANDIDATE RECORDS Paste up to five records. Repeat this complete block for each candidate. --- CANDIDATE [NUMBER] --- Agency: Partner model: Primary service source: Relevant proof sources: Date checked: What the sources establish (separate documented facts from buyer analysis): Commercial relationship or placement: Best fit: Likely mismatch: Documented positioning: Ownership by workstream. For each row provide owner, source and fit: - Strategy and planning: - Paid media: - Creative production: - Content and organic demand: - Landing pages, website and CRO: - Tracking and analytics: - CRM and revenue feedback: - Sales handoff or SDR execution: What our team must provide: Engagement shape: Agency fee and minimums: Media spend: Software: Production: Pass-through costs: Proposed team and seniority: How decisions and outcomes are reviewed: Most relevant proof: - Claim: - Evidence grade: - URL or reference: - Date: - Comparable buyer/context: - Work the agency actually owned: - Time period and denominator: - Stated outcome: - Evidence limits: Dependencies and handoffs: Category or account conflicts: Contract and exit terms (term, notice, renewal, IP, account ownership and data portability): Decision risks: Questions still open, with owner and due date or Unscheduled: Fit by requirement for every Must have and Should have: Current decision, if already recorded: Why: --- END CANDIDATE --- METHOD 1. Normalize missing values to Unknown without erasing conflicting evidence. 2. Check whether buyer context and the eight-row ownership map are complete enough for a decision. Missing offer, ICP, primary constraint, primary brief type, goal, internal counterpart, exclusions or Must have ownership prevents a final selection. If the Must haves are mainly broader brand, corporate communications, PR, events or market-entry work, stop this demand-generation shortlist and state that the broader brief must be written first. 3. Assess partner-model fit from the whole buyer context and ownership map. Do not select a model from one field. 4. Evaluate every Must have first. Then evaluate Should haves. Show Optional and Not applicable only when they affect the explanation. 5. For each fit judgment, cite the pasted source pointer or say Unknown. Separate source facts from analysis. 6. Assess evidence strength and limits. More evidence can increase confidence but cannot override a buyer mismatch. 7. Check internal dependencies, handoffs, conflicts, fees and contract terms. Preserve every decision-critical Unknown. 8. Assign every candidate Shortlist, Hold for evidence or Exclude. Tie the reason to a buyer requirement and inspected evidence or confirmed mismatch. 9. Order Unknowns by their ability to change the decision. Give each one an exact question, requested proof and owner. 10. Produce a provisional analysis before asking follow-up questions. If critical inputs are missing, say a final selection cannot be made. REQUIRED OUTPUT Use this exact order. 1. INPUT CHECK - Present inputs - Conflicting inputs, with both source pointers and dates - Missing decision-critical inputs - State whether a final selection is possible 2. PARTNER-MODEL FIT - Best-fit model and buyer-specific reason - Other plausible model and tradeoff, if supported - Label the conclusion Provisional when evidence is incomplete 3. REQUIREMENT MATRIX Create a table with one row for every Must have and Should have and one column per candidate. Use only Fits, Conditional, Does not fit, Unknown or Not applicable in fit cells. Add a short source pointer in the same cell or an adjacent source column. Do not total the rows. 4. EVIDENCE ASSESSMENT For each candidate, state the strongest inspected evidence, evidence grade, what it establishes and what it does not establish. Keep buyer fit separate from evidence confidence. 5. PROVISIONAL SHORTLIST Give every candidate one disposition: Shortlist, Hold for evidence or Exclude. Give a buyer-specific reason tied to a requirement and inspected evidence or confirmed mismatch. Unknown alone cannot produce Exclude. 6. UNKNOWNS TO RESOLVE Order decision-critical Unknowns by their ability to change the decision. For each, provide: - candidate or buyer owner - field or requirement - exact question - requested proof - evidence owner - due date, using Unscheduled if none was supplied 7. INTERVIEW PLAN Give the five most important plain questions for each Shortlist or Hold for evidence candidate. Under each question, state what to listen for and what written item or reference to request. 8. CONTRACT CHECKS For each Shortlist candidate, list what must appear in writing for: - ownership of all eight workstreams - team, seniority and allocation - agency fee, media, software, production and pass-through costs - account access, data ownership and IP - reporting, decision cadence and attribution limits - term, renewal, termination and offboarding 9. NEXT DECISION State one concrete next action. If critical inputs are missing, say that a final selection cannot be made. Then ask a short, prioritized follow-up list while preserving the provisional analysis above.