← Research Notes中文
Technology · 2026-07-15 · 8 min read · Close read

Giving an LLM a Persona: An Evidence Review of Product Requirements Judgment and Routing

"You are a senior product manager" — almost everyone who has used an LLM for requirements analysis has written an opener like that. Assign the model a role — role cognition (persona) — then have it understand requirements, judge priorities, and route tickets: intuitively, it makes sense. But the empirical research from 2023–2026 gives a strikingly counterintuitive answer: personas don't make judgment more accurate, and some personas make it worse; and when you organize personas into a pipeline that passes requirements along, independent testing found the "product manager persona" itself rewriting your requirements. This piece lays the evidence out, both sides of it.

Research statement

This article is an evidence review of "persona / role-based LLMs for understanding, judging, and routing product requirements": 25 conclusions in total, each of which survived three-vote adversarial verification (another 5 "plausible-sounding" claims were rejected after checking against the original sources). The evidence is graded in three tiers: [Verified] = checked word-for-word against first-hand papers/documentation; [Judgment] = analysis derived from the verified evidence. The uncovered gaps are disclosed honestly at the end. This is a snapshot of a fast-moving field (mid-2026), not a verdict on what to adopt.

The overall verdict up front: persona/role does not improve the "accuracy" of judgment — it only changes the judgment's "perspective and behavior distribution"; and for organizing roles into a pipeline that passes requirements along, the independent evidence shows the harm outweighs the benefit. The argument comes in four parts.

1. The Foundation: The Empirical Science of Persona Prompting [Verified]

2. Persona-Simulated Users for Requirements Discovery [Verified]

3. Role-Based Multi-Agent Pipelines: Self-Reported vs. Independent Testing [Verified]

MetaGPT (PM → architect → engineer → QA) and ChatDev organized "persona/role" into a software-company-style pipeline. The evidence on both sides is now hard:

The case for (all self-reported by the framework authors)

The case against (independent tests, all pointing the same way)

4. Uncovered Gaps (Disclosed Honestly)

The following three areas yielded no citable verified data in this research; where they come up, treat them with qualitative common sense: (1) concrete benchmark numbers for how well requirements classification / prioritization agrees with human experts; (2) industrial deployment numbers for ticket routing (official accuracy figures from Zendesk / Atlassian and the like); (3) primary quantification of LLM-as-judge position / verbosity / self-preference biases.

5. Feasibility Matrix and Practical Recommendations [Judgment]

TaskApproachRole of the persona
Requirements classification / clustering / first-pass routingSingle model, fully automated + structured output + eval set❌ No accuracy dividend — don't spend the budget on persona design
Perspective coverage check (who benefits / who loses)Rotate "new user / heavy user / paying user" perspectives through the same model✅ The only evidence-backed win zone: changing the perspective, not improving accuracy
Requirements-space exploration / interview hypothesis generationSimulated users grounded in verbatim real feedback⚠️ Exploration only, never validation; don't invent personas out of thin air
Priority recommendationsSingle model + confidence output, human makes the call❌; a reviewer role's pushback strength must be explicitly calibrated
Multi-role pipelines auto-passing requirements alongDon'tIndependent evidence: every role handoff is distortion, at 6–12x the cost
Final trade-offs / payment validation / strategyHumansSimulated intent ≠ willingness to pay; zero data supports substitution

The minimal setup for a solo founder / small team: (1) requirements triage pipeline = one strong model + structured output + a small eval set, with low-confidence cases forced to a human; (2) use personas in exactly one place — run the "who benefits / who loses" coverage check across three user perspectives for every candidate requirement; (3) for the review pass, use "calibrated, gentle pushback," not "forced disagreement"; (4) use simulated users only to generate interview questions — validation always goes to real people.

Primary sources (all verified against the original texts): Zheng et al., "When 'A Helpful Assistant' Is Not Really Helpful" (EMNLP 2024); Gupta et al., "Bias Runs Deep" (ICLR 2024); Kong et al. (NAACL 2024); Wang et al. (ACL 2025 Findings); Argyle et al., "Out of One, Many" (Political Analysis 2023); Park et al., "Generative Agent Simulations of 1,000 People" (2024); Wang et al. (Nature Machine Intelligence 2025); CoMPosT (EMNLP 2023); MetaGPT (ICLR 2024); ChatDev (ACL 2024); E2EDev (2025); MAST (NeurIPS 2025); Wang et al. (ACL 2024); Smit et al. (ICML 2024). Roughly 215 search-and-verification agents across three research rounds; 25 conclusions survived, 5 were rejected; "Verified / Judgment" tiers, self-reported numbers, and points of dispute are flagged item by item.

Amos · research.xishe.ai · Please credit when sharing