RSA Generator from Brief

Variables

Replace each placeholder before you run the prompt.

• {{product_brief}}Product, key features, differentiators, certifications, supporting proof.

  Industrial laser markers. Features: 100W fiber, integrated I/O, 24/7 service contract. Differentiators: US-built, ISO 9001, 4-week lead time. Proof: 2,400+ deployments, top OEMs.

• {{persona}}The buyer persona — title, pain points, internal vocabulary.

  Manufacturing engineer. Cares about: throughput, downtime, traceability compliance. Hates: vendor lock-in, long lead times, generic SaaS demos.

• {{campaign_intent}}What the campaign is supposed to do — the commercial outcome.

  Drive RFQ submissions from manufacturing engineers at OEMs evaluating new laser systems for line traceability.

Write Google Ads RSA copy for a B2B account. Stay in the persona's voice. Avoid marketing fluff.

PRODUCT:
{{product_brief}}

PERSONA:
{{persona}}

CAMPAIGN INTENT:
{{campaign_intent}}

OUTPUT:

15 headlines (≤30 chars). Mix of:
- Feature-led (3): name a specific feature with a number where possible
- Benefit-led (3): connect to a buyer pain point
- Proof-led (3): cert, deployments, lineage
- Process-led (3): lead-time, RFQ, spec-review hook
- Brand+vertical-led (3): "[Brand] for [vertical]" framing

4 descriptions (≤90 chars). Mix of:
- One feature-stack
- One outcome-led
- One proof-led
- One process / RFQ hook

Each output: write the asset, then a 1-sentence note on the angle and which character of the persona it speaks to.

Do NOT use: "leading", "premier", "innovative", "revolutionary", "cutting-edge", "best", "trusted", "industry-standard". Strike them on sight.

Expected output shape

15 headlines + 4 descriptions, each with a 1-sentence angle note.

Why we wrote it

Most RSA generation prompts produce broadly applicable copy that nobody loves. Anchoring to brief + persona + campaign intent forces specificity.

How to use

1. Open Claude or ChatGPT. The recommended model for this prompt is claude-sonnet-4-6 — opus when the prompt requires deep reasoning, sonnet for the rest.
2. Replace every {{variable}} with content specific to your account. The examples above are starting points, not templates to ship as-is.
3. Paste the prompt and run.
4. Read the output against the expected shape above. If the model produced a structurally different response, re-prompt rather than accept the drift.