GenStudio Email Prompts
Produce paste-ready prompts for Adobe GenStudio for Performance Marketing email experiences. Output the prompt only (plus a short GenStudio Parameters checklist when helpful)—not the email copy itself unless asked.
When to use
- User wants a GenStudio prompt for promotional, nurture, or educational email
- Single product or multi-product (multipod) emails using
Pod1,Pod2, etc. - Refining a weak prompt into GenStudio’s structured format
- Generated variants scored low on Brand / Content check and the prompt needs a rewrite
Products (GenStudio Parameters)
Use only these product guidelines unless the user specifies another. Full descriptions, value props, and messaging preferences: products.md.
| Product | Notes |
|---|---|
| Red Hat Enterprise Linux | Hybrid OS foundation; lead with why RHEL (not why Linux); hardened foundation and closed-loop remediation; hybrid consistency; Lightspeed for ops, risk, and compliance outcomes |
| Red Hat OpenShift Platform Plus | Prefer this name (not “OpenShift” alone); hybrid app platform at scale |
| Red Hat Ansible Automation Platform | Prefer this name (not “Ansible” or AAP); enterprise automation at scale; Lightspeed for skills gap, not the whole story |
| Red Hat Developer program | No-cost membership; peer-to-peer; Join / Start building CTAs |
| Red Hat product trial | No-cost, typically 60-day, full-subscription-value evaluation—not production; not the Developer program |
Map each pod to one product/offer focus when possible. Briefs that say “OpenShift” map to Red Hat OpenShift Platform Plus unless another edition is named. Briefs that say “Ansible” map to Red Hat Ansible Automation Platform unless they name community Ansible, Ansible Core, or AWX.
Personas (GenStudio Parameters)
Select from the GenStudio persona repository (WIP—not exhaustive). Full descriptions and messaging preferences: personas.md.
| Persona | Role examples |
|---|---|
| Champion | System Administrator, Data Science Lead, Automation Architect, AppDev ITDM, Lead Business Analyst |
| Technical Practitioner / Architect | Cloud Architect, DevOps Engineer, SysAdmin, Site Reliability Engineer |
| Developer | Enterprise Software Engineer, Full-Stack Developer, Cloud-Native Developer, Application Architect |
If the user says “Technical Practitioners & Influencers,” map to Technical Practitioner / Architect (and Champion when influencer/advocacy framing is needed). If they say Selectors (tactical acquire-and-use choices), map to Technical Practitioner / Architect. Note the assumption.
Email channel limits & style
Enforce Red Hat channel guidelines in pod directives and when reviewing generated copy. Full rules: channel-guidelines.md.
| Element | Limit |
|---|---|
| Subject line | 30–40 characters max |
| Preheader | 40–60 characters |
| Headline | 30 characters max |
| Sub-headline | 45 characters max |
| Body | Max 3 sentences per pod; outcomes over description |
| CTA | 20–25 characters max |
Style: scannable; customer benefit first; active voice; second person when appropriate; Oxford commas; contractions allowed in email; avoid vague words (e.g. “flexible,” “scalable”); entire email—no “free/win/unlock” (use “no-cost”); CTA verbs like Download, Register, Start, Explore, Watch—never “click here.”
Body rhythm (Brand → Channel guidelines → Email → Body—not the prompt): avoid long, information-dense sentences with similar length; do not start the body with “You.” Use these when reviewing generated copy.
Brand voice (Brand → Tone of voice and Brand values—not the prompt): stability and hybrid cloud engineering; community-driven innovation; practical, no-hype technical voice; operational problem then guidance; Red Hat as enabler. Values: Open, Authentic, Helpful, Brave (no competitor attacks). Canonical text: brand-guidelines.md. Do not put Tone’s word flexibility in email prompts—Channel still bans flexible.
Brand editorial (Brand → Editorial guidelines and Editorial restrictions—not the prompt): full product name first in body; approved abbreviations (e.g. RHEL) in subject/headline; Red Hat as “it” or “we/our,” not “they.” Restrictions: no uncited superlatives; no “secure/more secure”; no AI-typical wording (including testament / tapestry); no vague words without a direct object; no “the” before product names. Use when reviewing generated copy. Canonical text: brand-guidelines.md.
In prompts: bake length into pod lines when helpful, e.g. Pod1: In 300-400 characters… or Pod2: In 2 sentences maximum… (body still ≤ 3 sentences per pod). Do not paste Email channel field guidelines, Brand voice, Brand values, or Brand editorial lists into the prompt (Parameters already inject them). Do rewrite the brief so restricted words never appear in the prompt—GenStudio echoes prompt wording into copy, and Brand score is % of Brand guidelines passed vs tested on that copy. Substitution table: brand-guidelines.md.
Inputs to collect
If missing, ask briefly—or infer and note assumptions:
| Input | Notes |
|---|---|
| Email type | Single-section or multipod |
| Goal / CTA intent | Motivate, educate, drive trial, standardize, partner action |
| Persona | Champion, Technical Practitioner / Architect, or Developer |
| Product(s) | From the list above; one per pod when multipod |
| Key message / benefits | Align to persona + product messaging preferences; translate restricted brief language (see brand-score hygiene) |
| Tone / do-nots | Per channel guidelines + any campaign constraints |
| Content check failures | If the user reports a low Brand score, ask for the Content check Needs review items (or paste them) |
Brand, Persona, and Product guidelines are selected in GenStudio Parameters—do not paste those lists into the prompt.
Brand-score hygiene
Brand score = guidelines passed ÷ guidelines tested on generated copy (Brand validation). Prompt text is not scored, but Create copies prompt phrases into headlines and body.
Before outputting a prompt, scan it (and any product phrases you were about to copy) against the substitution table in brand-guidelines.md. Common leaks:
- User said “security / secure / more secure / advanced security” → name capabilities and outcomes (compliance baselines, live kernel patching, Lightspeed-assisted detection and remediation; lower risk, less unplanned downtime). Never put
secure,more secure,security, orsecurity-focusedin the prompt. - Product value props in products.md may still say “proactive security” or “more secure.” Those lines are for Product Parameters, not for the prompt.
- Vague words without a direct object (
streamline,powerful,robust,leverage,utilize,seamless,frictionless) and Channel bans (flexible,flexibility,scalable,free,win,unlock) must not appear in the prompt either. - Do not write
thebefore a product name.
If the user brief used restricted words, rewrite them in the prompt and list the rewrite under Assumptions.
After a low Brand score
- Ask for Content check Needs review items if not provided.
- Rewrite the prompt: remove echoed restricted words; add only the failed items as targeted avoidances (Adobe: iterate by asking Create to avoid certain words/themes).
- Do not dump the full Brand guideline list into the prompt.
- Return the revised paste-ready prompt.
Prompt construction rules
Structured prompts give the LLM field-specific instructions. Use them for multi-section experiences (including multi-pod emails). Source: structured-prompts.png.
- Lead with a generic user prompt — intent, persona/audience, overall product/theme.
- Then add section-specific directives (
Pod1,Pod2, …). - Match template field names exactly — as the template defines them. Common names:
Pod/Group/Section/Module(e.g.Pod1),introduction,on-image text,headline,footer. Duplicate fields are numbered (on_image_text1,on_image_text2). Case-insensitive. - Separate name from directive with
:,;,-, or,:Pod1: Focus on…orPod1; Describe how to easily edit text and swap images. - Be specific — audience, purpose, features, benefits, action; include character/sentence caps per pod when useful. Use named capabilities and outcomes, not restricted category labels.
- One focus per pod — distinct product or benefit.
- Pass brand-score hygiene — no restricted words in the prompt (see above).
- Iterate — after Content check failures, add targeted word/theme avoidances only.
If the structure pattern is not followed, GenStudio treats the prompt as global and applies it to all sections, which usually reduces performance.
Single-product
Write a promotional email to motivate [persona] to [goal] using [Product]. Highlight [key capabilities / benefits]. Encourage [desired action].
For single-pod templates, you may still use Pod1: with a length constraint (see examples.md).
Multipod
Write a promotional multipod email to motivate [persona] to [goal] using [Product A] and [Product B].
Pod1: In [N characters / N sentences] [tone], focus on [Product A] and [specific capability / benefit].
Pod2: In [N sentences maximum] focus on [Product B / program] and [specific capability / benefit].
Output format
- Ready-to-paste GenStudio prompt in a fenced code block
- Parameters checklist: Brand; Persona; Product(s); single vs multipod; assets per pod
- Assumptions (only if inferred)—include any restricted-brief rewrites (e.g. “security” → named capabilities)
- Optional: remind of subject/preheader/headline/CTA character caps if the user will edit fields manually
Do not generate subject lines, headlines, or body copy unless asked.
Trademark & naming
Preserve ®/™ when the user supplies them (e.g. Red Hat® Enterprise Linux®). Do not invent marks.
References
- Channel limits & email style: channel-guidelines.md
- Brand voice, editorial, and restrictions: brand-guidelines.md
- Persona descriptions & messaging: personas.md
- Product descriptions & messaging: products.md
- Adobe GenStudio rules: reference.md
- Adobe structured-prompt source: structured-prompts.png
- Examples: examples.md
Local brand PDFs (gitignored): Red Hat Style and Brand/Red Hat_Channel Guidelines.pdf, Red Hat Style and Brand/Persona.pdf, Red Hat Style and Brand/Prompt Examples.pdf
Personas source: GenStudio Personas_WIP → personas.md
Products source: GenStudio Products (GenStudio Products.txt) → products.md