Offer and proof
Supply the product, offer, audience, objective and destination page. List approved facts, evidence and excluded claims. Use a fictional or authorized brief; do not upload customer lists or private campaign data without permission.
Growth marketers, brand teams, and agencies
Turn an approved offer and audience brief into coordinated campaign messaging, channel-specific drafts and production-ready creative instructions. Keep facts, visual requirements and review decisions connected as you move between text and media tasks.
Practical guidance
A reviewable campaign pack: one brief, a message matrix, channel drafts, image/video directions and a launch checklist. Publishing and performance measurement happen in your own marketing tools.
Supply the product, offer, audience, objective and destination page. List approved facts, evidence and excluded claims. Use a fictional or authorized brief; do not upload customer lists or private campaign data without permission.
Provide voice examples, prohibited phrasing, target channels, placement specifications and approved brand assets. Specify what must remain exact, including prices, names, logos and product details.
Choose an owner for copy and visuals, define one primary conversion action and establish how your own tools will measure it. Set a generation budget yourself; this guide does not configure spending or launch ads.
Practical guidance
Ask AI chat to organize the offer, audience problem, promise, evidence and CTA. Separate confirmed facts from creative hypotheses and unresolved questions.
Checkpoint: Every factual claim should trace to an approved input rather than a plausible invention.
Generate a small message matrix with one benefit and supporting proof per angle. Select an angle using relevance, clarity and credibility before requesting many variations.
Checkpoint: Do not treat an AI score as a measured prediction of conversion performance.
Draft the email, landing-page or ad copy against the actual placement requirements. Ask for counts and alternate phrasings, then verify the counts in the destination tool.
Checkpoint: Names, offers and CTA destinations must remain consistent across channels.
Turn approved copy into a composition brief or shot list, specifying subject, framing, palette and non-negotiable product details. Use a compatible image/video model only after checking its inputs and app controls.
Checkpoint: A storyboard is a plan, not a finished edited video; compositing, typography and assembly may require external tools.
Review the campaign pack with the responsible people, then publish through your existing tools. Record the actual results before asking AI to interpret what to test next.
Checkpoint: Keep observed metrics separate from hypotheses and change one test variable at a time when practical.
Technical details
Channel documentation + editorial practice
Store channel, field names, length limits and destination alongside each draft. Google Ads responsive search ads use distinct headline and description fields; their rules are not universal social-ad or email limits. Verify the live placement requirements before launch.
App capability boundary
Plan the handoff explicitly: approved copy becomes an image brief or shot list, then a compatible generation input. This guide does not establish automatic orchestration, ad-account connections, scheduling or publishing inside EZ Ai Assist.
Model selection guidance
Resolution, reference inputs, aspect ratio and clip length depend on the exact route and exposed controls. Words in a prompt express intent, not a supported API setting. Check the linked model guides rather than applying one provider's limits to another.
W3C guidance
W3C distinguishes informative, decorative and functional images. Draft alternatives based on an image's actual purpose in the final page; decorative images and linked controls need different treatment. Inspect the finished asset instead of generating alt text from the prompt alone.
Technical details
These are editorial starting points, not measured comparisons. Check the exact model guide and current app access before using a model; linked catalogs may also contain historical entries.
Evaluate a chat model on your approved brief and a small set of voice examples. Ask for explicit reasons behind each angle rather than a large pile of slogans.
Tradeoff: A fluent draft can still invent facts. Compare it with the proof points you supplied.
Explore Anthropic modelsUse an image model for mood, composition and visual direction once the message is settled. Consult the exact guide before assuming reference-image editing is supported.
Tradeoff: A convincing image may alter product geometry or branding; preserve exact logos and copy in your production process.
Explore FLUX modelsChoose a video workflow around the source material you actually have. Some routes require an input image or video; confirm those requirements before writing the shot brief.
Tradeoff: Do not equate one generated clip with a complete multi-shot campaign edit or guaranteed synchronized audio.
Explore Runway modelsPractical guidance
Replace every bracketed placeholder with authorized context. These are original templates, not generated-output demonstrations or generation-tested recipes. Copying does not submit anything or incur a generation charge.
Where to use it: Paste into AI Chat with approved offer and brand facts.
Build a campaign brief for [OFFER] aimed at [AUDIENCE], with [PRIMARY ACTION] as the objective. Approved facts and proof: [EVIDENCE]. Voice examples: [EXAMPLES]. Channels and constraints: [CHANNELS/LIMITS]. Return the core proposition, three messaging angles, supporting proof, CTA and missing inputs. Separate facts from creative hypotheses. Do not invent testimonials, performance statistics, discounts or product capabilities.
Check: Approve the proposition and proof before expanding it into channel copy.
Where to use it: Paste into AI Chat after choosing an angle.
Adapt [APPROVED ANGLE AND PROOF] into [CHANNELS AND REQUIRED FIELDS]. Follow [EXACT LIMITS] and [BRAND VOICE]. Preserve [OFFER, PRICE, PRODUCT NAMES AND DESTINATION]. Return a table with channel, draft, proposed character count and evidence used. Add one separate image brief and one shot-list concept. Label media briefs as production instructions, not completed assets. Flag missing specs rather than guessing them.
Check: Verify character counts and whether the selected image/video route can accept the proposed input.
Where to use it: Paste into AI Chat with final drafts and human asset-review notes.
Audit [CAMPAIGN PACK] against [APPROVED FACTS], [PLACEMENT RULES] and [BRAND GUIDELINES]. Asset review notes: [WHAT A HUMAN OBSERVED]. Check unsupported claims, offer consistency, CTA destinations, text limits and accessibility needs. Return issues by severity with the exact affected draft, evidence and revision. Do not assume unseen images are correct or claim the campaign will achieve a particular conversion rate.
Check: Resolve concrete issues, then obtain human approval and publish in your own channel tools.
Illustrative example
An editorial example of structure, not a generated result, customer case study or measured outcome.
Example: introduce a neighborhood repair workshop. Objective: completed workshop bookings. Approved facts: the supplied session date and location. Unverified claims such as fastest repairs or guaranteed savings are excluded.
Angle: learn a practical skill. Email: invitation and booking CTA. Landing page: agenda and confirmed logistics. Image brief: tools arranged on a workbench. Shot list: arrival, demonstration and takeaway; all illustrative, not generated footage.
Record each asset's owner, placement, version, destination and approval status. Keep the measurement field blank until your own analytics or advertising tool reports an observed result.
Technical details
Add a specific audience problem and concrete proof point. Ask for fewer distinct propositions, not more adjectives or generic urgency.
Return to the approved offer sheet and identify immutable fields. Revise only the conflicting draft instead of regenerating the entire campaign.
Reject or correct the asset. A stronger prompt is not proof of fidelity; final inspection and any necessary external editing remain part of production.
Provider maximums, picker estimates and actual customer charges are different things. Read the pricing explanation.
Common questions
No. These are planning and production-assistance steps. This guide does not establish ad-account integrations, scheduling, publishing or automatic campaign-budget control in the app.
No. The example is an editorial illustration of an output format. No campaign was launched, no media was generated and no performance uplift was measured for this guide.
That depends on the models, usage, revisions and any external tools. Provider references and picker estimates are not actual EZ Ai Assist customer charges. Read the pricing explanation and confirm costs before generating assets.
Evidence and review
The linked technical references were reviewed on . Documentation and app availability can change.
The brief, prompts and campaign example are original editorial suggestions. Official sources support the attributed channel/accessibility guidance, not successful app execution or campaign outcomes. Model/provider pages explain their own evidence boundaries. No ad account or generation API was used.
Opening EZ Ai Assist does not prefill these templates or automatically run the workflow. Choose the relevant mode and review your input before submitting it.
Continue exploring
Create outlines, drafts, summaries, scripts, visuals, and publishing assets from one organized AI workspace.
Generate platform-specific hooks, captions, post series, image prompts, video ideas, and audio concepts.
Build scripts, scene plans, shot lists, video prompts, fallback concepts, and generated video previews.