# Build a brand guide so every AI output stays on brand > Every ads, social, blog, and email playbook you run afterwards writes in the client's voice, because they all read the same brand note. It is built once per client from the site and the live ad accounts, and the offering section adapts to whatever the client sells, from products to services to software. Source: https://www.liftedwork.com/automations/automate-brand-guide Published: 2026-04-20 Updated: 2026-09-15 ## What this does If you run AI content for more than one client, you've hit this wall: every task briefs the brand from scratch. The copy sounds slightly off. The image's blue isn't quite the client's blue. A headline uses language the client wouldn't. Across ten clients and dozens of content runs per client per month, that drift shows up in the work. This playbook fixes that with one per-client setup. It reverse-engineers the client's brand into a single note covering voice, visual system, photography direction, offering details, and ad creative style. Two halves of the same note: the Brand Overview (five voice adjectives, positioning) keeps AI-generated copy (ads, social posts, emails, blog articles) sounding like the client. The Image Generation Prompt Modifier, a 50 to 75 word paragraph, prepends to every image and keeps visuals consistent. A catalog of real reference images harvested from the client's own website gives image-generating playbooks real brand URLs to pass to Gemini as style anchors. The offering-details section adapts to whatever the client sells, so this works whether they're a physical product, a service, software, or a physical location. Run it once per client during onboarding. Every content run after that reads this note automatically. You stop re-briefing brand on every task. ## Best for Any client you generate content for, regardless of what they sell. Usually kicked off automatically by the [new-client onboarding playbook](https://www.liftedwork.com/automations/automate-new-client-onboarding), but you can also run it standalone at any point (onboarding an existing client, refreshing after a rebrand, etc.). Content-generating playbooks read this note on every run: image playbooks for visual consistency, copy playbooks for voice. ## How the chain works **Template 1: Research Brand.** Budget-bounded data collection. 3 page browses on the client's site (homepage, About, one product or services page) with `detail: full` to pick up the image list in one call, and one call each to Meta Ads and Google Ads for top-spending ads if those integrations are connected. Evaluates images and picks the 8-12 best candidates, dropping logos, icons, and decorative junk. Descriptions are inferred from URL, alt text, and page context (the AI cannot see the actual pixels), with a "subject unclear" fallback when signal is low. Writes everything into one "Brand Research Findings" deliverable. When done, books Template 2. **Template 2: Draft Brand Guide.** Reads the findings deliverable and the client profile (for hex values and fonts). Checks for an existing Brand Guide note and updates it in place, or creates a new one. This is the synthesis step, and its job is to make the **Image Generation Prompt Modifier** paragraph as precise as possible. That single paragraph is prepended to every AI image generated for the client, so its 50-75 words shape visuals more than any other field in the note. Creates or updates a client note titled exactly "Brand Guide". When done, books Template 3. **Template 3: Review Brand Guide.** You review the note, tweak the Prompt Modifier paragraph, prune any reference images the AI kept that don't belong, fix any wrong descriptions, and approve. The approved note is live for every image playbook from that point on. ## Requirements - Templates 1 and 2 need **Autocomplete** and **Run on complete** enabled. - The client's website URL must be available before Template 1 runs: paste it into the `[CLIENT WEBSITE URL HERE]` line in the Lifty Prompt before scheduling, or set the `website` field on the client profile. Pasting it into the prompt directly is more reliable. - The client note must stay titled exactly **Brand Guide**. Downstream playbooks match on that title; renaming it breaks the chain. - If the client profile already has brand colors and fonts extracted (automatic when the website is set), Template 2 uses them as the starting point for the Visual System section. - Optional but useful: Meta Ads and Google Ads integrations. Template 1 pulls the client's top-spending ads for the Ad Creative Style section if either is connected, and skips silently if not. ## How to use it 1. Create a task from "Research Brand (1/3)" for your client. Before saving, paste the client's homepage URL into the `[CLIENT WEBSITE URL HERE]` line in the Lifty Prompt (or leave it blank to fall back to the website on the client profile; pasting it is more reliable). 2. On its due date, it autocompletes: gathers site research and ad history into a "Brand Research Findings" deliverable, then books the Draft task for 5 minutes later. To run it earlier, mark it complete yourself, set the due date sooner, or open the task and click Run Prompt. 3. The Draft task autocompletes off the findings deliverable and writes the Brand Guide client note. 4. The review task lands in your queue. Open the "Brand Guide" client note on the client's Notes tab, tighten the Prompt Modifier paragraph, prune any reference images that don't belong, and mark complete. 5. The note is now live. When you run the [Ad Creatives playbook](https://www.liftedwork.com/automations/automate-client-social-media-posts) or the [Blog to Social playbook](https://www.liftedwork.com/automations/automate-blog-to-social-repurposing), the AI reads this note automatically and uses the Prompt Modifier and reference images on every generation. ## Tips - **Treat this like an onboarding asset.** One Brand Guide per client, set up once at the start of the relationship. Every content playbook you run for that client from then on reads this note automatically. Across your full roster, that's dozens of content runs per client per month that no longer need manual brand briefing. Refresh only when the client rebrands. - **The Prompt Modifier is the whole game.** The AI's draft is a starting point. Every image the client ever generates starts with this paragraph, so rewriting it in your own voice during review is the single biggest quality lever. Budget 5-10 minutes of real attention here. - **Add reference images the AI missed.** If the client has a photo library, a Dropbox, or product shots they use in email campaigns, paste those URLs into the Reference Images section by hand. The AI is limited to what it finds on the public website. - **Adapt the Offering Details section to the client.** The AI fills this section based on what the business actually sells, but it's worth adding specifics during review. Depending on the client, that might include any of: - **Physical products**: packaging materials, finishes (matte, gloss, translucent, embossed), label placement, product colorways. - **Services**: uniforms, vehicle or fleet styling, signage, before/after signatures, the feel of a typical work site. - **Software**: app icon, dashboard conventions, key UI patterns, color of the product UI vs the marketing site. - **Physical locations**: storefront look, interior photography style, signage, typical props and surfaces on-site. - **Ground the Ad Creative Style section in real spend.** If the client has Meta Ads or Google Ads connected, the AI pulls their top-spending ads over the last 90 days to anchor this section in what has actually worked, not guesses. - **Re-run after a rebrand.** If colors, fonts, or creative direction change, run this playbook again. It updates the existing note rather than creating a duplicate. - **Kicked off by the [New Client Onboarding playbook](https://www.liftedwork.com/automations/automate-new-client-onboarding).** That playbook creates a task from this one automatically during onboarding, so the Brand Guide is built before the kickoff call. - **Pairs with these downstream playbooks:** - [Ad Creatives](https://www.liftedwork.com/automations/automate-client-social-media-posts) (hard dependency): reads the Brand Guide for the Prompt Modifier, Image Dos and Don'ts, and Reference Images on every run. - [Blog to Social](https://www.liftedwork.com/automations/automate-blog-to-social-repurposing) (hard dependency on the image step, same as above; copy steps are unaffected). - [SEO Blog Post pipeline](https://www.liftedwork.com/automations/automate-seo-content-pipeline) (soft dependency): reads the Brand Overview section for voice. Writes the article either way. ## Credits The Brand DNA approach this playbook is built around, and the Image Generation Prompt Modifier concept specifically, come from a public prompt walkthrough shared by Will Sartorius and the Adcrate team: [Nano Banana 2 Prompt Document](https://adcrate.notion.site/Nano-Banana-2-Prompt-Document-31ed5575919580d4987ef6056b56122a) and [their YouTube walkthrough](https://www.youtube.com/watch?v=ibzfKRRSHhU). We have adapted it to fit our tools and workflow. ## Frequently asked questions ### What should an AI-ready client brand guide include? Include positioning, audience, voice rules, words to use or avoid, visual colors and fonts, photography direction, offering details, approved examples, and a short image prompt modifier. Separate confirmed brand rules from inferences that still need client approval. ### How often should an agency update a client's Brand Guide? Review it during onboarding, after a rebrand or material offer change, and whenever repeated outputs reveal a missing rule. A stable brand does not need monthly rewriting, but the note should change when its source material changes. ### Can the playbook understand reference images from a website crawl? The research step identifies image URLs and uses surrounding text, filenames, and alt text to describe likely subjects, but it does not inspect every pixel. A person should remove poor references, correct weak descriptions, and add approved assets from the client's library. ### Do Meta Ads and Google Ads have to be connected? No. The playbook can build a guide from the website and client profile. Connected ad accounts add evidence about the creative the client is already running, which can make the ad style section more specific.