Create a Campaign from Scratch
Execution
A controlled workflow for building campaigns from an initial brief.
Overview
This workflow takes you from a plain-language brief to a fully specified, ready-to-launch campaign — without letting the AI touch your live account until you say so. You describe the objective, audience, budget and constraints in natural language; OpenAds researches the opportunity, drafts the structure, keywords, targeting and creative, and assembles everything into a single proposal you can review in one place. The outcome is a complete campaign build that reflects your intent, backed by the reasoning behind every choice.
Because analysis, proposal, approval and execution are kept separate, nothing is created on the ad platform until you approve it. You are never handed a black-box campaign after the fact. Instead you get a structured draft — campaign settings, ad groups, keywords, negatives, bids and ads — that you can question, edit and approve piece by piece or all at once. The result is faster setup with full control retained.
When to use it
Reach for this workflow whenever you are starting a new campaign rather than adjusting an existing one. It suits both first launches and adding a new line of activity to a mature account.
Launching a new product, service or seasonal push from a written or verbal brief.
Entering a new market, geography or language where you have no existing structure.
Standing up a test campaign quickly to validate demand before committing larger budgets.
Rebuilding a legacy campaign with a cleaner structure and fresh keyword research.
Handing a rough idea to the AI and wanting a rigorous, reviewable first draft in minutes.
Example prompts
"Build a Search campaign for our new project-management app targeting startup founders in Bengaluru and Mumbai, ₹1,500 daily budget."
"Create a campaign to promote our winter sale across India with a total budget of ₹90,000 over three weeks."
"Draft a lead-generation campaign for our B2B accounting service, English and Hindi, ₹800 per day, and propose keywords."
"Set up a Performance Max campaign for our online store with a ₹2,000 daily budget and suggest asset group themes."
"Plan a test campaign for our fitness app in Delhi NCR only, ₹500 a day, focused on high-intent search terms."
"Create a brand-defence Search campaign on our own brand terms with a modest ₹300 daily budget."
"Build a campaign targeting competitor comparison searches in Pune, ₹1,200 daily, and include a strong negative keyword list."
What OpenAds prepares vs what you approve
OpenAds does the analysis — researching keywords, estimating volumes and competition, and checking account fit — and turns it into a proposal: a complete draft of campaign settings, structure, targeting, keywords, negatives, bids and ad copy, each with a short rationale. What it does not do is act on that draft. The approval step is entirely yours: you can accept, edit or reject any element, and nothing reaches the platform without your explicit sign-off. Only after approval does execution happen, when OpenAds creates the campaign through the MCP connection — usually paused, so you keep one last checkpoint before spend begins. This separation means the AI accelerates the work while you retain authority over every live change.
What to look at
SignalWhy it mattersWatch for Estimated reach & volumeConfirms the audience is large enough to hit the objectiveVolumes too thin to spend the budget Keyword relevancePoor matches waste budget and depress qualityBroad or ambiguous terms slipping in Negative keyword coverageBlocks irrelevant, costly clicks from day oneObvious exclusions missing from the list Daily budget vs bid strategyDetermines whether the campaign can actually competeBudget too low for the chosen bidding model Geo & language targetingEnsures spend lands where the audience isOver-broad regions or wrong language settings Ad copy accuracyMisstated offers erode trust and risk disapprovalClaims, prices or terms that do not match the brief Early spend & conversionsShows the live campaign is behaving as intendedZero impressions, runaway spend or no conversions
Common pitfalls & best practices
Vague briefs produce vague proposals. Spell out audience, budget and constraints up front so the draft needs less rework.
Don't skip the negative keyword review. A strong exclusion list at launch prevents weeks of wasted spend on irrelevant queries.
Match the budget to the bidding strategy. Smart bidding needs enough daily budget and conversion data to learn; starve it and performance stalls.
Launch paused, then verify. Use the paused state to sanity-check targeting and copy in the platform before enabling.
Start narrow, expand later. A focused first build is easier to read and optimise than a sprawling one; add ad groups once early signals are clear.
The workflow, step by step
Share the brief. Describe the objective, product, target audience, budget, geography and any hard constraints in plain language; the more specific the brief, the sharper the proposal.
Let OpenAds research. The AI pulls keyword ideas, search volumes, competition and auction signals, and checks how the theme fits your existing account so nothing overlaps or cannibalises.
Review the proposed structure. OpenAds drafts the campaign type, ad groups and their themes, then explains why each grouping was chosen so you can reshape it before anything is built.
Confirm targeting and budget. Check the proposed geo, language, devices, bidding strategy and daily budget, and adjust any figure that does not match your intent.
Approve keywords and negatives. Scan the suggested keywords, match types and negative list; prune irrelevant terms and add exclusions before they ever cost you a click.
Review the ad creative. Read the drafted headlines, descriptions and extensions for accuracy, tone and compliance, and edit any line that misstates the offer.
Approve for execution. Once satisfied, approve the proposal; only then does OpenAds create the campaign on the platform, typically in a paused state for a final look.
Enable and monitor. Turn the campaign live when ready and watch early signals over the first days to confirm it is spending and converting as expected.
Run this workflow through your AI assistant.