The short answer
AI can remove much of the mechanical work from a Demand Gen build, but it should not decide the strategy on its own. The useful pattern is to give it an approved brief, have it build the campaign in stages, inspect each layer, and enable only after a marketer has checked conversion goals, audience, inventory, creative, landing pages, and budget.
That staged approach matters because a Demand Gen campaign is not one object. It is a chain: budget → campaign → ad group → channel controls and targeting → assets → ads. A polished campaign shell can still be incomplete or unsafe to launch if one downstream layer is wrong.
Table of Contents
Demand Gen is not Search with images
Search captures existing intent. Demand Gen is designed to create and convert demand across visual surfaces such as YouTube, Discover, Gmail, Maps, and the Google Display Network. Google lets advertisers control the channels used by each ad group, while its systems combine and serve the eligible creative.
That changes the planning question. Do not begin with “create a Demand Gen campaign.” Begin with:
The business outcome: site visits, conversions, conversion value, or YouTube engagement.
The audience: who should see the message and what evidence supports that choice.
The creative idea: what will stop, explain, and persuade on the selected surfaces.
The measurement plan: which conversion actions are trustworthy enough to guide bidding.
The channel plan: broad inventory or a deliberate subset such as YouTube Shorts and Discover.
Before you ask AI to build anything
Prepare a compact campaign brief. If any of these inputs are missing, ask the AI to identify the gaps rather than inventing defaults.
1. Conversion goal and bidding
Choose the primary outcome first. Current Demand Gen bidding supports clicks, conversions, and conversion value; YouTube engagement is also available for appropriate campaigns. A target CPA, target ROAS, or target CPC is a constraint – not a substitute for a realistic budget and reliable historical data.
Google currently recommends a daily budget of at least 10 times target CPA for Demand Gen campaigns using target CPA. Treat that as a starting recommendation, then adjust for the account’s economics and expected conversion volume.
2. Budget and dates
Set a daily or campaign-total budget, start and end dates, and the currency context. Demand Gen cannot use a shared budget. If an AI assistant proposes an existing budget, verify that it is non-shared and belongs to the intended account.
3. Audience, location, and language
Demand Gen is audience-led. Define the audience you want to reach, but review its composition and size in Google Ads before launch. Also confirm locations, exclusions, language, and whether location matching should mean physical presence or presence plus interest.
4. Channel and creative fit
Channel controls sit at the ad-group level. You can use all channels, Google-owned-and-operated channels, or select individual channels. The creative must match that choice: a Shorts-only ad group needs credible vertical video; a carousel needs a coherent sequence; broad inventory needs enough image and video variation to avoid awkward crops and weak combinations.
5. Landing page and tracking
Confirm the final URL, mobile experience, consent behavior, conversion tag, and any tracking template or final URL suffix. Campaign creation does not validate the landing page or prove that conversion tracking fires correctly.
A six-step, review-first AI workflow
Step 1: Inspect the account before creating
Ask the AI to list current Demand Gen campaigns, relevant conversion actions, reusable audiences, and any existing creative assets. This prevents duplicate campaigns, weak audience assumptions, and accidental optimization toward the wrong action.
Prompt
In Google Ads account [CUSTOMER_ID], inspect the current Demand Gen setup.
List active Demand Gen campaigns, their goals and budgets, available audiences relevant to [OFFER], and the conversion actions that could support [OBJECTIVE].
Do not make changes. Flag anything I need to decide before campaign creation.Step 2: Turn the brief into a build plan
Have the AI translate the brief into a structured plan before it writes to Google Ads. The plan should show the campaign goal, bidding, budget, ad groups, channel selection, audience, locations, languages, creative formats, and required assets.
Prompt
Using this approved brief, draft a Demand Gen build plan.
Separate campaign-level settings from ad-group targeting and channel controls.
Map every creative asset to the placements it is meant to serve.
Call out missing inputs and do not create anything yet.
[PASTE BRIEF]Step 3: Create the campaign shell paused
Create only the budget and campaign-level settings first. Keep the campaign paused. Confirm the goal, bidding strategy, target, dates, device eligibility, geo matching, ad schedule, and tracking before proceeding.
Prompt
Create the approved Demand Gen campaign shell in account [CUSTOMER_ID] with the settings below. Keep it paused.
Return the new campaign ID, budget ID, goal, bidding setup, dates, device settings, geo target type, tracking settings, and anything Google rejected or changed.
[PASTE APPROVED CAMPAIGN SETTINGS]Step 4: Add ad groups with explicit inventory decisions
Create ad groups around real differences in audience, offer, or creative format – not merely to make the account look organized. Choose either a broad channel strategy or explicit channels. Confirm locations and languages at the ad-group layer.
Prompt
Create the approved Demand Gen ad groups under campaign [CAMPAIGN_ID].
Keep every ad group paused. For each one, use exactly the approved channels, audience resource, locations, exclusions, and language.
Return a comparison table of what was requested versus what was created.Step 5: Create assets and ads paused
Demand Gen supports multi-asset image ads, carousel ads, and video-responsive ads. Reuse standalone assets when the same logo, image, text, or YouTube video belongs in multiple ads; otherwise, inline asset creation may be simpler. Keep new ads paused while Google processes and reviews them.
For carousel ads, review the card sequence and destinations. For video ads, confirm the YouTube video IDs, aspect ratios, business name, headlines, descriptions, call to action, and final URL. Do not treat a successful API response as creative approval.
Prompt
Create the approved [MULTI-ASSET / CAROUSEL / VIDEO] Demand Gen ad in ad group [AD_GROUP_ID]. Keep it paused.
Use only the supplied creative and URLs. After creation, list the new ad ID, status, ad type, final URLs, and every attached creative asset so I can compare it with the brief.
[PASTE APPROVED CREATIVE]Step 6: Run a pre-launch readback
Do not rely on the prompts you sent. Read the created objects back from Google Ads and compare them with the brief.
Verify before enabling:
The campaign is paused and attached to the intended non-shared budget.
The optimization goal points to active, relevant conversion actions.
Every ad group has the intended audience, locations, language, and channel controls.
Every ad has the intended format, final URL, business name, copy, CTA, and asset references.
Creative previews and crops look acceptable in Google Ads across eligible placements.
Policies, brand safety settings, tracking, and landing-page behavior have been checked outside the build response.
Only then should a human approve enabling the campaign, ad groups, and ads.
Managing an existing Demand Gen campaign with AI
The same review-first pattern applies after launch: inspect current settings, propose a change, approve it, apply only the selected fields, and read the objects back.
Campaign-level changes
AI can help change the campaign name or status, budget assignment, goal and bidding targets, dates, devices, geo target type, schedules, and tracking. Schedule replacement deserves extra care: a replace operation can remove existing windows before adding the new set.
Ad-group changes
Channel strategy, individual channel selections, name, and status can be changed. Before narrowing inventory, compare performance and creative suitability; channel control is a strategic choice, not just a cleanup action.
Creative changes
Demand Gen creative should be replaced rather than edited in place. The safe sequence is to inspect the current ad, create a new ad with the approved creative, verify it, and then pause the old ad. This preserves a clearer review trail and avoids pretending the original ad was mutated.
Prompt
Inspect Demand Gen ad [AD_ID] in ad group [AD_GROUP_ID].
Show its current status, final URL, ad type, and creative asset references.
Then draft a replacement plan using the new creative below. Do not create or pause anything until I approve the complete replacement.
[PASTE NEW CREATIVE]Common mistakes
Letting AI choose a conversion action from its name alone. Verify the action’s setup, recent activity, and role in bidding.
Using broad channel coverage with only one creative shape. Supply placement-appropriate variants or deliberately narrow inventory.
Creating the campaign enabled. A campaign can start serving before every downstream setting has been reviewed.
Treating an audience as a substitute for an offer. Targeting cannot rescue weak creative or an irrelevant landing page.
Assuming creation validates tracking, policy eligibility, asset crops, or landing-page quality. Those remain separate checks.
Changing bidding and creative at the same time. Stagger material changes so performance shifts remain interpretable.
How HireOtto helps with this workflow
HireOtto turns the staged workflow above into actions inside the AI assistant you already use. It can inspect Demand Gen campaign, ad-group, and ad settings; create the paused campaign shell; add ad groups with audience, location, language, and channel controls; create reusable assets; and build paused image, carousel, or video ads. It can also update campaign settings, adjust ad-group channel controls, pause or enable entities, and replace creative by creating a new ad and pausing the old one.
The important review gates remain human. HireOtto does not decide whether the audience is strategically right, preview every crop, validate live conversion tracking, approve policy status, or determine whether a bid target is commercially sensible. Use it to make the build inspectable and faster – not to skip the brief, creative review, or launch approval.
Follow the step-by-step Demand Gen guide, check the Google Ads MCP tools reference for the current action surface, and connect HireOtto with the quickstart guide.
Frequently asked questions
Can AI create a complete Demand Gen campaign?
Yes, if the connected product supports write actions. Treat the build as a sequence – campaign, ad groups, targeting, assets, and ads – and inspect each layer. With HireOtto, new Demand Gen campaigns and ads default to paused for review.
Can I choose where Demand Gen ads appear?
Yes. Current Demand Gen channel controls can use all channels, Google-owned-and-operated channels, or individual channels such as YouTube in-stream, in-feed, Shorts, Discover, Gmail, Display, and Maps where supported. Channel controls are configured at the ad-group level.
Can AI edit an existing Demand Gen ad?
Status can be changed, but creative is best handled by replacement: create the approved replacement ad, verify it, and pause the old one.
Does creating the campaign prove conversion tracking works?
No. The campaign can reference a conversion goal, but live tag firing, consent behavior, attribution, and conversion quality need separate validation.
Should every Demand Gen campaign use all channels?
No. Broad inventory can increase reach and give Google more options, but a narrower selection may be appropriate when creative was designed for a specific surface or when the strategy calls for tighter control. Make the choice at the ad-group level and document why.
About Me
I’m Suyash – badminton junkie, ex‑GroupM ad‑ops grunt, first marketer at a B2B SaaS startup, and creator of Hiretto: Google Ads MCP Server.
My mission: less clicking, more thinking.
Let’s build leverage together.