How to Advertise on ChatGPT: The Step-by-Step Setup Guide for 2026 (While the Platform Is Still New Enough to Matter)

The $50,000 minimum spend requirement that kept most advertisers locked out of ChatGPT’s ad platform? Gone as of May 5, 2026.

OpenAI quietly opened its self-serve Ads Manager to beta advertisers, and suddenly one of the most talked-about new ad channels in years is accessible to businesses with a normal budget. If you’ve been watching from the sidelines waiting for the barrier to drop, it just did.

This guide walks you through the complete setup flow — from creating your advertiser account to monitoring your first live campaign. No filler, no “exciting opportunities ahead” commentary. Just the documented steps, the decisions you’ll actually face, and the things worth knowing before you spend a dollar.

Fair warning: this platform is moving fast. Steps and UI details in this post reflect the beta Ads Manager as of early June 2026. Bookmark it, but verify against what you see in-platform.

What You’re Actually Getting Into (Before You Touch the Interface)

If you haven’t read up on how ChatGPT ads actually work, do that first. We’ve covered what ChatGPT ads are and how they work in detail — but the short version is this: ads appear within the ChatGPT conversation interface, surfaced contextually based on what users are asking about. It’s not keyword-triggered in the traditional search sense. It’s intent-adjacent.

That distinction matters before you set up your first campaign, because the mental model you bring from Google Ads or Meta won’t map cleanly onto this platform. The bidding, the targeting logic, and especially the creative requirements are different enough that building campaigns the same way you would elsewhere will get you mediocre results at best.

Also worth reading before you spend: if you’re still deciding whether this channel makes sense for your business at all, the ChatGPT Ads vs Google Ads comparison lays out an honest framework for making that call.

With that context set — here’s how to actually set up a campaign.

Step 1: Create and Verify Your Advertiser Account

Go to ads.openai.com and apply for Ads Manager access. As of the May 2026 beta opening, this is a self-serve application — no sales call required, no minimum spend commitment at intake.

You’ll need an existing OpenAI account to apply. If you don’t have one, create it first at openai.com. Once you’re in the Ads Manager application flow, you’ll be asked to provide:

• Business name and primary website URL
• Business type (you’ll select from a category list)
• Country of business registration
• Advertiser contact information — a business email, not a personal Gmail

After submission, OpenAI sends a verification email. Click through and confirm. In the current beta, approval for most standard advertisers has been coming through within one to three business days. Some categories — financial services, healthcare, legal — are flagging for additional review, which is consistent with how those verticals get treated on Google and Meta.

Once approved, you’ll land in the Ads Manager dashboard. It’s clean, minimal, and currently less cluttered than Google Ads UI (enjoy it while it lasts).

Step 2: Add Billing

Before you can create a live campaign, you need a payment method on file. Navigate to Settings → Billing in the left sidebar.

Current accepted payment methods: major credit/debit cards and some regional payment processors depending on your country. Wire transfer is available for accounts spending above a threshold that OpenAI has not publicly specified — assume it’s not relevant for your first campaign.

Set a billing threshold here. This is the amount at which OpenAI charges your card automatically. The default threshold is relatively low during beta — watch it and adjust upward once you’re confident in your spend rate, or you’ll trigger multiple small charges that some finance teams hate.

There’s no prepaid budget model in the current beta. You’re running on credit, which means budget controls at the campaign level matter more than they would on a prepay system. Set hard daily budgets and don’t rely on monthly caps as your safety net.

Step 3: Create Your Campaign — Objective, Dates, Budget, and Geography

Click Create Campaign. The first decision is your objective, and this is actually an important strategic choice, not a formality.

You have two options:

• Views/Reach — This runs on a CPM (cost per thousand impressions) model. Use this for brand awareness, product launches, or when you’re trying to establish presence on the platform before you have conversion data.
• Clicks — This runs on a CPC (cost per click) model. Use this when you have a specific destination you’re driving to and a way to measure what happens after the click.

Don’t default to CPC out of habit. If you’re testing a new channel with no performance history, CPM can give you useful impression and engagement data at a lower risk profile. If you know your landing page converts and you have tracking in place, go CPC.

Next, set your campaign dates. OpenAI requires a defined start and end date — there’s no perpetual “always-on” toggle in the current beta UI. Set an end date far enough out to gather meaningful data, but put a calendar reminder to extend it before it lapses. Campaigns that expire and go dark will lose the performance signal they’ve accumulated.

Set your daily budget. There’s no official minimum daily spend documented in the beta, but campaigns with sub-$20/day budgets are showing limited delivery. Budget at least $30–50/day if you want meaningful data in any reasonable timeframe.

Finally, select your target countries. Geographic targeting in the current beta is country-level only — no region, DMA, or city targeting yet. If your business is hyper-local, this is a real limitation. Factor it in before committing budget.

Step 4: Build Your Ad Groups — Context Hints and Bids

This is where ChatGPT advertising diverges most sharply from what you know.

Inside each campaign, you create ad groups. Each ad group gets its own context hints and bid. Think of ad groups here similarly to how you’d think about them in Google — a way to organize related targeting and creative into coherent buckets.

Context hints are the targeting mechanism. They are not keywords. They’re descriptive phrases or topics that tell OpenAI’s system which conversational contexts your ads are relevant for. Examples: “project management software for remote teams,” “B2B SaaS procurement,” “outdoor home renovation,” “tax planning for small business owners.”

The mistake most Google Ads practitioners make here: they treat context hints like exact match keywords and write them like search queries. Don’t. Write them as descriptions of the conversational context where your ad would be genuinely helpful. The more semantically clear you are about the context, the better the system can place your ads appropriately. For a deeper look at how this targeting layer works, this breakdown of ChatGPT Ads targeting and audience selection is worth your time.

For CPC campaigns, set your maximum CPC bid at the ad group level. There’s no automated smart bidding equivalent in the current beta — you’re setting manual bids, which is actually refreshing. Start conservatively (test at $1.50–$3.00 CPC for most B2B categories) and adjust based on delivery and click-through data after the first week. For CPM campaigns, you’ll set a max CPM instead.

Create multiple ad groups with different context hint sets rather than one ad group with a long list of hints. This gives you cleaner performance data by context theme and lets you adjust bids where you’re getting strong results versus weak ones.

Step 5: Upload Your Image Creative and Write the Ad

Each ad requires four components: a title, body copy, an image, and a destination URL.

Image specs (current beta):

• Recommended size: 1200 x 628px (standard display/social crop)
• Format: JPG or PNG
• Max file size: 5MB
• No text overlaid on the image — OpenAI’s review policy rejects image-text combinations. Your title and body handle the copy; the image carries the visual.

On the image itself: clean, brand-consistent visuals perform better than busy stock photography. This is a conversational interface. A loud banner ad aesthetic is jarring in context. Think closer to a LinkedIn carousel image than a Google Display Network banner.

Ad copy guidelines:

• Title: Up to 30 characters. Clear, direct, benefit-led. Not clever. Not punny. The user is in the middle of a task — meet them there.
• Body: Up to 90 characters. One clean value proposition. This isn’t the place for feature lists.
• URL: Your final destination URL. Deep links to specific landing pages, not homepages. The more relevant the destination, the better your post-click performance and the less likely you are to see early policy flags for misleading landing experiences.

OpenAI’s ad content policies are strict and currently less transparent than Google’s. Superlatives (“the best,” “the only,” “#1”), unverified claims, and anything that could read as medical, financial, or legal advice is getting flagged in review. Write factually and specifically. “CRM software built for sales teams over 20 people” will clear review faster than “The most powerful CRM in the market.”

The copy discipline required here is similar to what good Google Ads copy demands — if you want a framework for writing ad copy that’s both compliant and converts, the principles in our guide on writing high-converting Google Ads copy translate well to this format.

Step 6: Submit for Review and Know What to Expect

Once your ad is written and your creative is uploaded, hit Submit for Review.

Current review timelines in the beta: 24–72 hours for most ads. Complex categories (anything touching finance, health, legal, or political content) are running longer — 3–5 business days in some cases. Plan your campaign start dates with this buffer built in.

If your ad is rejected, OpenAI will email a rejection reason. The reasons are currently less granular than what Google provides — you’ll often get a category-level flag (“content policy violation: claims”) rather than the specific line that triggered it. Rewrite conservatively and resubmit. There’s no appeal process documented in the current beta.

Common rejection triggers we’ve observed so far:

• Comparative claims (“better than [competitor]”)
• Urgency language (“limited time,” “act now”)
• Destination URLs that don’t clearly match the ad’s offer
• Image file sizes over the limit
• Any text embedded in the image file

Build a clean version and a conservative backup version of each ad before you submit. If the primary gets flagged, you’re not scrambling from scratch.

Step 7: Monitor Performance — What the Dashboard Tells You (and What It Doesn’t)

Once your campaign is live, the Ads Manager dashboard reports on: impressions, clicks, CTR, spend, average CPC or CPM, and reach. For CPM campaigns, you also get a frequency metric.

What it doesn’t tell you yet: anything that happens after the click. There’s no native conversion tracking in the current beta. You are fully dependent on your own tracking — UTM parameters, Google Analytics 4, your CRM, whatever you use — to understand post-click performance.

This is the most important operational point in this entire guide. If you don’t have UTM parameters on your destination URLs before the campaign goes live, you will have no idea what the traffic did. Set up campaign-level UTMs at minimum: utm_source=chatgpt&utm_medium=cpc&utm_campaign=[your-campaign-name]. Then verify in GA4 that the sessions are tagging correctly before you let spend accumulate.

Check delivery against budget pacing in the first 48 hours. If you’re running a $50/day budget and the campaign has spent $4 by end of day one, you have a delivery problem — either your context hints are too narrow, your bid is too low for the inventory available, or your ad is stuck in a second review pass. Don’t wait a week to catch this.

For the first two weeks, focus on CTR and delivery before drawing conclusions about ROI. The system needs time to find its footing with new advertisers, and your context hints may need adjustment based on which ones are actually triggering impressions.

Is This Worth Your Time and Budget Right Now?

That’s the actual question, and the honest answer is: it depends on what you’re trying to learn.

If you’re in a category where ChatGPT’s user base is highly active — software, B2B services, financial planning, anything where people are actively asking AI for recommendations — this is a channel worth testing now while CPCs are still forming and before the platform gets crowded with advertisers who figured it out six months ago.

If you’re a hyper-local business, a business dependent on phone calls, or a business where the geography limitation (country-level only) is a dealbreaker, wait for the platform to mature before committing real budget.

The early-mover advantage on a new ad channel is real — we saw it with Google Ads in the early 2000s, with Facebook Ads pre-2015, and with LinkedIn Ads before B2B advertisers piled in. Getting competent on this platform while it’s new is a legitimate strategic move. Just go in with UTMs, realistic expectations, and a defined test budget you’re prepared to learn from rather than demand immediate returns on.

Frequently Asked Questions

How much does it cost to advertise on ChatGPT?

There’s no documented minimum spend in the self-serve beta. Practically speaking, campaigns need at least $30–50/day to get meaningful delivery data. CPCs in most categories are currently forming in the $1.50–$4.00 range, though these benchmarks are early and will shift as more advertisers enter the platform. For a deeper look at cost expectations, see our ChatGPT Ads cost guide for 2026.

Do I need a special account to access OpenAI Ads Manager?

You need an OpenAI account and to apply for Ads Manager beta access at ads.openai.com. As of May 2026, this is open to self-serve applicants — no agency relationship or minimum spend commitment required at intake. Some business categories are subject to additional review.

What’s the difference between context hints and keywords?

Context hints are topical/conversational descriptions that tell OpenAI’s system where your ads are relevant. They’re not match-type keywords — there’s no exact, phrase, or broad match logic. Write them as descriptions of the conversation your ideal customer would be having, not as search queries you’d bid on in Google.

Can I run ChatGPT ads at the same time as Google Ads?

Yes, and for most advertisers, you should. These are complementary channels, not competing ones. Google captures active search intent at a moment of decision. ChatGPT ads reach people in an exploratory or research mindset, often earlier in the buying process. Run both, track them separately, and don’t let one channel’s metrics bleed into the other’s evaluation.

How long does ad review take on ChatGPT?

Most ads in standard categories clear review in 24–72 hours in the current beta. Regulated categories (finance, health, legal) are running 3–5 business days. Build this buffer into your campaign launch planning.

What happens if my ChatGPT ad gets rejected?

You’ll receive an email with a rejection category. Rewrite the ad more conservatively and resubmit. There’s no formal appeal process in the current beta. Common triggers include superlative claims, urgency language, text in the creative image, and destination URL mismatches.

Is there conversion tracking in ChatGPT Ads Manager?

Not natively in the current beta. You must use UTM parameters and your own analytics stack (GA4, your CRM, etc.) to track post-click behavior. This is non-negotiable — running spend without post-click tracking is flying completely blind.