.svg){kind=logo}
Buyers are researching companies inside ChatGPT now, not just Google. OpenAI opened ads to businesses of all sizes in 2026, and most B2B marketing teams don't yet have a plan for it, let alone someone running it properly.
We're Lever, a PPC agency already running Google, Microsoft and LinkedIn for a variety of SaaS and B2B scaleups, and ChatGPT Ads sits inside the same PPC strategy, with the same CRM-led measurement, and the same accountability for CAC, not clicks.
ChatGPT reached 900 million weekly active users in early 2026, per OpenAI's own disclosure reported by TechCrunch, which is a distribution scale no other AI assistant has matched, and one most B2B marketing budgets aren't yet allocated against.
That scale is exactly why OpenAI opened ads in the first place: keeping ChatGPT's Free and Go tiers fast and reliable takes real infrastructure spend, and advertising is how that gets funded. For advertisers, it means a genuinely new placement, and one that is visible the moment someone actively works through a decision, not after they've already formed an opinion and gone to Google to confirm it.
The catch is that almost nobody running Google or Microsoft Ads well is running ChatGPT Ads well yet, because the mechanics are different enough that copy-pasting a search campaign into it doesn't work. That gap is where an early, disciplined account gets disproportionate value relative to the spend required to test it.
Lever's team has run paid acquisition in-house at PayPal and Zettle before moving agency-side. That background shaped how we work: a four-stage process we call The Lever Method, applied to every account we run.
CAC reduction within the first two quarters for clients who move from click-based optimisation to ICP-aligned, SQL-based optimisation, without reducing spend.
Efficiency gains come from better targeting and measurement, not smaller budgets. That takes an agency fluent in both Google's technical requirements and the SaaS market itself, before you scale. Google's own documentation on offline conversion imports sets out how uploading enhanced lead data back into Google improves reporting and optimisation, provided the account feeds it clean CRM data.
Not every account should be on ChatGPT Ads on day one. We check three things before recommending it: whether your buyer's research process plausibly includes conversational AI (B2B SaaS and technical buyers, disproportionately yes), whether your landing pages and conversion tracking can support a context-based targeting model rather than exact-match keywords, and whether existing Google/Microsoft accounts are efficient enough that a new channel's learning-phase spend is a reasonable trade.
ChatGPT Ads still isn't built around exact-match keywords, but targeting is no longer just conversation context either. We now build campaigns across four layers:
1. Context hints: short, descriptive statements about your product or use case that guide delivery, not keywords to bid on
2. Custom audiences: matching your own customer or prospect lists for inclusion, exclusion or bid adjustment
3. Geo targeting: by country or specific location
4. Platform targeting: across ChatGPT's iOS app, Android app, or web.
We architect ad groups and context hints around the problem a buyer is actively describing to ChatGPT, not the search term they'd have typed into Google for the same need, then layer geo, platform and audience targeting on top the same way we would on any other channel.
We handle the whole setup directly with OpenAI, from registering the account, to setting the budget and uploading the ad creative. When it comes to how we tell the system to spend the budget, we set it up to aim for actual paying customers, not just clicks or views, because getting cheap clicks that never turn into business isn't the goal. We also put proper tracking in place from day one, so we can see exactly which ads are bringing in real leads. Most accounts are fully set up and running within about a week, the same as we'd expect on Google or Microsoft ads.
Where CRM data is connected, we optimise budget allocation against qualified pipeline, rather than treating ChatGPT Ads' own click and conversion numbers as the finish line.
Our clients see clear performance data, from impressions and clicks, to spend and conversions, so pipeline-level tracking still makes the channel accountable.
We're already running ChatGPT Ads for UK clients, including a SaaS client in the office real estate sector, where competition for ad space is fierce and clicks on Google can cost well over £100. The early results for that client are very encouraging:
on ChatGPT, vs £36 on Google Ads
on ChatGPT, vs £600 on Google
OpenAI doesn't let us see the actual conversation someone has with ChatGPT, so we lose some of the detail we'd normally get further down the funnel on other channels. But the early numbers are promising enough that it's a channel worth testing properly.
On ChatGPT, a click currently costs somewhere between $3 and $5, but there's also an option to pay just for how many people see the ad, rather than clicks, if the goal is getting your name in front of people rather than getting them to act straight away. One more thing worth knowing: these ads only show up for people using ChatGPT's free version, not people paying for a subscription.
ChatGPT Ads works best today as a test-and-learn channel layered onto an existing, CRM-connected account, not as a first channel for a company with no paid media history yet. If you're not already running Google or Microsoft ads with working conversion tracking, that's the more defensible place to start.
ChatGPT Ads is added as an additional channel at +£500/month, the same add-on rate as LinkedIn, since account structure and reporting integrate into the existing setup rather than requiring a separate engagement.
ChatGPT Ads is still in beta, so targeting, measurement and reporting can change while you're spending, and we don't think a client should carry that risk with a budget that has no other channel behind it. It also works best as one part of the funnel rather than the whole of it. That's why we only add it to an existing account, with a test budget carved out of your wider media plan, and read the results against everything else that's running.
Different ARR, different sales motions, different markets. What doesn't change is how we measure their success: qualified pipeline and CAC, not clicks.
Trial signups were growing, but the team couldn't tell which ones were worth chasing. We built a performance strategy around trial activation and low-cost acquisition rather than signup volume.
A white-label payment gateway based in Riga had invested in paid media for years without being able to optimise it. A previous agency had left the campaigns poorly managed. We audited the account, implemented enhanced conversion tracking to fix attribution and bidding, and shifted the campaigns toward lead quality rather than volume.
Sprintlaw's Google Ads account ran entirely on automated campaign types, with no strategy behind it, spending budget on Google's terms rather than the business's. We rebuilt the campaign around buyer intent over six weeks. Sprintlaw has since doubled monthly ad investment, and paid has overtaken every other channel for new business revenue.
ChatGPT Ads are sponsored placements that appear below ChatGPT's response to a conversation, clearly labelled as sponsored. Unlike Google Ads, targeting isn't built on exact-match keywords. OpenAI's system instead matches ads using conversation context, your landing page content, advertiser-supplied context hints, and broader usage signals, plus standard geo, platform and custom-audience targeting layered on top, through a relevance-weighted, second-price auction. The response itself is never influenced by advertising.
Yes. We're running live campaigns for UK clients already, with billing, targeting and reporting all working the same way as they would for any other market. Availability continues to widen as OpenAI opens Ads Manager to more businesses, but for UK B2B and SaaS advertisers specifically, it's live now, not a “coming soon.”
Only logged-in adult users on ChatGPT's Free and Go tiers. Plus, Pro, Business and Enterprise subscribers never see ads, and OpenAI excludes under-18 accounts based on account age and predictive signals. That's worth knowing before you budget: you're reaching ChatGPT's free-tier audience specifically, not its full user base.
OpenAI recommends starting bids of $3–5 per click for CPC and oCPC campaigns. Even oCPC, which optimises delivery toward conversions, still bills per valid click rather than per conversion, with CPM bidding available separately for reach-focused, impression-based campaigns. We typically recommend £1,000–£3,000/month in media budget for a first test, enough volume at those bid levels to draw a real conclusion within 8–12 weeks.
Yes, and we'd recommend it over running it standalone. The same CRM connection and CAC-based reporting we use for Google and Microsoft applies here, so a ChatGPT Ads pipeline shows up next to every other channel's, rather than living in its own dashboard.
It's worth testing if your buyers are the kind of technical, research-heavy audience likely to use ChatGPT during vendor research, which describes most SaaS ICPs, and if you already have working conversion tracking elsewhere. It's not yet the channel to lead with if you have no paid media history at all; get Google or Microsoft foundations right first.
No, and OpenAI doesn't provide it. Advertisers receive aggregated performance metrics (impressions, clicks, spend, CTR, average CPC/CPM, conversions), not access to individual conversations, chat history or personal details, which is why conversion event feedback is essential for ChatGPT ads to work. Conversion tracking runs through OpenAI's OAIQ pixel and Conversions API, or standard UTM-tagged links into your own analytics.
OpenAI is explicit that its context hints “are not exact-match keywords and do not guarantee delivery in specific conversations.” Practically, that means campaign structure is built around the problem a buyer is describing and the content of your landing page, rather than a keyword list, which is closer to contextual or topic targeting than search intent-matching.
Michéal has spent 10+ years leading B2B paid acquisition for SaaS, FinTech and marketplace companies, including in-house roles at PayPal and Zettle before founding Lever. He has managed £50M+ in ad spend across Google, Microsoft, LinkedIn and Meta, and writes regularly on SaaS PPC benchmarks, CAC measurement and multi-channel budget allocation.
Lever Digital is a Google Partner and a 2026 UK Paid Media Awards finalist.
You'll speak to the person who would run the account. Rolling monthly, no fixed term, no markup on media.
