Google Ads in 2026: PMax, AI Max or Search, which to use for B2B lead generation

A company that manufactures laser cutting machines opens Google Ads, accepts the platform’s recommendations, launches a Performance Max campaign and after a month sees 60 leads at a very low cost per contact. The salesperson calls back ten of them: three are students, two are looking for a job, one sells SEO services, and four are customers who wanted the support number. In Google’s report the campaign looks like it’s doing brilliantly.

The short answer to the question in the title is this. For B2B lead generation, the foundation is still a Search campaign with keywords you choose: it’s the format that reaches people searching for what you sell. AI Max is an extension of that same Search campaign, useful when searches are few, fragmented or phrased in ways you don’t anticipate, but it only delivers if Google receives the signal from the CRM about which leads are worth something. Performance Max comes third: it’s fine for broadening coverage when you have qualified conversions imported, brand exclusions active and customer lists excluded. Without these three elements, in B2B PMax tends to buy searches for your own brand, remarketing and forms filled in by people who will never buy.

How the two automated formats work is explained in our articles on AI Max and Performance Max. Here we decide which to use, in what order and with which signals.

Search, AI Max and Performance Max: what’s the difference if you’re after leads?

Search and AI Max work only on Google’s search network; Performance Max distributes ads across the search network, YouTube, Display, Discover, Gmail and Maps from a single campaign (Google documentation). The real difference for lead generation is how much control you keep over the query that triggers the ad.

With traditional Search, you decide the keywords and match types: if you sell “CNC press brakes”, you appear only on that family of searches. The price is maintaining keyword and negative lists.

AI Max is switched on inside an existing Search campaign. The keywords stay, but Google widens the net with search term matching, can rewrite parts of the ads and choose the landing page. Negatives, brand inclusions and exclusions and locations of interest remain available (Google’s guide to AI Max). It performs well when searches come in many variants with low volumes each, which is common in technical B2B.

Performance Max starts from a conversion goal and a package of assets. You don’t choose keywords: you provide search themes and audience signals, then the algorithm decides where to spend. It’s the format in which you have the least control over individual impressions.

Which format to choose: the decision matrix

If you have few conversions or conversions of uncertain quality, start with Search. If you have qualified conversions imported from the CRM and want more volume at the same quality, try AI Max on the Search campaign that’s already working. If you have robust signals, decent creative assets and Search is already saturated, add Performance Max.

Situation	Traditional Search	AI Max on Search	Performance Max
Technical niche, few searches a month	Yes, the essential foundation	Yes, with negatives and brand controls, to find variants	Usually not: too little data to optimise
B2B service with medium volumes (hundreds of searches)	Yes	Yes, after 4–6 weeks of clean data	Possible as an addition, with exclusions
Well-known brand, many brand searches	Yes, with a separate brand campaign	Yes, on non-brand campaigns	Only with brand exclusion, otherwise it takes the credit for searches on your name
Conversions are only “form submitted”	Yes	With caution	No
Qualified conversions imported from the CRM	Yes	Yes, performs better	Yes, it’s the prerequisite
No good-quality video or image assets	Makes no difference	Makes no difference	Hold off until you have them
Long sales cycle (months)	Yes	Yes, if you also import the intermediate stages	Only with intermediate conversions (e.g. qualified lead)
Goal: new customers, not people who already know you	Yes	Yes	Yes, excluding customer lists

The three formats aren’t mutually exclusive: in the B2B accounts we manage, there’s usually a brand campaign, one or more non-brand Search campaigns (some with AI Max) and, when the data allows, a PMax campaign with a defined scope.

The decision tree in six questions

To choose, just answer six questions in order. Every “no” stops you at a more controlled format.

1. Do you know which leads become sales opportunities? If not, sort out tracking and qualification in the CRM, and in the meantime work with Search on exact and phrase match keywords.
2. Do searches for your product exceed a few hundred a month? If not, stay on targeted Search and try AI Max on one ad group at a time, to discover variants.
3. Do you import at least one qualified event (qualified lead, opportunity, contract) into Google Ads? If not, no AI Max at scale and no PMax until the import is active.
4. Does non-brand Search already cover the main searches, without losing impression share to budget or rank? If not, work on that first. If so, the next step is AI Max on the Search campaigns that convert.
5. Do you have creative assets (at least images and a short video) and a clear offer for people who don’t know you? If not, PMax waits.
6. Can you switch on brand exclusions, exclude customer lists and optimise for qualified leads? If so, launch a PMax campaign with a separate budget and compare it with Search for 6–8 weeks. If not, stay on Search plus AI Max.

Three B2B use cases compared

The choice depends on the demand in the market: with few technical searches you start with Search, with medium volumes and many variants with AI Max, and with a brand people already search for you separate brand and non-brand before switching on PMax. Three typical examples.

Technical niche with few searches

Take a company that manufactures filtration systems for the electroplating industry. Specific searches number a few dozen a month, written in technical language (“electroplating wastewater treatment plant”). Here Search with phrase and exact match keywords is the only sensible foundation: an algorithm with five conversions a month has nothing to learn from. AI Max serves as a discovery tool on a single ad group, to turn good new queries into keywords. PMax, in such a small market, ends up on Display and YouTube looking for volumes that don’t exist.

B2B service with medium volumes

A predictive maintenance company gets a few hundred searches a month across many phrasings: “motor vibration monitoring”, “thermographic inspection of electrical panels”. This is AI Max territory: there are too many variants to cover by hand and Google, if it knows which leads turn into quotes, looks for similar queries. PMax comes later, once non-brand Search is covered and the qualified event imported from the CRM occurs a few dozen times a month.

A brand already known in the sector

A well-known manufacturer of hydraulic components receives thousands of searches for its brand. A PMax campaign without brand exclusion takes those searches, which you would have won anyway, and presents them as its own conversions: the report improves and revenue stays flat. The correct structure is a separate brand campaign, non-brand campaigns by category (with AI Max where volumes allow) and, if appropriate, a PMax campaign with the brand and customer lists excluded.

Pros and cons of each format

Each format trades control for volume: more automation brings more coverage, but requires better signals and more frequent checks on where the budget ends up.

	Pros for B2B	Cons for B2B
Search	Control over the query, clear reporting, works even with little data, easy to explain to the sales team	Requires constant maintenance, misses search variants you don’t anticipate, doesn’t scale much
AI Max	Finds relevant new queries, keeps keywords and negatives, reporting with a dedicated match type	With weak conversions it broadens towards generic searches; copy and landing pages changed by Google need reviewing
Performance Max	Coverage across all Google channels, reaches people who aren’t searching for you yet, good for volume	Tends to spend on brand and remarketing, limited control over individual placements, requires robust assets and data

For AI Max, Google reports on average 7% more conversions at a similar CPA or ROAS when all features are used together, compared with search term matching alone (source: Google). It’s an internal, global average: in our B2B accounts we verify it campaign by campaign on qualified leads.

What changed in 2026 for PMax and AI Max (as of October 2026)

In 2026 Performance Max has more control and reporting tools than in the past, and the automatic migration to AI Max has only partly started: Dynamic Search Ads have been postponed to February 2027.

Performance Max: more transparency

In 2026 Google announced new controls and reports for PMax (Google announcement):

• first-party data list exclusion: remove current customers and contacts already in negotiation from the campaign;
• budget reporting within the campaign, with a forecast of month-end spend;
• age and gender reporting on the audience reached;
• placement reporting segmented by network.

The tools that arrived earlier remain: channel and search term reports, brand exclusions, campaign-level negative keywords. The latter, according to Google’s documentation, apply only to Search and Shopping inventory: if a junk lead comes from a video, the negative won’t stop it.

AI Max: where the automatic migration stands

The original plan was for automatic migration to AI Max from September 2026 for DSA, automatically created assets and campaign-level broad match. On 11 June 2026 Google updated the announcement: the migration and retirement of DSA will begin in February 2027, while automatically created assets and campaign-level broad match were converted from September 2026 (source: Google).

If in September you saw AI Max appear on campaigns you hadn’t touched, that’s why: check which features are active and read the terms with the “AI Max” match type. If you still use DSA, you have a few months to migrate manually and choose the settings yourself.

Which signals does automation need to bring in good leads?

The algorithm optimises for whatever you tell it to count. If you count form submissions, it will bring you more forms; if you count qualified leads or opportunities, it will learn to look for people similar to those who buy. Which levers to leave to automation and which to keep under control on Google, Meta and LinkedIn is covered in AI in advertising campaigns.

The signals that count, in order of weight:

• Qualified conversions. One action for a lead qualified by the sales team, one for an opportunity, possibly one for a contract. Google itself, in its guidance on lead generation with Performance Max, suggests goals such as “Qualified lead” and “Converted lead”.
• Conversion values. A lead from a 200-employee company in your target market is worth more than one from a sole trader outside the sector. Different values, even estimated by band, open up value-based bidding strategies.
• Customer lists. Customer Match lists serve two purposes: as a signal (“look for people like these”) and, from 2026 in PMax too, as an exclusion (“don’t spend on them”).
• Brand exclusions. In PMax and AI Max they stop the automated format from taking over searches for your brand.
• Negative keywords. In Search and AI Max they remain the first line of defence. In PMax they work only on the Search and Shopping part.

Checklist before switching on AI Max or PMax

1. The CRM records the lead’s origin with the GCLID or with the data for enhanced conversions.
2. There’s a “qualified lead” conversion action set as primary, with “form submitted” as secondary.
3. The offline conversion import runs at least once a week and you’ve checked it against some real leads.
4. You have a list of current customers uploaded to Google Ads, ready to use as an exclusion.
5. Brand exclusions are configured (your own brand and, if needed, your distributors’ brands).
6. The shared negative keyword list includes jobs, training, free, used, spare parts if you don’t sell them, and terms in languages you don’t serve.
7. The form has at least one filtering question (sector, role, quantity) and an anti-spam system.
8. You’ve decided in advance how long you’ll allow for learning (6–8 weeks) and which metric you’ll judge on: cost per qualified lead.

How to import CRM conversions into Google Ads

Importing offline conversions means sending back to Google Ads, days or weeks after the click, which leads became qualified, opportunities or customers, so that automation learns from the commercial outcome.

Enhanced conversions for leads use data provided by the user, such as a hashed email address, to link the lead to the click even when the GCLID is missing. Since April 2026 Google Ads has accepted data arriving from tags, Data Manager and the API together, and since June 2026 web and leads have been a single setting (Google announcement). Google still recommends including the GCLID when you have it: it’s mandatory if you don’t collect user data with the tag (Google’s guide to upgrading offline imports). For setup, Google points to Data Manager as the simplest route.

In practice: the form saves the GCLID in a hidden field and passes the email and phone number to the tag; the CRM stores this data on the lead together with a status that the salesperson updates; at every relevant status change, the corresponding conversion is sent to Google Ads. The complete journey, from form to revenue, is in From lead to contract: connecting CRM and revenue; which statuses to measure is covered in our guide to B2B lead management metrics.

What changes for lead generation, CRM and performance

Choosing the right format shifts the reference metric from cost per contact to cost per qualified lead and, when the data is there, to cost per opportunity. It’s the metric the sales director recognises.

• Contacts fall, opportunities don’t. When a campaign switches from form submissions to qualified leads, cost per contact rises in the first few weeks: the algorithm stops buying easy forms.
• Marketing and sales share lead statuses. If the salesperson doesn’t update the CRM, the import stops and automation goes back to learning from forms. A simple rule (qualify within 48 hours) and a shared lead scoring criterion are worth more than any campaign setting.
• Formats are compared on CRM data, meaning qualified leads and opportunities, not on the conversions reported by Google, which in PMax include brand and remarketing.

Mistakes to avoid

The mistakes we find most often in B2B accounts have a common root: giving automation a goal that differs from the company’s.

• PMax living off brand and remarketing. Without brand exclusion and customer lists excluded, many PMax conversions come from people who were already looking for you. If the top queries in the search terms report are your brand, that PMax campaign isn’t acquiring anyone.
• Optimising for “form submitted”. A submitted form is easy to get even from people who will never buy, and the algorithm learns to find as many as possible at the lowest cost.
• Junk leads from forms. Forms that are too short, lead form assets with no qualifying questions, no reCAPTCHA. A mandatory “sector” field filters out a lot.
• Switching on AI Max without reading the search terms. In the first few weeks the report should be read every week, adding negatives for off-target queries (training, jobs, consumer products with the same name).
• Judging PMax after ten days. Google recommends at least one to two weeks, up to six for complex setups or low conversion volumes. In B2B you need 6–8 weeks before comparing it with Search.
• Overlapping formats without a plan. When the query matches an exact match keyword, the Search campaign takes priority over Performance Max; otherwise PMax and Search overlap, so decide in advance what each one should do.

If you work in manufacturing, many of these dynamics are intertwined with trade fairs, distributors and long purchasing cycles: we discuss this in our guide to industrial marketing.

Want to know which format your account needs?

A Google Ads account review starts from three questions: what the campaigns are optimising for today, how much brand and remarketing weigh in the reported conversions, and whether the CRM sends qualified leads back to Google. From there you can tell whether it makes sense to keep the current structure, switch on AI Max on some campaigns or launch (or switch off) a Performance Max campaign.

It’s the work we do in Google Ads campaign management and, for the automated side, in Performance Max campaigns. If you’d like us to look at your account, book a call or write to us via the contact form: we’ll tell you what we would change, and also whether, for your market, it’s better to stay on Search.

Frequently asked questions

Does Performance Max work for B2B lead generation?

It can work if it optimises for qualified leads imported from the CRM, with the brand excluded and customer lists excluded. If it optimises for form submissions, in B2B it usually brings in lots of low-quality contacts and conversions inflated by brand and remarketing.

Is AI Max a different campaign type from Search?

No, AI Max is a set of features switched on inside an existing Search campaign. Keywords, negatives and ad groups remain; Google adds query expansion, text customisation and final URL expansion.

Have Dynamic Search Ads already been converted to AI Max?

No. Google has postponed the automatic migration and retirement of DSA to February 2027. In September 2026, campaigns with automatically created assets and campaign-level broad match were converted instead.

Do negative keywords work in Performance Max?

Yes, at campaign and account level, but only on Search and Shopping inventory. They don’t block ads on YouTube, Display, Gmail or Discover.

How many conversions do you need to use AI Max or PMax in B2B?

Google doesn’t set a minimum for lead generation. In our experience you need at least a few dozen qualified conversions a month on the event you’re optimising for; below that threshold, Search with controlled keywords performs better.

Should I switch off Search if I switch on Performance Max?

No. In B2B, Search remains the foundation for high-intent searches and for the brand; PMax is added with a defined scope and judged on the additional qualified leads it brings.