Outbound Is Saturated: Why Decision-Makers Find Vendors Through AI Searches

Short answer: Cold outreach, LinkedIn messages, and Meta and Google ads bring back less because more and more vendors are targeting the same decision-makers at the same time. At the same time, buyers decide earlier and without sales: according to Gartner, 67% of B2B buyers want to buy without a sales rep, and according to 6sense, 94% of buying groups had already ranked their preferred vendors before the first contact. If you are then named in ChatGPT, Perplexity and Google AI, and have a website that backs up those answers, you receive inquiries instead of having to write cold messages.

The key points in brief

• Reply rates in outbound fall as volume rises (Hunter: 2.1% at 1,000+ recipients, 5.8% for small campaigns).
• The purchase decision is largely made before the first conversation, based on the buyer's own research, increasingly with AI.
• AI searches and AI Overviews shift clicks: according to Ahrefs, the click-through rate for position 1 drops by around 58% when an AI Overview appears.
• Your website is therefore no longer just a business card, but the surface where research, validation and inquiry take place.
• Outbound is not dead. It is a poor substitute for visibility where decision-makers search today.

What "saturated" means here

A channel is saturated when more vendors compete for the same attention than the target group can absorb. The price per contact rises, the response per contact falls, and every additional message worsens the average for everyone. This can be shown with individual figures, even though the data has gaps. We say openly where.

This guide covers five channels:

1. LinkedIn outreach: connection requests, messages, InMails, often with automation tools.
2. Cold email: sequences with personalization fields, increasingly with AI-generated copy.
3. Meta ads: reach and lead forms, in B2B often targeted by job title.
4. Google Ads: search ads on vendor and problem terms.
5. Cold calling: the oldest channel, with the same mechanics.

The thesis of this guide has three parts. First: these channels are getting more expensive and quieter because too many vendors use them at the same time. Second: decision-makers sidestep them and research on their own, more and more often with AI assistants. Third: this makes the website the most important sales surface, because that is where it is decided whether you land on the shortlist.

One caveat up front: many figures come from international studies, often from tool vendors or from the software sector. We flag that. Several series are missing for the German-speaking region, but the direction is consistent across the sources.

Who is pushing into your decision-makers' inbox right now?

A message to a managing director, purchasing manager or department head rarely comes from someone in your industry. Often it comes from software vendors, agencies or sales service providers who filter through databases such as LinkedIn Sales Navigator or Apollo and reach out by sequence.

That is an observation, not a measured statistic. There are indirect indications that the assumption holds, though:

• In an analysis of LinkedIn ad accounts reported by the portal seo-day.de, 97% of customers were B2B SaaS companies (around 700,000 US dollars in ad spend, May 2025 to May 2026). The sample is small and comes from a single agency, but it shows who dominates the paid B2B channels.
• The outreach studies from Hunter, Belkins and Expandi come from tool vendors. Their data basis is millions of automated messages.

Your message therefore competes not only against your competitors, but against everyone who wants to sell something to the decision-maker over the course of the week. Gartner reports that 73% of buyers avoid vendors who send them irrelevant outreach messages (Gartner, June 2025). Bad outbound is not just ineffective, it can actively do harm.

AI makes this worse: personalized messages are produced in minutes, which lowers the cost per message and raises the volume. There are no reliable figures on the growth of AI outreach volume. The next section instead shows what happens to the rates when a lot is sent.

What the numbers say about outbound channels

The following table summarizes the most reliable values. "Reliability" means: A is a primary source we retrieved ourselves, B is a benchmark from a tool vendor with a vested interest, whose sample is its own customers.

Metric	Value	Source	Year	Reliability
Average cold email reply rate	4.1% (95.9% go unanswered)	Hunter, State of Cold Email 2025	2024/25	B
Reply rate small campaigns (up to 50) vs. large (1,000+)	5.8% vs. 2.1%	Hunter	2024/25	B
Reply rate relative to all emails sent	0.45%	Belkins, Cold Email Response Rates	2025	B
Acceptance rate LinkedIn connection requests	28.5%	Expandi, LinkedIn Outreach Benchmarks 2026	05/2025 to 04/2026	B
Response rate to connection requests	from 3.5% to 2.2% in 12 months (relative minus 37%)	Expandi	05/2025 to 04/2026	B
Reply rate to LinkedIn messages	10.4%	Expandi	05/2025 to 04/2026	B
Reply rate C-level / founder on LinkedIn	7.0% / 6.4%	Belkins, LinkedIn Outreach Study	2025	B

A few points on reading this table, so you do not draw the wrong conclusions.

The figures seem to contradict each other. Hunter reports 4.1%, Belkins 0.45%. The difference lies in the calculation: since its new study, Belkins divides by all emails sent, not by opened or delivered emails. Instantly shows that studies land between roughly 5 and 9% depending on the measurement method. So never compare a figure from study A with one from study B, only values within the same study.

The clean trend value is LinkedIn. According to Expandi, the response rate to connection requests fell from 3.5% to 2.2% within twelve months. That is the only reliable year-long trend we have.

Volume eats the rate. At Hunter, small, targeted campaigns reach 5.8%, large ones with 1,000 or more recipients only 2.1%. That is exactly the saturation pattern: anyone who sends more to compensate for the decline lowers their rate further.

Not everything is bad. A 10.4% reply rate on LinkedIn messages shows that well-tailored outreach works, and C-level replies at around 7%, no less often than others. Outbound is not dead. It only reaches those who happen to be open to your topic right now, and that is a small part of the market.

Valuable messages are rare. According to Hunter, only 24% of surveyed decision-makers get a cold email that is genuinely useful to them at least once a week, and 20% never get a relevant one (survey of 217 decision-makers, 2025).

And the ad costs?

There is no reliable series for cost-per-click figures in the German-speaking region. As an order of magnitude, the seo-day.de analysis (predominantly B2B SaaS accounts, itself created with AI support) cites 11.12 US dollars per click for LinkedIn ads and 12.48 US dollars for Google non-branded search. For Meta ads in Germany, the sources contradict each other, so no figure is given. In the calculator further below, enter your own cost-per-click figures instead. That says more than any average.

How decision-makers really research today

Now the other side: where do the people who ignore your message get their information? Four studies point in the same direction.

They want to buy without sales. In the Gartner survey (646 B2B buyers, surveyed August to September 2025, published March 2026), 67% prefer a buying experience without a sales rep. The year before it was 61% (Gartner, June 2025, 632 buyers). 70% want a fully digital self-service purchase.

They use AI for vendor search. In the same Gartner series, 45% of buyers said they used generative AI in their last purchase, mainly for vendor and product research (Gartner, May 2026). Forrester cites even higher values: 94% of business buyers used AI in the buying process, up from 89% the year before (Forrester, Buyers' Journey Survey 2025). Forrester does not disclose the sample and region in that post, so take the 94% as a directional statement. The software market is furthest along: G2 finds among more than 1,000 software buyers that 50% start their research in an AI chatbot instead of Google, and 47% prefer ChatGPT for it (G2, August 2025). This explicitly applies only to software, not to mechanical engineering or the trades.

They have their favorites before you notice. 6sense surveyed more than 4,000 buyers in North America, EMEA and Asia-Pacific (Buyer Experience Report 2025). The result: buyers complete around 60% of their journey before contact with a vendor (the ratio has shifted from 70/30 to 60/40), 94% of buying groups had already ranked their preferred vendors before first contact, and the favorite wins the deal in around 80% of cases. The press release elsewhere cites 77%, hence "around 80%" with a caveat. The same study reports that 94% of buyers used large language models in the buying process.

Half of Germans already search with AI. Bitkom reports for Germany (1,156 respondents aged 16 and older, by telephone, September to October 2025) that 50% use AI chats for internet search at least occasionally, 7% predominantly and 5% exclusively. Among 16 to 29-year-olds it is two thirds. This is a population value, not a B2B value, but tomorrow's decision-makers sit in exactly these age groups.

They do not trust AI blindly, and that helps you. 69% of buyers still want to validate AI answers with a sales rep or vendor (Gartner, May 2026). On average they use seven information sources per purchase. And according to Forrester, decisions typically involve 13 internal participants and 9 external influencers (State of Business Buying 2026, January 2026). In other words: even if an assistant names you, a whole committee then looks at your website, references and reviews.

What follows from this

The classic outbound idea is: I find the decision-maker and approach them. The studies show: the decision-maker has already formed their circle of possible vendors by the time your message arrives. You are being sorted, not discovered. If you are missing in the first round, you have to write against the other vendors' favorite status, and that is the most expensive route.

Why clicks collapse even when you rank

Ahrefs analyzed 300,000 keywords (December 2025 against a baseline of December 2023, desktop, global). When an AI Overview appears for a search, the click-through rate for position 1 is 0.016 instead of a projected 0.037, around 58% less. For informational searches without an AI Overview, the position 1 click-through rate halved between 2023 and 2025 from 0.076 to 0.039 (Ahrefs, update on AI Overviews, February 2026).

Pew Research sees a similar pattern (900 US adults): with an AI Overview, 8% click on a result, without one 15% (Pew, 2025, reported by 9to5Google). Google disputes the methodology. Neither study comes from Germany.

For you, this means: visibility and clicks are no longer the same thing. If an AI assistant names you and does not link to your page, you have still left an impression that later leads to a direct search or inquiry. How you can measure such mentions is covered in the guide Measuring AI Visibility. We explain the difference between classic SEO and GEO in SEO vs. GEO.

The website as the most important sales surface

Research, shortlist and validation happen before the conversation, on pages that you control or that report about you. Your website is the only one you can shape entirely yourself. Instead of sending more messages, you strengthen the surface decision-makers come to anyway.

What AI systems need from your page

• No special trick needed. In its documentation on AI features, Google writes that nothing beyond normal SEO basics is required for AI Overviews and AI Mode: the page must be indexed and eligible for snippets, run cleanly from a technical standpoint, offer important content as text, be well linked internally, and structured data must match the visible text (Google Search Central, AI features and your website). So there is no secret "GEO switch".
• Backed-up statements are cited more often. The KDD study on generative search engines (Aggarwal et al., 2024) found that methods such as source citations, quotes and statistics can increase visibility in generated answers by up to 40%, depending on the field. It was tested with research data, not with German company websites. Take it as a directional statement, not a guarantee.
• Freshness helps to a limited degree. An Ahrefs analysis of 17 million citations shows that AI assistants cite newer pages on average than classic search does, most strongly ChatGPT. An artificially changed date without a substantive revision does not help (Ahrefs, Do AI assistants prefer to cite fresh content).
• llms.txt is a small building block, not a switch. In June 2025, Google said that no AI system currently uses llms.txt (John Mueller, via Search Engine Roundtable). Treat it as a quick-to-implement building block among several, but do not expect miracles and never use it as a substitute for good content and clean structure.

Checklist: Is your website ready for decision-makers who research beforehand?

Check	Why it matters	Typical shortcoming
Clear statement in the first sentences of every service page	Assistants pull short, self-contained answers	Marketing slogans instead of statements
Concrete numbers, references, industry focus	Buyers validate AI answers (69% according to Gartner)	References without context or result
One page per question instead of a catch-all page	Whoever asks a question gets the matching page named	Everything on one service page
Comparison and decision content	Shortlists come from comparison questions	Only self-promotion, no assessment
Industry pages with real technical terms	Buying groups search for their own field	One generic page for everyone
FAQ with real buyer questions	Assistants adopt question-answer pairs	Questions nobody ever asked
Visible authors, date, sources	Trust and verifiability	Anonymous texts without a date
Fast, technically clean page	Basic requirement for crawling and use	Heavy media that drive up load time
Contact route without hurdles	67% want to move forward without sales, but need a self-service entry point	Only a "get in touch" form

PageSpeed and AI Findability.

The 7 industries at a glance

The mechanism is the same everywhere, the specifics are not. Each industry has a companion guide with more depth; here is the short overview. Where we found no reliable figure, it says so.

IT & Software

This is where the shift is furthest along. G2 reports that 50% of B2B software buyers begin their research in an AI chatbot, and AI chats are the most important source for shortlists (G2, 2025). At the same time, the industry is the largest sender of outbound, which explains the flood of messages for decision-makers. Separate entry points for technology, business unit and purchasing, security pages and comparison content are important. More on this in the companion guide Outbound Is Saturated: IT & Software, matching the industry page IT, Software & SaaS.

Industry & Mechanical Engineering

Long cycles and technical buyers: according to Contentful (900+ decision-makers in manufacturing, industry and tech, EMEA, 2025), 84% consider self-service tools critical when choosing a vendor. Friction points are missing price transparency (39%), cumbersome reordering (35%) and slow vendor response (24%). There is no current DACH study on online supplier research in mechanical engineering. The requirements for the website are still clear: data sheets, certificates, application examples, inquiries with product reference. More on this in the companion guide Outbound Is Saturated: Industry & Mechanical Engineering, matching the industry page Industry & Engineering.

Real Estate

The available data mainly concerns people searching for housing. In a Scout24 survey of 500 people (March 2026), 74% believe personal contacts are decisive in the apartment search, and 69% want to know early which apartments are becoming available. For project developers and portfolio holders as a B2B target group, there is no current source, and we say so openly. More on this in the companion guide Outbound Is Saturated: Real Estate, matching the industry page Real Estate.

Energy & Sustainability

The EY/BDEW municipal utilities study 2024 (100 utilities) shows what occupies decision-makers: 87% name digitalization, heat planning and skilled-labor recruitment as their most important topics, and 79% consider BEW funding insufficient. Anyone who wants to reach municipalities and utilities runs into committees and tender processes in which outbound messages land poorly. Substantive specialist pages, funding overviews and proof of reference projects are more likely to land. The study gives no figures on the role of the website in vendor search. More on this in the companion guide Outbound Is Saturated: Energy & Sustainability, matching the industry page Energy.

Mobility & Aviation

The market is small, international and under pressure. The VDA reports almost 120,000 lost jobs in the German automotive industry since 2019 (currently around 716,000 employees) and a 24% decline for suppliers between August 2019 and August 2025 (VDA, October 2025). Suppliers are therefore looking for new markets, such as aviation, and buyers check suppliers heavily through certificates and proof. In a small target group, broadly scattered outbound is especially inefficient: every relevant decision-maker already knows the senders. There are no reliable figures on vendor search in aviation. More on this in the companion guide Outbound Is Saturated: Mobility & Aviation, matching the industry page Aviation.

Trades & Commercial Businesses

Here the core question is often twofold: finding customers and finding skilled workers. According to KOFA Kompakt 07/2026 (IW Köln, 2025 data), 209,444 open positions in the trades faced 139,239 unemployed skilled workers, a gap of 97,482, about every second open position. 59 of 115 trades occupations are shortage occupations. There is no current source on how commercial customers research trades businesses. More on this in the companion guide Outbound Is Saturated: Trades & Commercial Businesses, matching the industry page Trades.

Consulting & Law Firms

Consulting is a trust business in which recommendations and first conversations count. For outbound that is a difficult environment: in the Belkins analysis, response rates in finance and legal are slightly above average (around 0.56 to 0.60% versus 0.45%), but the level stays low. There are no industry-specific studies on AI-based vendor search for consultancies and law firms. What applies is the general finding: whoever is recommended gets checked before the conversation, and that happens on the website and in AI answers. More on this in the companion guide Outbound Is Saturated: Consulting & Law Firms, matching the industry page Consulting & Agencies.

Cost per lead: comparing channels properly

The sensible comparison is: what does a qualified inquiry cost me in each channel, and how long does it take to materialize?

For that there is the Cost-per-Lead Calculator. You choose your industry and enter your own assumptions for Google Ads, outbound, organic search and AI visibility. The calculator shows cost per inquiry and per customer over 12, 24 or 36 months.

Two things to keep in mind when comparing:

1. Time structure. Ads and outbound work immediately but end with the payment. SEO and AI visibility take months and remain as an asset. If you only compare the first three months, paid channels always come out ahead.
2. The AI factor is uncertain. For a single website, Ahrefs reports that visitors from AI searches converted much better (around 23 times) at a share of 0.5% of visitors and 12.1% of sign-ups. The author himself warns against generalizing. The calculator therefore works with a conservative multiplier by default, which you can change.

What you can concretely do in 90 days

A sensible approach in four steps:

1. Test the current state (weeks 1 to 2). Ask 15 to 20 real customer questions in ChatGPT, Perplexity and Google AI Mode, for example "Which vendors for [service] in [region/industry] are there?" and "What should I look for in [decision]?". Note who is named, who is cited and what the sources have in common. That is your baseline.
2. Close gaps in the answer structure (weeks 3 to 6). For every question where you are missing, check whether you have a page that answers it directly. If not, create one: with a short answer at the start, numbers with sources, a comparison, an FAQ.
3. Make proof visible (weeks 5 to 9). References with context, reviews, awards, team, certificates. After the AI answer, validation runs through these elements.
4. Rebuild outbound instead of switching it off (in parallel). Less volume, more relevance: small lists, reference to a specific event or problem, a link to a real specialist page instead of a home page. The Hunter figures (5.8% for small campaigns versus 2.1% for large ones) speak for this.

You can measure this through direct inquiries with the question "How did you find us?" in the form, through referral sources such as chatgpt.com in your analytics, and through repeated prompt tests. We explain the method in the guide Show Up in ChatGPT.

When outbound still makes sense

An honest assessment is part of this. There are situations in which outbound still does good work:

• New categories. If nobody is searching for your solution, no AI can recommend you either. Then you have to create demand actively.
• Very small, known target groups. If your market has 200 relevant companies, prepared personal outreach is more efficient than any visibility strategy. The principle only holds with real personalization, though, not with sequences.
• Prior relationship. If you have the contact through a trade show, a referral or earlier work, the message is no longer a cold contact.

If you need inquiries in three months, ads are faster. That is legitimate, as long as you build substance in parallel.

Conversely, visibility in AI answers is not self-propelling. The data on citation frequency and impact on B2B deals is young, mostly from the US, and partly from vendors who sell products themselves. The direction is well documented, the order of magnitude for your industry is open.

How we implement this

Webwavers is a premium agency for branding, web design and web development with locations in Günzburg and Munich, and is one of seven Webflow Premium Partners in Germany. For visibility in AI answers, we work with a website structure that answers questions directly, with industry pages, comparison content and a technical foundation that supports crawling and citability. We advise neutrally on which platform fits (Webflow, WordPress, TYPO3 or headless). More on our approach is on the page GEO and AI Visibility.

Frequently asked questions

Does cold outreach still work in B2B at all?

Yes, but with diminishing effect at high volume. According to Hunter, small cold email campaigns of up to 50 recipients reach a 5.8% reply rate, large ones from 1,000 recipients only 2.1%. On LinkedIn, the response rate to connection requests fell from 3.5% to 2.2% within twelve months (Expandi). Well-prepared, small campaigns continue to work, mass sending keeps getting worse.

Why don't decision-makers reply to LinkedIn messages and emails anymore?

Many get too many similar messages, mostly unrelated to their business. According to Gartner, 73% of buyers avoid vendors who send irrelevant outreach emails (2025). According to Hunter, only 24% of surveyed decision-makers get a cold email with value at least weekly. On top of that, they prefer to research on their own: 67% prefer buying without a sales rep (Gartner, 2026).

Do decision-makers really use ChatGPT and Perplexity for vendor search?

Increasingly. According to Gartner, 45% of B2B buyers used generative AI in their last purchase, mainly for vendor and product research (May 2026). G2 reports that 50% of software buyers start in an AI chatbot (2025). For Germany, Bitkom shows that 50% of internet users use AI chats for search (November 2025), though as a population average, not just in B2B.

What matters more, SEO or AI visibility?

They build on each other. Google writes that AI Overviews and AI Mode need no special optimization, but a good SEO foundation plus clear, text-based content (Google Search Central). Visibility in AI answers rests in large part on pages that also rank in the classic way. The difference lies in the structure: answers, evidence, comparisons. Details in the guide SEO vs. GEO.

How do I measure whether AI assistants name me?

With fixed test prompts in ChatGPT, Perplexity and Google AI Mode, with referral data from your analytics and with the question in the inquiry form about how someone found you. Specialized tools track mentions automatically. We describe the method in the post Measuring AI Visibility.

Should I stop my ads if I want to become visible in AI answers?

No. Visibility takes months, ads deliver immediately. It makes sense to keep ads as a bridge and build content in parallel that later brings inquiries without a cost per click. Use the Cost-per-Lead Calculator to check, with your own values, from what point the build-up pays off.

Does this also apply to the trades or small businesses?

The pattern yes, the specifics no. For trades businesses, local search, reviews and recommendation play a big role, and the second front is skilled workers (KOFA: 97,482 missing skilled workers in the trades in 2025). We lack a reliable source on the search behavior of commercial customers in the trades. Details in the companion guide Trades & Commercial Businesses.

Conclusion

Outbound is saturated because too many vendors, often from entirely different industries, want to reach the same decision-makers with the same tools. Rates fall as soon as volume rises, and buyers avoid senders who are not relevant. At the same time, vendor search is shifting to the buyers themselves: without sales, with Google, with review portals and more and more often with AI assistants. Whoever is named there and delivers answers, evidence and an easy next step on their website gets inquiries. The first step is a test: ask the AI yourself whom it names for your topic, and see whether you are among them.

Support with AI visibility

If you want to know where your website stands for vendor questions in AI assistants, get in touch via the contact page. For orientation beforehand, the service page Growth gives you an overview.

Sources

1. Gartner: 67 % der B2B-Käufer bevorzugen ein Kauferlebnis ohne Vertrieb (09.03.2026) (67% of B2B buyers prefer a rep-free buying experience)
2. Gartner: 61 % der B2B-Käufer bevorzugen ein Kauferlebnis ohne Vertrieb (25.06.2025) (61% of B2B buyers prefer a rep-free buying experience)
3. Gartner: 69 % der B2B-Käufer validieren KI-Erkenntnisse beim Vertrieb (20.05.2026) (69% of B2B buyers validate AI insights with sales)
4. 6sense 2025 B2B Buyer Experience Report, Pressemitteilung (12.11.2025) (press release)
5. Forrester: B2B Buyers Make Zero-Click Buying Number One
6. Forrester: The State of Business Buying 2026 (21.01.2026)
7. G2: AI search surging for B2B buyers (2025)
8. Bitkom: Internet-Suche im Wandel, die Hälfte nutzt KI-Chats (20.11.2025) (Internet search in transition, half use AI chats)
9. Hunter.io: The State of Cold Email 2025
10. Belkins: Cold Email Response Rates (2025)
11. Belkins: LinkedIn Outreach Study (2025)
12. Expandi: LinkedIn Outreach Benchmarks 2026
13. Instantly: Cold Email Reply Rate Benchmarks
14. seo-day.de: LinkedIn Ads CPC Benchmarks (Bericht nach Search Engine Land) (report based on Search Engine Land)
15. Ahrefs: AI Overviews Reduce Clicks (Update, 02/2026)
16. Ahrefs: Besucher aus KI-Suche und Conversion (Visitors from AI search and conversion)
17. Ahrefs: Do AI assistants prefer to cite fresh content?
18. 9to5Google: Pew-Studie zu Klicks bei Google AI Overviews (23.07.2025) (Pew study on clicks in Google AI Overviews)
19. Google Search Central: AI features and your website
20. Aggarwal et al.: GEO, Generative Engine Optimization (KDD 2024)
21. Search Engine Roundtable: Google zu llms.txt (17.06.2025) (Google on llms.txt)
22. Contentful: 2025 B2B Buyer Benchmark Report
23. Scout24: Umfrage Wohnungssuche (März 2026) (Apartment search survey)
24. EY/BDEW Stadtwerkestudie 2024 (Pressemitteilung) (EY/BDEW municipal utilities study 2024, press release)
25. VDA: Stellenabbau in der Automobilindustrie, Verkehrsrundschau (22.10.2025) (VDA: job cuts in the automotive industry, Verkehrsrundschau)
26. IW Köln: KOFA Kompakt 07/2026, Handwerk (KOFA Kompakt 07/2026, trades)