AI in Sales

Executive Summary

57 percent of German companies now use AI, nearly three times as many as two years ago (Bitkom, 2026). Globally, however, only 37 percent of the surveyed companies report a positive impact on operating profit, and only 6 percent achieve a noticeable one (McKinsey, 2026). In sales, this gap is particularly relevant, as this is where the promises are the loudest: autonomous AI agents, automated lead generation, and a sales pipeline at the push of a button.

This e-book summarizes what independent studies on AI in sales actually demonstrate. It highlights why projects fail, where AI delivers measurable results, and the five questions decision-makers should address before making their next investment.

Here’s the key finding up front: Success is rarely determined by the model itself, but almost always by the data set, the integration
into sales processes, and what teams do with the time they save.

High usage, little impact: The gap in numbers

Adoption of AI is now widespread in Germany. 91 percent of companies consider AI to be the most important technology of the future, and its use in sales has nearly tripled within a year: Of the companies that use AI, 14 percent deploy it in sales, up from
5 percent the previous year (Bitkom, 2026). At the same time, not a single one of the companies surveyed says it is fully tapping into AI’s potential. According to their own assessments, 59 percent are not utilizing it at all, and one-third of users openly admit that they use AI primarily out of fear of falling behind.

The global picture is similar. In the latest McKinsey survey of 1,719 respondents, 80 percent report that AI has improved their personal productivity. The percentage of respondents who see a positive impact on EBIT, however, stands at 37 percent – practically the same as the previous year – even though more companies are scaling up their use of AI (McKinsey, 2026). Noteworthy for sales managers: Revenue gains from AI are most often attributed to marketing and sales. The potential, therefore, lies precisely where most misinvestments occur.

Personal productivity is not the same as business results. Just because you write emails faster doesn’t automatically mean you’ll win more contracts.

Why AI Fails in Sales

The independent sources are remarkably unanimous on one point: The problem is rarely
the technology. The MIT report cites a learning gap as the main cause. Many AI tools
do not learn from feedback, do not adapt to specific workflows, and therefore remain
siloed solutions. McKinsey arrives at the same conclusion via a different path: The
small group of companies that are seeing real returns are fundamentally
reimagining workflows instead of applying AI to existing processes.

Three patterns emerge particularly frequently in practice.

The data foundation is insufficient. The new Bitkom survey shows where AI costs actually
arise. Not in licenses, but in infrastructure, data preparation, and
system integration.

Bitkom President Ralf Wintergerst sums it up succinctly: “Anyone who just buys a license has
purchased AI, but hasn’t changed anything yet.” For sales, this means specifically: AI-powered
lead scoring is only as good as the CRM data it’s based on. Incomplete
contact histories, duplicate accounts, and missing reasons for closing deals render any model ineffective.

The time saved goes to waste. Gartner surveyed 210 sales managers: AI saves salespeople an
average of 4.8 hours per week. However, 72 percent of organizations hardly invest this time in
value-adding sales work (Gartner, May 2026). The result is a striking ROI gap.
25 percent of sales organizations report a return of 50 percent or more on their AI investments, while 20 percent report an equally high negative return. Same technology,
opposite results.

Tools are piling up. Each team is introducing its own assistants, and each tool comes with its own
prompts and interfaces. Gartner explicitly warns that additional AI tools in
already complex workflows overload salespeople instead of lightening their load.

AI Agents and AI SDRs: When Hype Meets Reality

The most visible promise of the past two years is the autonomous AI agent: an “AI SDR” that
researches target customers, writes personalized messages, follows up, and schedules appointments, all
around the clock. In Germany, 11 percent of companies are already using such agents, and another
60 percent are planning or discussing their implementation (Bitkom, 2026).

Analysts’ forecasts are significantly more sobering than vendors’ promises. Gartner
expects that by 2028, there will be ten times as many AI agents as human salespeople,
but fewer than 40 percent of salespeople will report that these agents have improved their productivity
(Gartner, forecast, 2025). Across all industries, Gartner anticipates
that more than 40 percent of agentic AI projects will be discontinued by the end of 2027 due to
rising costs, unclear business benefits, or a lack of risk controls. At the same time, Gartner estimates
that only a small fraction of vendors advertising “agentic AI” actually deliver agentic
systems. The rest engage in what is known as “agent washing.”

A high-profile case illustrates just how wide the gap between promise and reality can be. In 2025,
TechCrunch reported that 11x, an AI-SDR provider funded by Andreessen Horowitz and Benchmark, listed companies on its website that, according to the company itself, were not
its clients. ZoomInfo stated that the product had performed significantly worse in a one-month test
than the company’s own sales representatives. 11x disputed parts of this account and removed the
relevant logos. The case is not evidence against AI agents per se, but it is a strong argument
for always verifying providers’ references and metrics yourself.

Search interest has also cooled: For the term “AI SDR,” the Google Keyword
Planner in Germany most recently showed a decline of around 90 percent compared to the previous year
(own analysis, September 2026; Google estimates).

The core problem with autonomous sales prospecting runs deeper. An agent that operates without reliable signals
primarily produces more of the same irrelevance—only faster. Added to this in
Germany is the legal framework: According to Section 7 of the Unfair Competition Act (UWG),
promotional emails to business customers generally require prior consent. This does not change if an AI
writes and sends the message.

Where Artificial Intelligence Really Delivers in Sales

Identifying and prioritizing signals. The most valuable question in sales is: Which account is
relevant right now, and why? AI can identify patterns in website behavior, CRM history, and external
market signals much faster than a human. Signal-mining approaches, such as those
used by Synthetic Leads, a product from ALEX & GROSS, analyze exclusively public
sources: job postings, tech stack changes, funding rounds, or
regulatory deadlines. Predictive lead scoring falls into the same category. The improvement rates cited in vendor studies
vary widely and should be validated internally against your own historical
closing rates rather than accepted at face value.

Prepare for calls. AI consolidates account information, competitive context, and past
interactions into a briefing before the salesperson picks up the phone. EVERLEAD, the marketing and sales platform from ALEX & GROSS, provides this context—for example, via AI Battle Cards—directly
within the workflow. The human conducts the call; the AI ensures they’re prepared.

Streamline routine tasks. CRM maintenance, call summaries, follow-up drafts: This is where
the hours saved—as measured by Gartner—come from. However, their value only materializes when
managers actively direct how the freed-up time is used.

All three areas have one thing in common: AI supports human decision-making rather than
replacing it. And all three stand or fall on the quality of the underlying data.

The Human Factor: What B2B Buyers Expect

B2B buyers have long been using AI themselves. In a Gartner survey of 645 procurement decision-makers,
45 percent reported using generative AI during their most recent purchase, primarily to
research vendors and products. 67 percent prefer a purchasing process that involves no
contact with sales representatives at all (Gartner, May 2026).

However, it would be wrong to conclude from this that sales teams will become obsolete.
69 percent of these same buyers turn to sales representatives to verify AI-generated information. Trust is strikingly divided in this regard: 51 percent consider misleading information
more likely to come from generative AI, while 49 percent believe it is more likely to come from salespeople. Buyers, therefore, do not blindly trust
either the machine or the human. They want both and cross-check them.

For sales organizations, this shifts their role. The salesperson is less often the first
source of information and more often the validator at crucial moments. This is precisely why
they need time, context, and credibility—in other words, the very things that well-deployed AI can unlock
and poorly deployed AI consumes. Gartner puts it as a clear recommendation: The future
of sales belongs to organizations that combine human empathy with AI-powered insights.

Five Questions to Ask Before Your Next AI Investment

A simple evaluation process can be derived from the available research. It is no substitute for a business case,
but it does help prevent the most common bad investments.

A realistic note on the time frame: Data cleansing and process adjustments typically take
longer than the technical implementation. Anyone expecting significant pipeline effects
within a few weeks is usually measuring too soon.

Frequently Asked Questions About AI in Sales

Is AI even worth it in sales?
Yes, if the fundamentals are right. According to McKinsey (2026), revenue gains from AI are
most often attributed to marketing and sales. At the same time, only 37 percent of
companies report a positive impact on EBIT. What makes the difference is data quality,
process integration, and the strategic use of the time saved—not the choice of AI model.

Will AI agents replace sales teams?
Not according to current studies. While Gartner predicts that by 2028, there will be ten times as
many AI agents as salespeople, it expects that fewer than 40 percent of salespeople will see a
productivity improvement as a result. 69 percent of B2B buyers want to verify AI-generated
information with a human sales representative.

Is it true that 95 percent of all AI projects fail?
The figure comes from a preliminary, unreviewed MIT report (2025). It states that
in approximately 95 percent of the organizations studied during the observation period, no measurable
effect on profit and loss was detectable. This is a serious warning sign, but it is not
synonymous with technical failure.

What is the biggest cost when implementing AI?
According to Bitkom (2026), AI users in Germany most frequently cite infrastructure (51 percent),
data preparation (50 percent), and system integration (41 percent) as major
cost drivers. Licenses for AI software are a major expense for only 21 percent.

How do I measure the ROI of AI in sales?
Before you start, define a pipeline metric with a baseline value, such as the conversion rate
of leads to opportunities or the pipeline contribution per sales representative. Activity metrics such as
emails sent or hours saved indicate efficiency, but do not yet reflect business success.

Is AI-powered cold calling permitted in Germany?
The same rules apply as for human cold calling. According to Section 7 of the German Unfair Competition Act (UWG), promotional emails to business customers
generally require prior consent; the GDPR also applies. A legal review should be conducted before
using automated outreach.

How ALEX & GROSS Supports Companies in This Process

Signal-based prioritization, clean data foundations, sales-marketing alignment, and AI where
it truly lightens the load for sales teams: ALEX & GROSS has been planning and implementing these strategies operationally
for B2B companies for 25 years, supported technologically by EVERLEAD® and Synthetic Leads. We
don’t start with the tool, but rather with the question of which sales problem needs to be
solved.

Sources

• Bitkom Research on behalf of Bitkom e. V. (September 14, 2026): For the first time, a majority of companies are using AI.
  Press release, n = 603 companies with 20 or more employees (industry association study, self-reported data)
• McKinsey & Company (August 25, 2026): The State of AI in 2026: On the Road to ROI, n = 1,719
  (Consulting study, self-reported data)
• Gartner (May 19, 2026): Gartner Survey Finds AI Saves Sellers Nearly 5 Hours Per Week, Yet 72% of Sales
  Organizations Fail to Reinvest Time in High-Value Activities, n = 210 (Analyst survey)
• Gartner (May 20, 2026): Gartner Survey Finds 69% of B2B Buyers Turn to Sales Reps to Validate AI-Generated
  Insights, n = 645 (analyst survey)
• Gartner (November 18, 2025 / July 28, 2026): Gartner Predicts That by 2028, AI Agents Will Outnumber Sellers by a Factor of 10
  (analyst forecast)
• Gartner (June 25, 2025): Gartner Predicts That Over 40% of Agentic AI Projects Will Be Canceled by the End of 2027
  (analyst forecast)
• MIT NANDA (2025): The GenAI Divide: State of AI in Business 2025 (preliminary research report, not
  peer-reviewed)
• TechCrunch (March 24, 2025): a16z- and Benchmark-backed 11x has been claiming customers it doesn’t have
  (journalistic investigation)
• Google Keyword Planner (September 2026): Search volume in Germany, ALEX & GROSS’s own analysis
  (Google estimates)
• EVERLEAD / ALEX & GROSS GmbH: Feature descriptions (everlead.ai, as of September 2026)
  Synthetic Leads / ALEX & GROSS GmbH: Methodology description for Signal Mining (synthetic-leads.com, as of
  September 2026)

Note: Statistics from analyst, consulting, and industry association studies are based primarily on self-reported data from
respondents. Forecasts describe expectations, not measured results. All figures should be viewed as indicative trends
rather than precise forecasts for your own company.