How to use AI in prospecting but not lose the humanity: a practical guide for B2B companies
- Martin Hurych
- 17. 6.
- Minut čtení: 8
AI can significantly speed up your journey to new customers. But if you use it indiscriminately, you risk losing what's most valuable - human contact and trust. In this article, I'll show you how to use AI as a smart assistant in B2B commerce, not as a substitute for a real salesperson.
You hear from all sides that you should start using AI - it will save you time, it will speed up the whole business process, it will do in minutes what used to take hours. It finds the right companies, sorts out relevant contacts, writes outreach proposals. Sounds great. But how do you do it in a way that doesn't lose the most important part - the humanity?
Recently, I watched a debate of experts from leading B2B companies who discussed this very topic in depth. In the article you will find not only the most important things from their experience, but also specific recommendations on how to use AI in business today in a smart way and what to avoid.
Further in the article:
What is AI prospecting and why companies are tackling it now
It used to be about quantity. Reach as many people as possible, expect someone to respond. Cold calls, mass e-mails, contact lists and trying to get the most out of trade shows. But the world has changed, and so has the way companies find new customers.
AI prospecting for B2B companies in 2025 is no longer about "shooting blind". Instead of endless lists of contacts, successful companies today work with signals - that is, information that indicates that a given company may have a reason to start considering your service or product. For example: changing suppliers, expanding a team, opening a new office, looking for a solution to a specific problem. AI can find these signals, connect them to other data, and intelligently prioritize them - faster and more accurately than any marketer could do manually.
What exactly does AI assisted prospecting mean?
AI-assisted prospecting is based on using artificial intelligence as a sales assistant to help select the right targets and prepare personalized outreach. Some of the most commonly used tools today include::
Clay - helps you build customized lists of companies or contacts. It collects data from different sources and can combine them smartly according to your criteria.
Apollo.io - a tool for managing contacts and sending personalized emails on a larger scale. It also tracks responses so you know what works.
Clearbit or ZoomInfo - provide a quick overview of the company: size, industry, technology, key people or current changes.
ChatGPT - a clever assistant for preparing outreach. It can summarize what the company does, suggest the first sentence of the message or check the tone. It can also help you prepare for a business meeting.
The combination of these tools can dramatically increase the effectiveness of your sales team - but only if you know how to use them.
And this is where the risk comes in. If you use AI without control, it can easily lead you in the wrong direction. Not every signal is relevant, not every output makes sense. And most importantly, no technology can yet replace the trust and relationship that develops between two people.
The biggest mistakes in AI prospecting
In discussions with sales teams and in working with clients, I very often find that they repeat four fundamental mistakes. They have one thing in common: salespeople are using AI as a shortcut instead of engaging it as a smart assistant. AI is not a magic new-lead button that does everything for you. It's a tool. And a tool only works as well as you know how to use it.
You don't know your ideal customer profile (ICP)
A lot of companies start with AI before they get the most basic thing straight in their heads: who our ideal customer is. AI then works with too broad a brief and reaches out to companies that aren't interested in the offer. Marketers are wasting their time and ruining their reputations. Until you know who you want to reach, AI will only speed you down the wrong path.
You want personalisation in a big way, but without control
How many times have you had AI write a sales email for you? Just insert a link to a website and let it create it. The AI can retrieve information, prepare an excerpt from the site, or design an outline. But the marketer has to decide what makes sense to that customer and why they're writing it in the first place. Otherwise, it's just another email that no one wants to read and most likely won't.
You rely too much on "hot signals" from the market
Position changes, recruitment, revenue growth - these are signals that AI can find quickly. But so can the competition. If you're only working with this "public data," you won't get any edge. Much more valuable are your own signals: who downloaded a price list, opened a newsletter, or signed up for a webinar.
4. You see AI as a shortcut, not an enhancement
AI should enable you to do your job better, not take away from it. If you set up automation, AI will do a lot of the tasks for you. However, if you don't have a thought out process, it will do them poorly. And you don't want that. If you take AI under your wing as more of a partner, it will help you do things better: offer data, show context, suggest what to do next. But the final decision is always yours.
How to properly set up AI tools for prospecting
Now that you know where companies most often make mistakes, let's see where AI really makes sense. How do we set it up to help discover new opportunities, support marketers, and not just make them operators of automation? Here are four specific areas where AI can save sales teams dozens of hours per week.
1. Targeting refinement
AI tools won't help you if you don't give them the right input. So the first step is to create a dynamic ICP - an ideal customer profile that responds to changes over time. And that's where AI can help you.
How to do it in practice:
Enter a list of your top 50 clients into a tool (e.g. ChatGPT+ corporate data/CRM output) and let the AI find commonalities:
Company size,
field,
number of branches,
growth phase,
technological maturity,
site visits.
Prompt Tip:
"Analyse these companies and suggest what other companies might have a similar profile. Give specific reasons why they might be interested in our solution [add a brief product description]."
2. Evaluation of signals
Signals from the market are not equivalent. If the AI evaluates equally a career site visitor and a
of a person who downloaded the case study and then clicked on the contact form, the result won't be worth much.
How to do it in practice:
Use AI to segment visitors based on behaviour. For example:
Score 90+: downloaded the e-book, visited the price list, came back again
Score 50-89: clicked on newsletter, visited case study
Score <50: short visits without interaction
Tools such as Clay, Factors.ai or even a simple prompt in ChatGPT can help you effectively evaluate combinations of behaviors and direct traders to where they are most likely to succeed.
3. Preparing a personalised address
Forget the generic "we're interested in your company". As long as AI works with well-structured input (e.g. from the web, company news or LinkedIn), it can prepare a quality pitch for you.
How to do it in practice:
Enter AI prompt type:
"Based on the company's website [url] and recent posts on LinkedIn, prepare a summary of the company's current situation + propose 3 hypotheses on what could be the company's biggest concern in terms of [the solution you propose]."
You can use the output as a springboard - but the personalisation must be fine-tuned by the marketer.
4. Effective follow-up
Salespeople often waste time manually taking notes on calls, planning their next move or transcribing emails. This is where AI clearly has its place.
How to do it in practice:
Send the call recording to a tool like Otter.ai or have it transcribed by another AI tool.
ChatGPT or another tool will generate a summary, key points, recommendations for the next step.
You can also automatically generate a personalized follow-up email from the output.
What AI can't do and where a good trader will always have the upper hand
Successful salespeople know that selling is not about forcing the customer to buy. It's about helping. The modern salesperson is more of an advisor and guide than an aggressive salesperson. And this is where AI can't compete. Your goal is to become the person the customer wants on their side. AI will supply you with data, summarize information about the company, pull signals from the market, but it won't know when a customer is pausing because they are considering risk. It doesn't sense uncertainty or hidden opportunity between the lines.
A practical example:
"We see that you are expanding production. Maybe you are dealing with digitalization. We would like to introduce you to our
Solution."
"That is now being solved by the management in Germany." End of call? Not for an experienced salesman.
"I see. We also encounter this in our practice. Does it make sense to send you a short summary that you can possibly forward to your colleagues?"
This is where the added value lies. AI will help you be more prepared, but it won't build trust, relationships and humanity for you. And in B2B business, where not only logic but also trust between people is the deciding factor, that's still what most contracts are based on.
AI in the sales team: How to set up processes to work
Introducing AI into the sales team is not about buying tools and waiting for "something to happen". If you want AI to actually help, you need to be smart about it. Innovation is for everyone, and with this proven approach, you too can combine business with technology:
1. Start small but purposeful
Select two to three salespeople who have a willingness to try new things and a natural curiosity. Give them the space to test specific use-cases (e.g., preparing pre-meeting documents, finding contacts, structuring outreach). Let them make mistakes, but record what works.
2. Choose tools based on what you are really solving
It is not important to have the most expensive AI software. What's important is that marketers actually save time or refine targeting. Somewhere ChatGPT will help, somewhere Clay, Apollo, somewhere a custom script in Notion. Good selection starts with the problem you want to solve - not what's trending on LinkedIn right now.
3. Set a clear metric for success
What does AI realistically bring? Faster preparation? More answers? Less unnecessary outreach? Establish simple metrics (e.g. time spent preparing before a call, % conversions for AI-prepared outreach) and evaluate after four weeks. Without that, you won't know if something is working.
4. Don't underestimate emotions, fears are normal
Some traders will be afraid of AI. They will feel that it is meant to replace them, or that they "can't do it". This is where open communication is important: AI is a tool, not a competitor. And everyone has the right to get to grips with it at their own pace. Give the people on your team time, support and ideally micro-training directly on a specific scenario from their agenda.
5. Leader must lead by example
If the sales director doesn't use AI and refers to it as a "toy", the team won't take it. But when a supervisor sends a thoughtful prompt or shows how AI has saved them an hour a week, it works as the best motivator on the team because, as with any change, the leader has to get moving first.
How to start using AI effectively
Start with yourself. Get clear on who your ideal customer is, what signals matter to your business, and where AI can really save you time. No technology can save you without a solid foundation. Even in an age full of smart tools, roughly 80% of Czech and Slovak companies can't do business.
The tools are changing fast. But the approach, the ability to ask, understand and guide the customer remains yours. So perhaps the time is coming when the best salespeople will be those who are great at combining technology and humanity. And those who learn to do this today will gain an edge in the marketplace that will be hard to catch. Will that be you?
