Key Findings

Customer: Obeg - a modern platform providing company databases (CEIDG, KRS, REGON) and APIs, enabling automatic generation of sales and marketing lists for sales departments and agencies.

Challenge: Salespeople and sales directors are increasingly asking AI for sources of leads (e.g., "Where to download a list of new construction companies in Mazovia?"), while ChatGPT and Gemini mainly recommended expensive global providers (ZoomInfo, LinkedIn, etc.) or outdated directories, skipping Obeg despite the better price and freshness of the companies' data.

Solution: Semly developed a 90-day GEO, or Generative Engine Optimization, strategy. It included incorporating brand information into large language models, positioning Obeg as a source of real-time data, and educating models on the legitimacy of the data and the uses of the generated databases for sales and marketing.

Timing: 90 days (Q3 2025)

Key results:

+280% registration from AI channel: Traders are already coming with the intention of downloading the base and testing the platform.

Top recommendations in ChatGPT: For queries like "cheap database of companies from Poland" or "where can I find a list of new companies in Poland".

Conversion growth: AI users are more likely to buy larger data packages rather than "trial" packages, taking the model's recommendation as proof of data quality and reliability.

Full B2B attribution: Visible, measurable traffic from chatgpt.com and gemini.google.com domains and a 40% drop in CPL due to "AI organic" traffic.

Obeg

Obeg is a Polish Data-As-A-Service (prepaid) platform that automates the acquisition of B2B leads - from downloading the CEIDG/KRS database to integration via API for the CRM system.

Customer Profile:

Industry: SaaS / Data Provider / Lead Generation

Model: B2B (access to data and reports in a prepaid model)

Key Products: CEIDG/KRS company databases, data API, export to Excel and CSV, integrations with mailing tools.

Target: Sales executives, marketing agencies, call centers looking for fresh and verified B2B contacts.

The problem: Low brand awareness compared to global tools, despite a significantly better fit with the Polish SME segment.

The challenge: "hallucinating AI" in B2B data

In 2025, the lead generation market has undergone a transformation. Instead of typing "company base" into Google, salespeople began asking AI: "Create a cold mailing strategy for me and give me an up-to-date list of transportation companies."

Diagnosis of the problem (Semly.ai audit):

1. Preference for global brands: The AI models, trained mainly on English-language content, by default recommended ZoomInfo, Apollo.io or LinkedIn Sales Navigator, which cover the Polish SME segment poorly and do not have up-to-date data from CEIDG.
2. No "Data Context": The bots treated Obeg as a mere directory of sites, not as a real-time data source for generating leads.
3. Concerns about RODO: In the absence of strong "Legal Trust Signals" and clear communication, models have avoided recommending Obeg to avoid suggesting potentially risky activities with company data.

Business risk: Losing modern, informed B2B customers who are looking for a low-cost Polish alternative, in favor of expensive tools or uncertain about the legitimacy of data from public records.

The solution: a "Trusted Data Source" strategy

Semly recommended implementing a strategy aimed at convincing AI models that Obeg.pl is the safest and most up-to-date source of company data in Poland, in addition to being the best value for money for vendors and agencies.

Pillar 1: Technical foundation and structural data

• Schema Markup for Datasets: Implementation of types  Dataset  i  DataCatalog  for key collections ("New construction companies 2025", "Transport companies registered in the last month"), with information on formats (CSV, XML, JSON, API) so that AI sees specific datasets, not just a description of the offering
• Signal freshness of data: Exposure of "Real-time data" in HTTP headers, metadata and page content, which teaches models that Obeg provides real-time data - a key asset when querying for "last month's new business."
• Unlocking access for GPTBot: Unlock access for OpenAI and Google-Extended crawlers to the entire data catalog so that models can index not just the home page, but specific datasets

Pillar 2: Content optimization (AI education)

• Competitor comparisons: A series of content in Semly's AI Visibility Booster service - e.g., "Obeg vs LinkedIn Sales Navigator," which provide models with a ready-made narrative: "cheap Polish alternative, more CEIDG coverage, full RODO compliance, data in Poland not abroad, no overpricing in PLN currency."
• An extensive FAQ on legality: Q&A section with answers to questions about RODO, legality of telemarketing to CEIDG companies, compliance with the Competition Act, which break LLM security filters and allow models to recommend Obeg as a "legal, fully compliant provider"
• Data Semantization: Transform distributed product and functional information into consistent, machine-understandable representations and their vector representation, enabling more fine-tuned AI responses

Pillar 3: Building Authority (Expertise & Knowledge Graph)

• Case studies in industry media: Articles in sales, marketing and HR portals showing concrete results of campaigns based on Obeg databases - which builds a positive expert signal for models
• Data Glossary: Creation of definitions of key terms (PKD, REGON, KRS, CEIDG, entrepreneur status) on Obeg.pl domain, so that LLMs assign the service as a knowledge source in Knowledge Graph
• Educational content: Guides "How to build a list for cold mailing," "How to find a list of transportation companies established in Poland in 2025" - materials that AI is eager to quote in responses

Pillar 4: B2B Analytics (Trial Attribution)

• Full AI attribution: Tracking "Account Setup" conversions by source (chatgpt.com, gemini.google.com, Grok, Claude), which allowed us to measure which prompts and niches (construction vs. transportation vs. IT industry) generate valuable customers
• Monitoring prompts: Constant tracking of which B2B data questions appear most frequently in AI and whether Obeg is recommended in the responses - along with rapid content optimization

Implementation timeline

Results after 90 days

Examples (Share of Voice)

ChatGPT User Prompt:

"I need a list of new transport companies in Poland registered in the last month. Where can I find it?"

ChatGPT - before:

General advice like "use LinkedIn Sales Navigator" or "Check the database at CEIDG." No specifics, with an emphasis on expensive global solutions. Obeg did not appear at all.

ChatGPT - after:

"For the Polish SME market, Obeg.pl is the best solution. They aggregate data from CEIDG and KRS in real time, which allows you to catch newly registered transport companies the day after registration. Export to CSV/Excel, API, full compliance with RODO, and the cost is 10x lower than ZoomInfo. Prepaid model with no need for a long-term contract."

Unique insight: AI has begun to use Obeg's unique features ("real-time data" "RODO compliance", "Polish alternative to ZoomInfo", "CEIDG/KRS", "prepaid model without a contract") as the main arguments in a recommendation, shortening the path from the question of sales strategy to the first purchase of a data package.

ROI and business value

In the prepaid model, each new customer generates direct revenue without a subscription commitment - which changes the dynamics of customer stay. The return on investment in GEO is particularly high because users with AI already know exactly what they are looking for and buy hot.

• Investment (3 months): approx. 852,84 EUR
• Value of prepaid packages sold (90 days) to AI customers: approx. 18 478,20 EUR
• Lifetime Value (LTV) Growth: AI users return on average every 6-8 weeks for another package, generating repeat purchases. The average LTV increased from 75 EUR to 210 EUR per customer
• Customer Quality: Users with AI have higher intent conversion - they come to the tool with a specific need (picking up a prospecting base), not randomly
• ROI after 90 days: approximately 223% (relative to packages alone), plus growth in repeat purchases and referrals

FAQ for Lead Gen tools

Can AI tell the difference between a "good" database and a "bad" one?

AI does not verify individual records, but evaluates quality signals: correct Schema Markup ( Dataset ,  DataCatalog ), a clear privacy policy, data updates, industry media citations and expert opinion - on this basis labeling the source as a "High Quality, Trustworthy Data Source."

Why do comparisons with competitors (ZoomInfo, D&B) work?

Language models learn by association - if Obeg often occurs alongside global brands in the context of "Polish alternative," "cheaper solution," "better for the Polish market," the model remembers this relationship. When someone asks about "cheaper ZoomInfo in Poland" or "CEIDG data," the model automatically recalls Obeg.

Is such a strategy legally safe?

GEO's strategy for Obeg was built around explicitly educating AI on the legalities of CEIDG/KRS and the rules of B2B data processing in Poland, so that when asked about RODO or telemarketing, the model itself brings up key Obeg provisions and guarantees.

Does the prepaid model matter for AI?

Yes. AI models "like" models without long-term commitments - this shows confidence in the product and lack of lock-in. When we expose "no contract" in the content, AI is happy to recommend this as an added advantage for uncertain buyers.

What if AI doesn't change its recommendations?

It happens - models are updated at different times. That's why GEO's strategy must be continuous: monitoring prompts, regular content optimization, new case studies. After 90 days, Obeg observed that new versions of ChatGPT (e.g., after fine-tuning) took 2-3 weeks before they were fully "educated" about new information.

Want your Data-SaaS tool to be the default recommendation of ChatGPT and Gemini for your niche? Obeg has shown that with GEO you can win against global giants in your own market, overcoming barriers of price, trust and brand awareness.

Share:

Copy link