Question 1

What is Zorka?

Accepted Answer

Zorka builds, verifies and maintains a machine-readable model of what a company knows, has done and can prove, so its people and its AI can act on it. The model is built from the company's own documents and owned by the company. The first application, B2B sales, is live.

Question 2

Is Zorka an AI SDR?

Accepted Answer

No. Zorka doesn't send messages. It researches and qualifies accounts from what your company has already delivered, maps the buying committee and drafts the briefs, then hands them to the tools you already use. Most tools start from the prospect; Zorka starts from your business.

Question 3

What is the Graph Steward?

Accepted Answer

The Graph Steward is Zorka's governance layer. Every fact carries its source, a date and a status (hypothesized, reported, verified or rejected). Fact, inference and hypothesis are kept apart, and contradictions are flagged. People steer the work; nobody can overwrite the evidence.

Question 4

Our systems already hold this information. What does Zorka add?

Accepted Answer

Your systems keep the records. Zorka keeps what they mean together, with the evidence for each claim. It reads what you've already written — delivery reports, proposals, case studies and specifications — and connects it into one model: what you delivered, for whom, what it changed and what proves it. Nothing is migrated; your systems stay where they are.

Question 5

How is Zorka different from enterprise search or a chatbot on our documents?

Accepted Answer

Search tools find and summarise documents. Zorka checks what the documents actually support: every fact keeps its source and date, and contradictions between documents are flagged. AI built over documents rebuilds its context for every answer; Zorka gives every model and every use the same sourced context. The two can work together, with a search tool on top of a governed graph.

Question 6

How is Zorka's knowledge graph different from others?

Accepted Answer

It is built from what your company documented, not from what people describe. Many tools build their model from interviews, workshops, forms or prompts; people forget, hope and fill gaps, so the model inherits all of that. Zorka's graph is ontology-based: it reads delivery reports, proposals, case studies and specifications, and an ontology defines what a valid fact looks like. Every fact keeps its source and date; a claim with nothing behind it is not accepted as fact. People still set the direction and review the work; nobody can overwrite the evidence.

Question 7

What can Zorka do beyond sales?

Accepted Answer

Sales is the first application, not the boundary. Most of the graph describes the company itself — what it delivered, what it can prove and the problems it solves — not only the pipeline. That is why the same graph can serve other decisions. Today the Evidence Report is available on your own material; compliance & audit and bids & proposals are next.

Question 8

Who is Zorka for?

Accepted Answer

Business owners, founders and the managers who run a company's key processes, in mid-market B2B companies whose sale rests on work they have delivered and where each deal involves several stakeholders. Sales leaders usually start, because sales is the first application. Technical and engineering teams gain too: their delivered work is what the model reads, so solved problems stop being solved again. The one requirement is that you have written your work down; a five-minute fit check tells you whether there is enough to build on.

Question 9

What do we need to give you?

Accepted Answer

As many documents about your business as you can share. Proposals, delivery reports, case studies, specifications, technical documentation and meeting notes all work, in any format; they do not have to be tidy or written for sales. The more you share, the more complete and reliable your graph becomes. You do not write an ICP: Zorka builds it from your documents. The only other thing we ask is where not to look — markets or sectors you will not sell to.

Question 10

How is the model built?

Accepted Answer

From documents the company already has: delivery reports, technical documentation, specifications, meeting notes, proposals and case material. Construction agents extract the facts and map them onto the model, each with its source, and the Graph Steward checks them. There are no workshops and no forms about your own business. How fast it goes depends on access to the documents, not on meetings.

Question 11

Do I need technical skills?

Accepted Answer

You need to be able to say who you will not sell to. Tools in the orchestration layer ask you to encode what you know as tables, prompts and enrichment logic, then maintain it as it drifts. Zorka reads what your company has already written.

Question 12

What if our documentation is thin or out of date?

Accepted Answer

Then that is the first thing you will learn. Before you are asked for anything else, you get a graph-completeness score: a percentage against what a graph of your business has to contain, with every gap named. A company whose knowledge lives only in people's heads has nothing for Zorka to read, and we say so before anyone commits.

Question 13

How long does a graph build take?

Accepted Answer

Typical delivery is one week on Starter, two weeks on Growth, three to four weeks on Scale and six to eight weeks on Enterprise. Construction agents do the extraction and mapping, so the time depends on how quickly we get access to the documents, not on meetings.

Question 14

Can we try Zorka before we commit?

Accepted Answer

Yes, in two ways. The free plan is €0 a month, with no trial clock, no expiry and no card needed: a graph built from your public materials, plus ten researched companies and five buying-committee briefs every month. Or start with the Evidence Report: a short, light run of the whole approach on your own documents, before any full graph build.

Question 15

What is the Evidence Report?

Accepted Answer

A report on what your company can and cannot prove about itself: every contradiction, every evidence gap and every claim your material makes that your evidence does not support, each with its source and date. It costs €2,500, credited in full against a graph build within 90 days; if you stop, you keep the report. Pointed at one client's case portfolio without being asked to audit anything, it found the same testimonial, with the same client's name, in four case studies about unrelated industries.

Question 16

How do we start?

Accepted Answer

1. A five-minute fit check. Is enough written down to build on? 2. A graph-completeness score. What your documents support, and what is missing. 3. The Evidence Report. €2,500, credited in full against a graph build within 90 days. 4. The graph build. Owned by you, from €1,500, paid once, with half credited back over twelve months. 5. Applications on the same graph. Sales from €149 a month on self-serve plans, or from €1,200 a month by contract.

Question 17

What does the Sales application do?

Accepted Answer

It starts from your graph: Zorka builds your ICP, with personas, pains and triggers, from what you have actually delivered. Then three steps nest. A screen tells you whether a company is worth looking at, with the rule that failed or the signals that matched. Research gives a qualified profile with its evidence chain, a pain hypothesis and a fit score with its confidence. A brief maps the buying committee (typically three to five people) to who owns which metric, researches each named member and drafts a personal LinkedIn invitation, follow-up and email for every one of them, each traceable to its evidence. Ask us to show you a company we rejected, and why.

Question 18

Do we need to give you our ICP?

Accepted Answer

No. Zorka builds it from your documents: the cases you delivered, the problems they solved and the results they produced become a profile of who has the same problem, with personas, pains and triggers. You only tell us where not to look, such as markets or sectors you will not sell to.

Question 19

Do you send the emails?

Accepted Answer

No. We do everything before the send: the graph, the ICP, the account list, research, qualification, the buying committee, individual profiles and drafted messages. Then we deliver it into the tools you already use — your CRM or sales engagement platform — through the API or as files. We feed your stack; we do not replace it.

Question 20

Which tools does Zorka work with?

Accepted Answer

Any external solution. Zorka has a native HubSpot app that prepares research and briefs inside the CRM; results go into any other CRM or tool through the API or as files. Every paid plan can add an integration seat for CRM webhooks, API access and scheduled batches. Scale adds API read and connectors; Enterprise adds SSO/SCIM, an audit log and an SLA.

Question 21

Do I need HubSpot?

Accepted Answer

No. Zorka has a native HubSpot app, but it does not depend on it. Bring a list from any CRM, a CSV or another external source, and take the results into whatever you use, through the API or as files.

Question 22

How does pricing work?

Accepted Answer

You can buy three things. A graph build — the model of your business that you own — costs from €1,500, paid once, with half credited back over twelve months. The Evidence Report costs €2,500, credited in full against a build within 90 days. A plan for the Sales application starts at €0 a month (free plan) and goes up through self-serve and contract plans. Annual billing is available. Prices exclude VAT. Full plan details and unit rates are on the Pricing page.

Question 23

Why is screening free?

Accepted Answer

Because charging for a “no” makes people screen less, which is exactly backwards. Screens are never invoiced: you pay for the research and briefs on companies that pass, and only for the highest step a company reached, once. Every plan includes a monthly volume of screens.

Question 24

Who counts as an operator, and who is a viewer?

Accepted Answer

An operator runs research, briefs and exports and, on paid plans, brings a monthly block of work. A viewer reads, comments and shares, free on every plan. Zorka charges for the people who run the work, never for the people who read it.

Question 25

What happens when the monthly block runs out?

Accepted Answer

Further companies are billed at your plan’s rate. You get alerts before the block is exhausted. Usage beyond the block can be capped or set to a hard stop. Unused work expires monthly. Current rates are on the Pricing page.

Question 26

Are there credits or packs?

Accepted Answer

No. Each plan includes a monthly block of research and briefs, screens are never invoiced, and work beyond the block is billed per company at your plan’s rate. Unused work expires monthly.

Question 27

Why pay for a graph build if half comes back?

Accepted Answer

Because it is real work with a real deliverable: a model of your business built from documents you already have, a completeness report, and an asset you keep. Half the fee is credited against your subscription over the following twelve months, and the credit shows on every invoice.

Question 28

Who owns the graph?

Accepted Answer

You do. Your knowledge graph, with every extracted fact and its provenance, is yours, and so are your source documents and outcome data; you can download, copy and use the graph at any time. Zorka owns the ontologies, the extraction and mapping logic, and the agents that build and maintain it. The subscription is what keeps those facts true and reasons over them.

Question 29

Can we download the graph and use it without a subscription?

Accepted Answer

Yes. The graph is yours; you can download it and use it however you like. But it will not stay current on its own. What you have delivered stays true, while the accounts around it change: people move roles, companies reorganise, and buying signals lose their value quickly. The subscription pays for the Graph Steward, which re-checks your accounts on your plan’s schedule and looks for news and events that match the triggers you set. Without it, the graph is a snapshot of the day you downloaded it.

Question 30

How is our data handled?

Accepted Answer

Before any build, you get a written answer on where your data is processed, which model providers may handle it, how it is stored and protected, and how it can be exported or deleted. Zorka is model-agnostic: each task goes to a model that fits it. The providers are listed in your data processing agreement, and you are told before that list changes. A fixed set of models for one client can be discussed as a custom option. Operators steer and run the work, viewers read and comment, and nobody can overwrite the evidence.

Question 31

Do you use our data to improve Zorka for other clients?

Accepted Answer

Only if you agree to it in your contract, and never for sensitive information, intellectual property or know-how. Without that agreement, nothing from your graph is used for any other client.

Question 32

Can Zorka run in our own cloud?

Accepted Answer

Zorka runs as a service in its own cloud; that is the standard setup. Hybrid or stand-alone installation is available on request and is scoped and priced as a custom project.

Question 33

How do we know Zorka's answers are right?

Accepted Answer

Every fact carries its source, a date and a status, and fact, inference and hypothesis are kept apart. Zorka reduces the risk of hallucination by breaking reasoning into small, ontology-constrained steps over a governed graph. Verification is never skipped silently: beyond your plan’s verification volume, new facts are published as Unverified and listed in the Steward digest. Key claims can also be cross-checked across models before they are accepted. No system is right every time; the point is that you can see what each answer rests on.

Question 34

What is a triplet?

Accepted Answer

A triplet (also called a triple) is one fact with a subject, a relationship and an object: this offering / improves / this metric. Every one carries its source, its date and its status. That is why Zorka can tell you what it does not know, which matters when you are deciding whether to trust a score.

Question 35

How up to date is the information?

Accepted Answer

The Graph Steward re-checks accounts, including news and events that match your triggers, on a schedule: monthly on the free plan, weekly on Starter, and daily for priority accounts (weekly for the rest) on Growth and Scale. Enterprise is set by contract.

Question 36

Who is behind Zorka?

Accepted Answer

Zorka Lab Sp. z o.o., a Warsaw company founded by five people from semantic technology, ontology research and B2B sales. The CEO, Andrei Kuryan, is an ontology engineer and co-initiator of the TRIZ Ontology. The Chief Architect, Vasili Paulik, has spent 25 years building commercial semantic technology. All five founders are on the About page.

Question 37

Where is Zorka Lab based?

Accepted Answer

In Warsaw, Poland. The company is Zorka Lab Sp. z o.o.; you can write to contact@zorkalab.com.

Privacy Policy

1. Introduction

2. Data We Collect

3. How We Use Your Data

4. Third-Party Services and Data Processors

5. Data Sharing and Disclosure

6. Data Retention

7. Data Security

8. International Data Transfers

9. Your Privacy Rights

10. Children’s Privacy

11. Changes to This Privacy Policy

12. Contact Us