Frequently Asked Questions

Expert Consilium turns a business question into a structured review of options, risks and next steps. It draws on multiple separate AI perspectives, then analyzes their individual answers and synthesizes them into one clear report to support your decision - the Decision Brief.

A chat with a single AI tool gives you one answer from one perspective, based on what you just typed. Expert Consilium has multiple separate AI perspectives analyze the same question independently, and then each one anonymously evaluates and ranks the others' answers. Only then do you receive one synthesized, structured report.

No. Expert Consilium organizes the analysis and gives you a structured view of options, risks and arguments - the decision, and responsibility for it, always stays with you. A higher analysis level never guarantees the right decision; it only widens the scope of the analysis.

Essential (4 AI perspectives) is for relatively bounded decisions where you want a fast, structured second look. Consilium (8 AI perspectives) is for important business decisions that benefit from broader strategic, operational, financial, market and risk analysis. Grand Consilium (12 AI perspectives) is for complex or high-consequence decisions where the broadest analysis and closer scrutiny of assumptions are justified.

The price of a level is fixed and covers your main question, one additional scenario (for example "what if X changes" or another option you want us to consider) and the written report. After the report is delivered you can ask one short question about it in the app (up to 250 characters); that is also included in the price. There is no surcharge for the additional scenario. Current prices for levels and packages are on the Pricing page.

A perspective is a predefined role from which the same business decision is examined, each with its own logic - for example strategy, execution, risk, finance, the customer or contractual exposure. Each perspective is a separate call to an AI model, with its own instructions and short principles for that role. The roles at each level are described on the Perspectives page.

Not in the first stage - each perspective first gives an independent answer without seeing the others. Only in the second stage does each perspective anonymously evaluate and rank the others' answers, and in the third stage everything is synthesized into one report.

One clear Decision Brief: a summary of the situation with a clear recommendation, the key arguments and risks from multiple perspectives, open questions and one concrete next step - not the raw answers of individual perspectives or internal rankings.

Yes. Before synthesis, each AI perspective anonymously evaluates and ranks the others' answers, showing where they agree, where they diverge, which risks deserve attention and which assumptions should be verified. Only then is the result synthesized into the final report.

Yes. All questions and answers are automatically screened for inappropriate content. A request containing offensive or vulgar language will be declined, and you will be asked to rephrase it.

The check works in two layers. The first is a local word list for Serbian and English, which runs instantly and without sending any data to outside services. The second is the OpenAI Moderation API, which covers the other languages and more complex cases. A request containing inappropriate content is declined before processing; if problematic content appears in a model's response, that part of the text is automatically removed or replaced.

No. The content check happens before the request is processed, so a declined request does not use up a credit and is not charged. You receive a message asking you to rephrase the question and resubmit it.

The local check (Serbian/English) runs on our own servers, without sending any data to outside services. When a broader check is needed, the text is sent to the OpenAI Moderation API - a dedicated service that, per OpenAI's published policy, does not use data to train models; any retention period is set by OpenAI under its own terms. The text is sent masked. This is separate from the AI models that generate the report itself, which are covered by the policy described in the answer on ZDR.

Expert Consilium supports analyses in 42 languages. The system preserves the language and script of the original request throughout the analysis and final report unless another output language is explicitly selected.

No. Just describe the situation, your objective, constraints and the question you want examined clearly, in plain language - the system structures the analysis itself.

Partly. The system automatically replaces the data it can reliably recognize with placeholders: email addresses, phone numbers, IBANs and account numbers, card numbers, personal ID, social security and passport numbers of many countries (when the format is valid or a label such as passport no or ID number precedes the number), and tax or company registration numbers when they are labeled. Some countries' formats are not covered. We do not automatically recognize names of people and companies, addresses, amounts or dates. To hide a name, put it in double square brackets, for example [[Acme Ltd.]].

Expert Consilium is designed with data confidentiality as a key requirement. Analysis requests are routed according to active privacy requirements, with a Zero Data Retention requirement and a ban on data collection. ZDR is evaluated at the specific provider-endpoint level rather than assumed merely because a particular model or routing platform is used. There are documented exceptions: the inappropriate-content check, the model that searches the internet, and models that run at Google, where the provider may retain a request for safety monitoring, for a period the provider sets.

ZDR (Zero Data Retention) means the provider does not retain request content after processing. It is evaluated at the specific provider-endpoint level rather than assumed merely because a particular model or routing platform is used. Analysis requests require ZDR and a ban on using data for model training. Exceptions to ZDR are: the inappropriate-content check (OpenAI Moderation), the model that searches the internet, and models that run at Google (including Gemini), where the provider may, under its own rules, retain requests for safety monitoring and abuse prevention. The duration and terms of that retention are set by the provider, and we do not control them.

Yes, one of the perspectives in the report may search public sources on the internet. To do that, it may use the text of your question, masked in the same way as for the other perspectives, as a search query. The search itself is carried out by that perspective's provider under its own data retention rules, so we cannot guarantee the same level of protection as for the others. No perspective opens your documents or verifies facts about your company. If you do not want a piece of data to appear in the query, put it in double square brackets, for example [[Company name]], and it will be replaced with a placeholder.

No. Analysis content is used only to prepare your report and is not shared with other clients or published.

No. Expert Consilium provides analytical support for business decisions and does not replace licensed legal, tax, accounting, investment, medical or other regulated advice. Verifying key facts and assumptions remains the client's responsibility.

No. No analysis level, including Grand Consilium, guarantees the right decision or business outcome. A higher level means a broader range of perspectives and closer scrutiny of assumptions, not a guaranteed result.

Work begins once payment is confirmed, and you receive the report once the analysis is done. We do not offer deferred delivery.

Yes. Essential is designed exactly for relatively bounded, easily reversible decisions where a fast, structured check from four separate AI perspectives is enough.

If you are not sure which level you need, we first suggest one. Separate AI perspectives then examine the question, each one anonymously evaluates and ranks the others' answers, and everything is synthesized into a structured Decision Brief.

Logging in, creating an account and things like resetting your password happen on app.expertconsilium.com - the separate application where you work with your analyses. This site (expertconsilium.com) describes the service; your account and the analysis process itself are on app.expertconsilium.com.

For founders, executives and advisors who want a disciplined second look before making an important commitment - for example when choosing a growth path, partner, supplier, price or new market, when comparing options for a team, or when a professional advisor wants help structuring the questions for a client.

We are not the right choice if you need urgent medical or safety help, fact verification, or regulated legal, tax, investment or accounting advice.

A single AI answer is a single view. Expert Consilium sends your question to 4, 8 or 12 separate AI perspectives, has each one anonymously evaluate and rank the others' answers, and compiles one Decision Brief - worthwhile when a mistake would be costly and hard to undo.

You can, and that is better than relying on a single answer. The difference is that you would have to put together the question, the different approaches and the comparison yourself, and each tool sees only your question without challenging the others. Expert Consilium puts the same question to separate AI perspectives. Each one then anonymously evaluates and ranks the others' answers, and everything is assembled into one structured report with options, risks and open questions.

The report supports your decision; it does not make it for you. The decision is always yours. The assumptions and open questions in the report show what the recommendation rests on, so you can check whether your situation differs from those assumptions.

For decisions that carry a real cost and are hard to reverse: signing a contract, entering a partnership or a new market, changing prices, making a major investment. The more important the decision, the more it is worth looking at the options, risks and open questions from several angles before choosing.

It's not random. Each perspective is given a clearly defined analytical style before it answers - for example a focus on risk, on numbers and cash flow, or on long-term strategy - based on recognizable, publicly known approaches from business strategy, negotiation and risk management, not on training with private sources, and without any claim that a specific individual took part.

No. Agreement among multiple AI perspectives is not fact-checking - the conclusions rely on the information you provide, so we recommend verifying key facts and figures before you decide.

No, not under the current settings. Logging of request and response content, sending data to external monitoring platforms, training models on request data and publishing prompts are all disabled. Technical metadata may be processed only for the operation and security of the service, and Zero Data Retention is mandatory for provider routing.

You first choose a level of analysis on the Request an analysis page. If you are not sure which level you need, you describe the situation and we can suggest one. Payment follows on a secure payment page, and work starts once payment is confirmed. You submit the question for analysis in the app and receive the structured report as soon as the analysis is done.

You order on this website, on the Pricing page, or in the app. You pay on a secure payment page, and your credit is activated as soon as the payment is confirmed. If you have a question, write to us at support@expertconsilium.com.

Yes, you can order wherever you are. You order on this website, on the Pricing page, or in the app. You pay on a secure payment page, and your credit is activated as soon as the payment is confirmed. If you have a question, write to us at support@expertconsilium.com.

A question with inappropriate content is declined without using up a credit, and you can rephrase it and send it again.

Your conversation and report remain available in the app for 30 days after the last activity, and you also receive the report by email, so be sure to save a copy. While the conversation is visible in the app, you can delete it yourself with the Delete conversation button, which removes the conversation, the report and all copies of them from our server (server backups are deleted automatically within 60 days). After 30 days the conversation is no longer shown, so we delete it from the server at your request, sent to office@expertconsilium.com.

Yes. You describe the offer and add what you have. The report compares options and risks and shows which assumptions the offer rests on. The report is not legal advice. For legal questions, please consult a lawyer.

You provide the plan and the numbers. The report shows which assumptions the numbers rest on and what changes if one of them does not hold.

No. You order a single analysis or a package, which is paid once and does not renew automatically.

Write to us at support@expertconsilium.com or visit the Support page. For anything not covered here, we're happy to answer personally.