Terms and limitations

What is explicitly out of scope

• Non-standard data preparation: we work with tables as given; reshaping messy formats into usable schemas costs time we do not bill here
• Document extraction: contracts, PDFs, and unstructured records are out of scope
• Period churn: "logo churn" calculated from a snapshot is unverifiable and disabled; churn measurement needs a starting cohort and time-series events
• FX conversion: if your table mixes currencies, we report the currencies present; we do not convert between them
• Revenue recognition: we sum what your table labels as recurring; we do not apply accounting standards
• Source authenticity: we verify that the arithmetic matches your table; we do not audit whether the table itself is accurate or current
• Any statement about whether the deal is good: we report numbers and reconciliation only

Data handling preconditions

Before processing real, confidential data, we commit to the following in writing:

Data and processors:

• An explicit list of which data elements your table contains
• Which elements are personally identifiable (names, IDs, email, etc.) and which are not
• Who processes the data: whynot core engine only; no third-party ML vendors or LLM training

Geography and access:

• Permitted geographic region(s) for processing (local, EU, US, other)
• Who can view the table and report: named individuals on your team only
• No storage in shared repositories or multi-tenant environments

Contractual basis:

• Your authorization to share the table with PreDealCheck
• Your data retention and deletion requirements
• Any contractual restrictions (e.g., from your own customers) that affect our processing

Training and re-use:

• Explicit ban: your table and results will never be used to train AI models or build public datasets
• If we later want to use anonymized patterns from your case for model improvement, we request a separate written right

First demos are synthetic:

• All published examples, walkthroughs, and sample reports use synthetic (made-up) data, not real companies or real clients
• Your name, your metrics, and your numbers do not appear in any external material

Technical limitations:

• Local computation (running the reconciler on your table without uploading to cloud services) limits data transfer and can reduce certain compliance risks
• Local computation limits data transfer but does not determine compliance. Your counsel and data-protection officer must review the actual service, processor agreements, and your obligations before processing any real data.

Boundaries we keep

PreDealCheck's role is limited to factual reconciliation and open questions. We will never:

• Recommend: "You should accept this deal" or "reject it." No buy/sell advice.
• Value: "This company is worth X" or "the price is fair." No valuation.
• Certify safety: "This deal is safe" or "this metric is accurate." No assurance that passing reconciliation means the deal is sound.
• Condition fees on outcome: No part of our fee is contingent on whether the deal closes, fails, or meets your expectations.
• Make legal claims: We do not allege fraud, only discrepancy. We do not render legal opinions or audit opinions.

On disclaimers and legal status:

A disclaimer alone does not establish what a service legally is. If PreDealCheck's reconciliation reports might inform investment decisions, securities law (including SEC investment-adviser rules in the US and equivalent rules in your jurisdiction) may apply. Before running a commercial pilot, your legal counsel must review the actual service we provide, the wording of our reports, and our fee structure, and confirm that our role fits your regulatory requirements.