Our public commitment
Your data is never used to train our models.
When you upload documents, transcripts, or other materials to UmergenceAI, that data is used exclusively to deliver analysis and insights to you. It is not used to train, fine-tune, or improve any AI model — ours or anyone else's.
This commitment applies to all data processed through our platform, including documents uploaded to deal workspaces, meeting transcripts, financial data, investor communications, and any other materials you share with us. For clarity, this commitment does not prevent us or our service providers from using limited operational telemetry — such as error logs, performance metrics, and security event data — to operate, secure, and monitor the services. However, such telemetry is not used to train, fine-tune, or improve any machine learning or AI model and is handled in accordance with our Written Information Security Plan and Data Processing Agreement.
We hold our technology partners to the same standard. Our large language model providers — Anthropic, OpenAI, and Google — each provide API-level commitments that data submitted via their APIs is not used for model training. We verify these commitments as part of our ongoing vendor oversight process.
How we handle your data
Processing scope
Your data is processed only to provide the services you have contracted with us. It is not analyzed for business intelligence, product improvement, or any purpose beyond delivering your requested analysis.
Isolation
Each customer's data is logically isolated. Your documents, transcripts, and analysis outputs are not visible to or accessible by other customers' accounts.
Sub-processors
When your data is sent to an AI model for analysis, it is transmitted to the relevant provider's API endpoint under terms that prohibit that provider from using your data for model training. Current LLM sub-processors: Anthropic (Claude), OpenAI (GPT-4o), Google (Gemini). Full sub-processor list available on request.
Contractual commitment
Our Master Service Agreement and Data Processing Agreement include explicit zero-training provisions. Contact us to review these terms before signing.
Contract language
The following clauses appear in our Master Service Agreement and Data Processing Agreement as a binding commitment to our customers.
Zero-Training Commitment
Customer Data processed by or through the UmergenceAI platform is not used, and will not be used, to train, fine-tune, improve, benchmark, or evaluate any machine learning model, large language model, embedding model, or artificial intelligence system, whether operated by UmergenceAI or any third-party sub-processor. Customer Data is processed solely for the purpose of delivering the contracted Services to Customer. This zero-training commitment is implemented in our customer contracts and public communications.
Processing Limitation
UmergenceAI processes Customer Data only as instructed by Customer, consistent with the terms of this Agreement and any applicable Data Processing Agreement. UmergenceAI will not process Customer Data for any purpose beyond the scope of the contracted Services without Customer's prior written authorization.
Sub-Processor Commitments
UmergenceAI represents that each sub-processor authorized to process Customer Data — including large language model providers (currently Anthropic, OpenAI, and Google, as applicable) and cloud infrastructure providers — has provided contractual or published commitments materially equivalent to the zero-training commitment above with respect to API-accessed model inference. UmergenceAI will notify Customer of any material change to sub-processor commitments within thirty (30) days of becoming aware of such change.
No Cross-Tenant Data Use
Customer Data is logically isolated from data belonging to other UmergenceAI customers. Customer Data will not be combined with, compared against, or used to derive insights on behalf of any other customer's account without Customer's express written consent.
Verification
Upon Customer's written request (no more than once per calendar year), UmergenceAI will provide documentation confirming compliance with these commitments, which may include applicable sub-processor terms, SOC 2 audit reports (when available), or written attestation from UmergenceAI's Qualified Individual.
Regulatory basis
This commitment supports our obligations under the GLBA Safeguards Rule and FINRA's vendor management expectations, and reflects what has become standard practice for enterprise procurement in the financial services sector. UmergenceAI's sub-processors — Anthropic (Claude), OpenAI (GPT), and Google (Gemini) — each have no-training commitments for API usage, which we verify against current terms and disclose to customers in our documentation.