Meet Vera AI VendorBenchmark is now Vera AI. Same team, same numbers. See what changed →
Request a demo Free Trial
Security and data FAQ | Vera AI by VendorBenchmark
Vera AI · Security

63 questions security, procurement and legal teams ask us, answered with numbers. Written for the reviewer working through a questionnaire: every answer describes something implemented today, and the last section lists what we have not built. Most questionnaire lines map one to one.

THE SECURITY REVIEW Security overview Security and data FAQ Data processing addendum Subprocessors Trust center Responsible disclosure Security review package Security and data review guide (PDF)
STEP 1
Validated on arrival
Type allowlist, 200 MB cap, byte-signature check.
STEP 2
Isolated in your tenant
Row-level security and private storage, enforced by the database.
STEP 3
Every touch logged
Views and downloads, yours and ours, in an audit trail you can read.
STEP 4
Deleted for real
Your retention policy, your export, and a hard delete that removes rows and files, with a certificate.
ON THIS PAGE

11 sections, 63 questions

01Architecture and tenant isolation502Identity and access903Data lifecycle, upload to deletion804AI and model data flow505Benchmarks and data anonymity306Integrations and API407Monitoring, audit and incident response708Infrastructure and subprocessors509Compliance and legal910People and process411What we do not have yet4
01 · ARCHITECTURE AND TENANT ISOLATION

Architecture and tenant isolation

5 questions
1.1How is tenant isolation actually enforced?
At the database, not in application code. More than 270 customer tables carry your organization ID on every row, and Postgres row-level security policies enforce the boundary on every query. A session from your organization physically cannot return another organization's rows, even if there were a bug in the application layer. Storage policies on the file buckets mirror the same organization rule. This is the control everything else stands on.
1.2What happens if there is a bug in your application code?
The database still refuses the rows. Row-level security runs inside Postgres, below the application, so a cross-tenant read attempt returns a row-level denial from the database, not an application error a bug could route around. You can reproduce this yourself in a trial: create two organizations, upload to one, and watch the other organization's session get denied at the API.
1.3How do you prove isolation keeps working as you ship?
39 automated cross-tenant isolation test suites run in CI on every change. The failing cross-org read is a permanent regression test, not a one-time demo, and we demonstrate it live in security reviews on request.
1.4Do customers share a database?
Yes, and we say so plainly: this is a multitenant Postgres database, not a database per tenant. The boundary is enforced by the database engine through row-level security on more than 270 customer tables, and re-proven by 39 CI isolation suites on every change. What you get is a boundary that is tested continuously rather than assumed per instance.
1.5Are files isolated the same way as database rows?
Yes. Contract files live in private storage buckets whose policies mirror the organization rule from the database. There is no public URL to any customer file, ever, and access outside your organization is refused even with a direct storage path.
02 · IDENTITY AND ACCESS

Identity and access

9 questions
2.1What multi-factor authentication do you support?
App-based TOTP, with session step-up enforcement for sensitive actions. Organization owners can enforce MFA org-wide by policy, and MFA is mandatory for every member of Vendor Benchmark staff. Passkeys and WebAuthn are not built yet; TOTP is the second factor today.
2.2Do you support single sign-on?
SAML and OIDC SSO per organization. The Require SSO switch is a real enforcement: once it is on, password and magic-link sign-in are refused for your domain. Organization owners keep a break-glass path so a broken identity provider cannot lock the tenant out.
2.3Do you support SCIM provisioning?
SCIM 2.0, working today with Okta, Microsoft Entra and any SCIM-capable identity provider: provision, list and deactivate from your directory. Offboarding in your directory is offboarding here. SCIM tokens are stored as SHA-256 hashes and expire within one year.
2.4What are the password rules?
Twelve characters minimum, and every new password is screened against known breach corpora before acceptance, using a privacy-preserving k-anonymity range query in the style of Have I Been Pwned, so the password itself never leaves our server. Under Require SSO, password sign-in is refused entirely for your domain.
2.5What are the session lifetimes and session controls?
Users see every active session under Settings, can revoke any single one, or sign out everywhere else with one action. A sign-in from a new device emails the account owner. Organization owners can also set a session policy under Settings: a maximum session age and an idle timeout, so your own limits apply to every member.
2.6Can we restrict access to our corporate network?
Yes. Organization owners can set an IP allowlist in CIDR notation under Settings, so sessions outside your listed ranges are refused for your organization.
2.7How do roles and per-contract access work?
Owner and member roles per organization, managed by your own organization owners. Sensitive actions, member management, data export, retention and deletion among them, are owner-only. Below the role layer, individual contracts can be restricted to named people within your organization, so need-to-know applies inside the tenant boundary too.
2.8Can a customer account become a platform admin?
No. Platform admin rights live in a separate table that no customer-facing code path can write to. There is no privilege-escalation route from a customer account, and admin status cannot be self-assigned.
2.9How are API keys and SCIM tokens protected?
Both are stored as SHA-256 hashes: we can verify them, nobody can read them back, including us. Both expire within one year, so a forgotten credential ages out on its own. API key management is owner-only.
03 · DATA LIFECYCLE, UPLOAD TO DELETION

Data lifecycle, upload to deletion

8 questions
3.1What is validated when a file is uploaded?
Three checks before anything else touches it: document types only (PDF, Word, Excel, CSV), a 200 MB cap, and a check of the file's actual byte signature rather than its extension, so a renamed executable is rejected even if it claims to be a PDF. The upload is then written to a private bucket under your organization's path and recorded in the audit trail.
3.2What information do you collect about individuals?

For your authorized users: name, work email, the timezone they set, and the sign-in credential (a password hash, or nothing when you use SSO). Beyond that, the platform records only what operating it securely requires, and nothing for marketing or profiling.

Audit and session records
the IP address, user agent and time of each sign-in and each action on contract data, and the devices a user is signed in on, kept so your owners can inspect the trail
Activity attribution
which user ran which benchmark, AI analysis or export, in your own tenant
Voice and meeting features, only when used
audio and transcripts of the calls and meetings you seat Vera in, including participant names
Your uploaded documents
whatever personal data your contracts contain, under your control; the service requests no special-category data

No government identifiers, no payment card data, no home addresses, no biometrics, no behavioral advertising profiles. Annex 1 of the data processing addendum states the categories of data subjects and personal data in the processor's terms.

3.3How do file downloads work?
Every download uses a signed URL minted per authorized request, expiring after roughly 60 seconds, and never pre-signed into HTML. A leaked link or a saved page is dead before it can be abused. There are no public or long-lived links to any customer file, and every download lands in the audit log.
3.4How long do you retain our data?
As long as you need the service, and no longer. Organization owners set a retention policy under Settings, Data controls, and the same screen shows the whole lifecycle: collected, isolated, retained on your policy, deleted on request.
3.5What happens when we delete a contract?
The database rows and the stored files are both removed. Deletion is a data path built into the product, not a support ticket. Every contract deletion also writes a durable, timestamped deletion certificate your organization can download and keep for its own records.
3.6What happens when we delete our whole organization?
Every storage bucket belonging to the organization is swept along with the database rows, and the deletion writes the same durable, timestamped certificate. Nothing waits in a queue for a human to action.
3.7Can we export everything we have put in, including the files?
Yes. Organization owners export the organization's data from the data controls screen without a support ticket, and the export includes the file bytes through 1-hour signed links, not just the rows that describe them. The audit log has its own full CSV export.
3.8What happens at the end of our contract with you?
Export, then deletion, in that order and on your initiative: owners pull the complete export including files, request full deletion, and the rows and files are removed with a deletion certificate to prove it. Your data does not linger because leaving was inconvenient to us.
04 · AI AND MODEL DATA FLOW

AI and model data flow

5 questions
4.1When does AI touch our documents?
At two points. When a document is uploaded it is read once, automatically, to index it for search and to fill in its key terms: dates, values, notice periods. Everything after that, a benchmark, a brief, a question, a call, runs only when a member of your organization starts it. No training job reads your documents at any point.
4.2What exactly leaves your platform, and to whom?

Only what a feature needs, and only to the provider that runs it.

Anthropic (Claude)
the document text and the question for every written analysis, called from our servers only
Voyage AI
document text in passages, to build the search index, and the search questions
OpenAI
audio, transcripts and the deal facts Vera looks up, when you use her voice on a call or in a meeting
ElevenLabs, Recall.ai, HeyGen
call audio, meeting audio and Vera's spoken audio, only for the call copilot, the meeting seat and the video face

A live voice session streams audio from your browser to the voice model, and the browser never holds a provider credential. Nothing goes to an ad network or a third-party analytics service. Results come back to your tenant under the same row-level security as the document they came from. The full list with roles is on the subprocessor page.

4.3Is our data used to train AI models?
Not by us, ever, and not by the model providers that run the analysis and the voice: Anthropic and OpenAI do not use API data to train models. The zero-data-retention addendum with Anthropic is signed. The remaining providers, for search embeddings, transcription and the meeting seat, are listed with the data each one receives on the subprocessor page.
4.4Are AI prompts and outputs logged and attributable?
Every AI run is metered and attributable to the user and organization that ran it, so your admins can see who analyzed what. Outputs are stored in your tenant under the same row-level security as the source data.
4.5How do we know an AI answer is grounded and not invented?
Figures in AI-generated reports are grounded in your stored data and citation tagged, and AI reports pass a groundedness check before they reach you. Curated analyst reports are verified by hand. When the assistant cites a clause, the citation points at your actual document.
05 · BENCHMARKS AND DATA ANONYMITY

Benchmarks and data anonymity

3 questions
5.1What goes into the benchmark pool from our contracts?
Anonymized aggregates only, and the pooling is one-way: market low, median, high and percentile positions come out, but nothing goes in attached to a name. Nothing that could identify your company, your contract or your specific terms is exposed to another customer.
5.2Could another customer identify us from a benchmark?
No. Benchmark output is pooled and anonymized before any other customer sees it. What other customers see is the market shape: ranges and percentiles, never a data point traceable to a company, a contract or a set of terms.
5.3Is our pricing visible to other customers?
Never in identifiable form. Your pricing contributes to pooled, anonymized aggregates, and only those aggregates are visible to anyone outside your organization.
06 · INTEGRATIONS AND API

Integrations and API

4 questions
6.1How are integration and connector credentials stored?
Encrypted at rest with AES-256-GCM, scoped to your organization, and used only server-side. No connector credential is ever sent to a browser.
6.2How is inbound email and webhook traffic authenticated?
Inbound email and webhooks verify HMAC signatures before anything is processed, and each organization's inbound email addresses carry unguessable capability tokens, so mail sent to yours can only land in your workspace. A forged or misdirected message is uploaded, not filed.
6.3Are your outbound webhooks signed?
Yes. Events we send you are HMAC-signed with timestamps, so your systems can verify us the same way we verify inbound traffic, and replayed payloads are detectable.
6.4How is the API rate limited?
Per user, per IP and per API key, and in production the limiter fails closed: if the rate-limiting backend is unreachable, requests are refused rather than waved through. Authentication, upload and AI routes carry their own limits.
07 · MONITORING, AUDIT AND INCIDENT RESPONSE

Monitoring, audit and incident response

7 questions
7.1What does an audit event contain?
Actor, action, entity, organization, IP address and user agent, written for every view, download, create, update and delete of contract data. Reads are logged, not just writes: a view of a contract file is an audit event.
7.2Is Vendor Benchmark staff access logged too?
Identically. Analyst views and downloads of your files flow through the same audit logging as your own members' actions, in the same trail your owners can inspect and export. Staff access is never quiet access.
7.3Can we export the audit history?
Yes. Organization owners export the complete audit history as CSV at any time, without a support ticket.
7.4Can we stream audit events to our SIEM?
Yes. Subscribe an audit webhook and signed batches of audit events arrive every 15 minutes, cursor-tracked so nothing is skipped and a missed delivery resumes where it left off. Setup lives in the in-app Security Center.
7.5Do you detect and alert on unusual activity?
An hourly anomaly detector reads the audit trail and emails your organization owners when one account's downloads, exports or permission changes look unusual. It runs on our side, so it works even if you never wire up the SIEM stream.
7.6What is the Security Center?

A live posture page inside the app under Settings, Security, Posture. It exists so your team can verify the claims on these pages against your own tenant instead of taking our word for them.

MFA coverage
your organization's second-factor coverage, member by member
Audit export
the complete audit history as CSV
SIEM webhook
the signed 15-minute audit stream, set up here
7.7Have you had a breach, and what is your incident commitment?
No customer data breach to date. The commitment if that ever changes: you hear it from us directly within 72 hours of us confirming an incident that affects your data, followed by a written post-mortem, and the audit trail shows exactly what was touched. Short-lived file links and per-request access bound the blast radius by design.
08 · INFRASTRUCTURE AND SUBPROCESSORS

Infrastructure and subprocessors

5 questions
8.1Where does the platform run, and how is the edge hardened?
The application runs on Vercel and the data lives in Supabase, Postgres for records and private buckets for files. Both providers are SOC 2 Type II audited. Every response carries HSTS, nosniff, frame-deny and a strict Content-Security-Policy with a fresh per-request nonce, and no secret is ever stored in the codebase; keys live in server-side environment variables only.
8.2Where is our data located?
In the United States, all of it, today. EU data residency is not built; it sits in the Not yet column of the claim ladder rather than in a footnote.
8.3How is data encrypted, and who manages the keys?
TLS for everything in transit. At rest, the database and storage are encrypted by our infrastructure provider, and the keys are provider-managed. We do not offer customer-managed keys (BYOK) yet, and we say so rather than blurring the line.
8.4Who are your subprocessors, and where do we track changes?

11, each with one job.

Supabase
Postgres and file storage
Vercel
application hosting
Anthropic
document analysis and written AI answers
OpenAI
voice, transcription and live market lookups
Voyage AI
document search embeddings
Recall.ai
the meeting seat
ElevenLabs
live call transcription
HeyGen (LiveAvatar)
the optional video face
Cloudflare
meeting audio relay and the public site
Resend
transactional email
Upstash
rate limiting and a short-lived cache

The current list with purpose, data and region always lives on the subprocessor page; we notify customers before adding a provider that touches contract data.

8.5What is your backup posture?
Provider backups are enabled on the database and storage, with continuous backup and daily snapshots, and a second, independent backup copy runs alongside them. The first restore from backup has been completed, so the backups are exercised and not only enabled, and the recovery objectives, RPO and RTO both under 24 hours, are in place. Restore tests run quarterly, and the record is in the trust center's evidence table.
09 · COMPLIANCE AND LEGAL

Compliance and legal

9 questions
9.1What certifications do you hold?

None issued to Vendor Benchmark yet, and we say so rather than borrowing a badge. What exists today:

SOC 2
Type I targeted Q4 2026, Type II after the observation window; the controls are implemented and mapped to evidence, and the report goes into the NDA document room when the auditor issues it
Infrastructure
Supabase (database and storage) and Vercel (hosting) are SOC 2 Type II audited; their reports are available from them
Independent validation
the first third-party penetration test is complete, annual from here; the executive summary is in the NDA document room
ISO 27001
not started; sequenced after SOC 2

The trust center carries the dated status of each, and moves an item only when the report exists.

9.2Where are your Master Subscription Agreement and the other agreements?

Public, on the agreements page, so legal can read them before anyone signs anything.

Master Subscription Agreement
the commercial terms every subscription sits under (PDF)
Data processing addendum
the processor terms, with the annexes and standard contractual clauses (PDF)
Service level agreement
availability, support response and credits (PDF)
Acceptable use policy
what the service may and may not be used for (PDF)
Mutual NDA
for the security review itself; we sign yours or bring ours (PDF)
Order form
the template an order is written on (PDF)

The executed copies come from your account team; to run the agreements through your own paper or redlines, email info@vendorbenchmark.com.

9.3What is your SOC 2 status?
SOC 2 Type II controls are implemented and mapped to evidence, evidence collection is underway, and the report will be available under NDA when the auditor issues it. We do not claim SOC 2 until there is a report to show.
9.4Where are you on GDPR and CCPA?
GDPR
the DPA is public at /dpa, the subprocessor list at /subprocessors, and deletion is a live product path; the records of processing are being finalized with counsel
CCPA
the privacy notice is live, and we do not sell personal data
9.5How do we execute your DPA?
Read the standard DPA at /dpa, public rather than gated behind sales, and download the addendum as a PDF. To execute it, email info@vendorbenchmark.com and we complete signature as part of your procurement, on our paper or through your process.
9.6How do we get your policies and questionnaire answers?

Through the NDA document room: request access on the trust center.

The links
token-scoped, gated by an NDA click-through, and every open is audit-logged
Inside
the policy set, a pre-completed security questionnaire, the penetration test summary, and the SOC 2 report when it is issued
9.7How do we report a vulnerability?

Through the responsible disclosure policy, referenced from our security.txt.

Acknowledgment
within 3 business days
Good-faith research
gets safe harbor
Security is owned by
Fredrik Filipsson, founder, reachable at info@vendorbenchmark.com
9.8Do you have a third-party penetration test?
Yes. The first independent penetration test has been completed, and the cadence from here is annual. The executive summary is available under NDA through the document room, where every open is logged. Between tests, the standing controls are the CI isolation suites and the public disclosure program with safe harbor.
9.9How do you handle law enforcement and government requests?
We require valid legal process before disclosing customer data, and we notify the affected customer unless we are legally barred from doing so. We do not volunteer customer data, and we interpret requests narrowly.
10 · PEOPLE AND PROCESS

People and process

4 questions
10.1Who at Vendor Benchmark can access our contracts?
The analysts who produce your deliverables, for that purpose. Their access runs through the same audited paths as yours: every analyst view and download is logged with the person, the action, the time, the IP and the user agent, in the trail your owners can inspect and export. MFA is mandatory for every member of our staff.
10.2How do you secure your own staff?
MFA is mandatory for every staff member, access follows least privilege, and admin actions are logged in the same audit trail as everything else. Platform admin status lives in a separate table that no customer-facing code path can write to, so it is granted deliberately, never inherited or self-assigned.
10.3How do you secure your development process?
39 cross-tenant isolation suites run in CI on every change, so an isolation regression cannot ship quietly. Changes go through code review, and no secret is stored in the repository; keys live in server-side environment variables only.
10.4What guards external sharing of reports?
Sharing outside your organization is an explicit action by someone in your organization, never a default. Anything shared externally passes a confidentiality scan for sensitive terms first, and external share links expire on their own.
11 · WHAT WE DO NOT HAVE YET

What we do not have yet

4 questions
11.1Are you ISO 27001 certified?
No, and we have not started. SOC 2 Type II comes first; ISO 27001 is sequenced behind it.
11.2Do you support customer-managed keys (BYOK)?
Not built. Encryption at rest is provider-managed today. If BYOK is a hard requirement for you, tell us; it is exactly the kind of signal that reorders a roadmap.
11.3Do you offer EU data residency?
Not built. All data is hosted in the US today, and we would rather you plan on that than discover it in a data-flow diagram later.
11.4Do you support passkeys or WebAuthn?
Not built. TOTP is the second factor today, with step-up enforcement and org-wide policy. Passkeys are on the same honest Not yet list as everything else we have not shipped.
STILL HAVE QUESTIONS

Three ways forward.

1. Email the security owner
Fredrik Filipsson, founder, at info@vendorbenchmark.com. Send the questionnaire as it is; we answer all of it.
2. Request the NDA document room
On the trust center: the policy set, the pre-completed questionnaire and the penetration test summary, through scoped, expiring links.
3. Book the architecture walkthrough
Request access and mention the review. We demonstrate the failing cross-tenant read live.
A QUESTION WE DID NOT ANSWER

Send your security questionnaire. We answer all of it.

We complete questionnaires in your format, walk your team through the architecture, and demonstrate the isolation model live, including the cross-tenant access attempt failing.

Set up the review →