Privacy Policy

1. Overview

This Privacy Policy describes how Agentic Assets (“Agentic Assets,” “we,” “us”) handles information in connection with EQUIRE (the “Service”). EQUIRE is a business-to-business platform for commercial real estate teams. For most Customer Content, our customers (the organizations that use EQUIRE) are the controllers of the data and we act as their processor, handling the data on their behalf and under their instructions.

This policy works alongside our Terms of Service.

2. Information we collect

We collect the following categories of information:

• Account information. Name, work email, organization, role, and authentication data when you create or are invited to an account. If you sign in with a third-party provider such as Google, we receive basic profile information from that provider.
• Customer Content. The documents and data you and your team upload to or generate within the Service, such as offering memoranda, rent rolls, financial statements, leases, diligence files, valuation models, and memos, along with the data extracted from them.
• Usage and device data. Log data, IP address, browser and device information, pages viewed, features used, and timestamps, collected to operate, secure, and improve the Service.
• Communications. Information you provide when you contact us, request a demo, or request support.

3. How we use information

We use information to:

• provide, maintain, secure, and improve the Service;
• process Customer Content to deliver features such as extraction, analysis, valuation, and drafting;
• authenticate users and manage accounts and organizations;
• provide support and respond to your requests;
• monitor for, prevent, and address security incidents, fraud, and abuse;
• comply with legal obligations and enforce our agreements; and
• communicate with you about the Service, including service-related and, where permitted, product updates.

We do not sell your personal information, and we do not use Customer Content to advertise to you.

4. Artificial intelligence processing

The Service uses third-party large language models and related AI services to process Customer Content and your instructions and to generate output. When Customer Content is sent to these providers for processing:

• the providers of the AI models we use by default do not use your Customer Content to train their models, under their terms with us or with the Vercel AI Gateway;
• providers may keep inputs and outputs for a limited period under their standard terms; and
• we send Customer Content to them to run the Service’s AI features, including a background check that sends figures read from your spreadsheets, with their labels, to an AI model to test whether each figure was matched to the right field. This check does not change what you see.

Some features, such as IC memo drafting and prospecting runs, let you choose a different AI model. The first two points above do not cover a model you choose. Its provider handles the Customer Content sent to it under its own terms, which may allow it to keep that content or use it to train models.

These commitments do not cover web search and research requests, which section 5 describes.

AI output may be inaccurate or incomplete and should be independently verified, as described in our Terms of Service.

5. Web search

Some features search the public web, including market research, comparable sales search, prospecting runs that look for listings matching your investment criteria, building your company profile from public sources, and web research requested through EQUIRE’s in-product assistant. Most of these searches run through the Vercel AI Gateway, which sends the search queries to web search providers (currently Parallel and Perplexity). For these searches, an AI model accessed through the Gateway also processes each search request in order to submit it to the search provider. Prospecting runs can also send the web address of a public page found during the run, such as a listing, to Exa to find similar pages.

Some research features use Corbis, a research service operated by Agentic Assets, the company that operates EQUIRE. Deep Research in EQUIRE’s assistant, which is off unless you turn it on, sends research questions to Corbis. Prospecting runs can send market screening requests, market data requests, and research questions to Corbis. Economic Insight, which is on by default in EQUIRE’s assistant for a deal, sends requests for Federal Reserve (FRED) economic data, such as search terms and data series names, to Corbis. Corbis can run these requests through the web search, AI, and data providers it uses.

Market research queries, including research started by an AI assistant you connect to EQUIRE, are built from the market, submarket, property type, research topic, and time period, together with any custom request text that you or the assistant enter. Other searches are built from the details of the request, such as a property’s name and street address, a city, a property type, a company name, or terms that EQUIRE’s assistant takes from your conversation or your deal records. We do not attach your uploaded documents to search or research requests, but these requests can include details taken from them, such as a property’s name and address or a tenant’s name.

The search providers named above, the AI model that submits searches through the Vercel AI Gateway, and the providers Corbis uses handle these requests under their own terms. We do not currently guarantee that they do not keep these requests or use them to train models, and the commitments in section 4 do not apply to these requests.

6. AI assistants you connect to EQUIRE

You can connect EQUIRE to a third-party AI assistant, such as ChatGPT or Claude, through the EQUIRE connector. Nothing is shared with an assistant until you approve the connection on EQUIRE’s consent screen. There you choose the organizations the assistant may reach, which of the permissions it requests it receives, and its access level: Read Only, Reads & Staged Changes, or, where available, Routine Workflow Updates. The permissions you approve decide what the assistant can read, and making changes also requires an access level that allows them. The assistant can reach only the records your EQUIRE role already permits in those organizations. Each organization’s administrators decide whether its members can connect assistants to it and set the highest access level any connection may have there. Your connection cannot go beyond that level. While assistant access is off for an organization, members’ assistants cannot read or change its data, starting with their next request. Administrators can see each member connection that includes their organization, with the member, the assistant application, the permissions and access level approved, when it was approved, and the time of its last activity and the number of tool calls, calls that make or stage a change (counted whether or not they succeed), and errors attributed to that organization over the past 30 days. They can also remove their organization from a connection. A request already in progress when an administrator turns access off or removes a connection still completes, and research the assistant started before then still finishes in EQUIRE. The organization’s audit log records when a member connects an assistant or revokes a connection, when a member tries to connect while assistant access is off in every organization they belong to, when the organization’s setting refuses a connection a member submits, and when an administrator changes that setting or removes a connection. A member’s repeated refused attempts with the same assistant within an hour are recorded once. The same audit log includes the tool-call records described below for calls attributed to the organization.

Depending on the permissions you approve, EQUIRE returns the following information to the assistant when it asks:

• Account details. Your name, email address, and EQUIRE user ID; the organizations you selected, with your role in each and its subscription plan; and the access granted to the connection. Every connection can read these details.
• Organization profile. Your organization’s company profile, website, and the public sources cited in that profile; its investment thesis and criteria; its underwriting, sourcing, and memo style preferences; recurring investment committee questions; common reasons for passing on deals; and a list of its most recently updated deals.
• Deal records. Deal and property names, addresses and location, property type, size, year built, status, key dates, pricing and cap rate, and portfolio summaries, including total value, net operating income, returns, and leverage.
• Deal health issues. Issues flagged on a deal with their descriptions, supporting evidence (which can include short passages and values from your documents), resolution notes, due dates, and the name and user ID of the team member each is assigned to.
• Document excerpts. Passages of text from your deal documents, with the document name and page or section. The connector returns excerpts, not complete files.
• Valuation outputs. Saved valuation snapshots with their assumptions and the basis of their inputs, returns, deal comparisons, and what-if scenarios. What-if scenarios are calculated on request and not saved.
• Diligence and activity. Diligence items with their descriptions, status, priority, due dates, and notes, and the name of the team member each is assigned to; and a deal’s activity history, which lists changes to the deal, exports, and requests that team members’ connected assistants made about it, with the email address of the team member behind each entry, the names and roles of team members added to, removed from, or given a new role on the deal, and the names of documents uploaded to or removed from it.
• Notes and proposals. Deal notes and staged proposals created through the connector.
• Prospects. Prospecting results and their sources.
• Listing data. Commercial real estate listings, including pricing, income and occupancy figures, lease terms, tenant and guarantor names, property details and location, photos, links to listing documents such as offering memoranda, text from the brokerage’s listing page, figures from offering memoranda such as rent roll entries with tenant names, and the broker contact details published with each listing (name, title, brokerage, email address, phone number, and links to the broker’s profile and contact card). This is information brokers publicly post with their listings on brokerage websites; EQUIRE shows it as posted and refreshes it.
• Market research. Research plans, budgets, and run status, source links, the market evidence and figures gathered for a deal, metro trends for its market, and how the deal’s rent, occupancy, pricing, and valuation assumptions compare with that evidence.

If you approve an access level that allows changes, the assistant can also act in EQUIRE on your behalf: add notes to deals, create and update diligence items, add listings to your prospecting pipeline, start market research and attach report links to deals, and stage valuation or review proposals. Staged proposals do not change a deal until an authorized person reviews them in EQUIRE. Changes remain limited by your role and permissions and are recorded under your account in EQUIRE’s audit trail. The connector cannot delete records or record investment committee decisions. EQUIRE marks each tool that changes data so that your assistant can ask you to confirm before it runs; whether and how it asks is up to the assistant.

We record the tool calls an assistant makes through the connector, including the tool used, your account and organization, the deal involved, the assistant application, the time, the outcome and any error message, and the number of results. A tool call about a single deal also appears in that deal’s activity history with your email address and the time, where anyone who can see the deal, and any assistant they use, can see it. These records do not keep the body text of notes, search queries, prompts, or written reasons, and links given to a tool, such as report web addresses, appear in them only as one-way hashes. The records can keep short labels and settings, such as a note subject, a diligence item’s title and assignee, a scenario name, or a search filter. They do not contain the information EQUIRE returns to the assistant.

You can review and revoke a connection at any time under Connected Assistants in your EQUIRE settings. Revoking a connection ends the assistant’s access for every new request immediately, and access to an organization also ends when your membership in it ends. A request already in progress when you revoke still completes, and research the assistant started before then still finishes in EQUIRE.

Information returned to an assistant leaves EQUIRE. Once the assistant receives it, the assistant provider’s own terms and privacy policy govern how it is used, stored, and shared, including whether it is used to train models. We do not control how an assistant provider handles that information, and the commitments in section 4 do not apply to it. Please review your assistant’s terms before you connect it.

7. Service providers and sub-processors

We share information with vendors who process it on our behalf to run the Service, under contractual confidentiality and data-protection obligations except where noted below. Key categories include:

• Cloud hosting and delivery: to host and serve the application (e.g., Vercel).
• Database and authentication: to store account data and Customer Content and manage sign-in (e.g., Supabase).
• AI providers: to process Customer Content for AI features (e.g., Anthropic, Google, OpenAI, and xAI models accessed through the Vercel AI Gateway), under terms that do not allow them to train on your data, except for a model you choose, as section 4 explains.
• Web search: to run the web searches described in section 5, through the Vercel AI Gateway and the providers it uses for search (currently Parallel and Perplexity, with an AI model that submits each search), and through Exa. Parallel, Perplexity, Exa, and that AI model handle search requests under their own terms, as section 5 explains.
• Email and communications: to send transactional and account email (e.g., Resend).

Corbis, the research service described in section 5, is operated by Agentic Assets rather than a third-party vendor. The web search, AI, and data providers that Corbis uses handle research requests under their own terms.

We may update our sub-processors as the Service evolves. To request the current list, contact us at admin@agenticassets.ai.

8. How we share information

We share information only as described in this policy, including:

• within your organization, with other authorized members and administrators of your account;
• with the service providers and sub-processors described above;
• with an AI assistant that you connect to the Service, at your direction, as described in section 6;
• in connection with a merger, acquisition, financing, or sale of assets, subject to appropriate confidentiality protections; and
• when required by law, legal process, or to protect the rights, property, or safety of Agentic Assets, our users, or others.

9. Data retention

We retain Customer Content for as long as your account is active or as needed to provide the Service, and as instructed by your organization. We retain account and usage data as needed for the purposes described in this policy and to meet legal, security, and operational requirements. When data is no longer needed, we delete or de-identify it. Backups are purged on a rolling basis.

10. Your choices and rights

Depending on your location and your relationship with us, you may have rights to access, correct, export, or delete personal information, or to object to or restrict certain processing. Because we often act as a processor on behalf of your organization, we may direct certain requests to your organization (the controller), or ask you to submit them through your administrator.

You can update much of your account information within the Service, and your organization administrator can manage member access. To make a privacy request, contact us at admin@agenticassets.ai. Account deletion is available within the Service and removes associated data in accordance with our deletion process.

11. Cookies and similar technologies

We use cookies and similar technologies that are necessary to operate the Service, such as keeping you signed in and maintaining security. We may also use limited analytics to understand and improve how the Service is used. You can control cookies through your browser settings, though disabling necessary cookies may affect functionality.

12. Data location and international transfers

We and our service providers may process and store information in the United States and in other countries where our providers operate. Where information is transferred across borders, we rely on appropriate safeguards as required by applicable law.

13. Security

We use reasonable technical and organizational measures designed to protect information, including access controls, encryption in transit, and tenant isolation. No system is completely secure, and we cannot guarantee absolute security. Please use a strong password, protect your credentials, and notify us promptly of any suspected unauthorized access.

14. Children’s privacy

The Service is intended for business use by adults and is not directed to children under 18. We do not knowingly collect personal information from children.

15. Changes to this policy

We may update this Privacy Policy from time to time. If we make material changes, we will provide reasonable notice, such as by posting the updated policy with a new effective date or by notifying you through the Service. Your continued use of the Service after the changes take effect constitutes acceptance of the updated policy.

16. Contact

For privacy questions or requests, contact us at admin@agenticassets.ai, or write to Agentic Assets, the company that operates EQUIRE.