Security & Trust

Which data, processed where, under whose agreement.

Funnel Economics has two components, and their data paths are not the same. This page states each path separately — including what is built, what is not yet in force, and what we decided against.

Three data paths, three trust boundaries

Path A · Funnel Economics Core

What it is: The native Salesforce 2GP managed package in your tenant.
Where it is computed: In Salesforce (Apex). Salesforce Enterprise Edition, no Data Cloud.
Which data: Your pipeline data.
What leaves Salesforce: The analysis runs inside your tenant. Its AI commentary is off by default; when enabled, the package masks personal values before the callout leaves. With Agentforce, the Einstein Trust Layer applies on top.
Under whose agreement: Your Salesforce and Agentforce terms.

Path B · Core + Funnel Economics Strategist (optional)

What it is: Core plus the Strategist server and knowledge base. The analysis is done by the language model you choose.
Where it is computed: Figures in Salesforce; context prepared on a service we operate; analysis in your language model.
Which data: Your pipeline data as prepared context — with the names of sales staff and opportunities, and record links.
What leaves Salesforce: The context passes through the Strategist server we operate, and from there to the language model you have chosen.
Under whose agreement: The model: your own agreement with Anthropic, Google or Microsoft. The leg through our service: a data-processing agreement is required; we are preparing it.

Website and Discover

What it is: This public website. Discover is built for a demo org that holds fictitious data.
Where it is computed: A server in the EU is intended for the Discover backend.
Which data: Intended for Discover: registration (business email, first name, company, role, country) and consent records. No customer data.
What leaves Salesforce: Discover is not connected to your Salesforce.
Under whose agreement: Our privacy policy.

Three statements that always go with Path B

Optional, authorised per user
Strategist is not active by default. Access requires a per-user authorisation against the Funnel Economics External Client App in your Salesforce. Your Salesforce administrator decides which profiles or permission sets may authorise it, and can revoke the tokens.
Data-processing agreement
The leg through our service requires a data-processing agreement, which we are preparing.
Our recommendation
The language model processes the data under your own agreement with the provider; we strongly recommend an enterprise agreement with the language model provider, including a data-processing agreement with that provider that excludes the use of the data for training. The exclusion has to come from that agreement, not from Funnel Economics.

What the Strategist server does — and does not do

Strategist reads only

Strategist's Salesforce access is read-only. Core writes only its own package objects, never standard objects such as Opportunities.

Stores no analyses

The service keeps no result store of its own.

Sharing through your own connections

A result reaches Slack, Microsoft Teams or email through your assistant's own connections, not through our service — under your identity, with a person confirming the recipient. This says nothing about the data path: the context passes through the Strategist server we operate to the language model you have chosen.

Model choice without lock-in

The language model is yours to choose. The technical precondition is an environment that can attach an MCP server. We strongly recommend an enterprise agreement with the language model provider, including a data-processing agreement with that provider that excludes the use of the data for training — a recommendation, not a precondition.

Controls and their status

Stated per component: in force, built but not yet in force, or deliberately absent.

Control	Core	Strategist server
Org-wide AI data-transfer switch, off by default	In force	Built, not yet in force
Per-user permission set “FE AI Data Transfer”	In force	Built, not yet in force
Per-user authorisation against the External Client App	—	In force — the effective control today
Masking of personal data before the language model	In force — by the package, before the callout	None — a decision, see below

The server-side check of the switch and the permission set is complete and waits for an addition to the Salesforce package. Until it is in force, the switch and the permission set do not stop Strategist. What stops it today: revoking the user's authorisation in Salesforce, or removing the connection in the assistant.

Core in detail

Requires Salesforce Enterprise Edition. No Data Cloud.
The calculation logic is compiled code in the package's protected namespace. Neither your users nor your administrators can inspect or modify it.
Core writes only its own package objects, never standard objects such as Opportunities.
AI data transfer is off by default and is enabled per org and per user.

Personal data, stated openly

Pipeline data contains the names of sales staff and opportunities.

Path A

The names stay in your tenant. For AI commentary, the package replaces them with placeholders before the callout and restores them inside your org.

Path B

The names reach the language model by design: evaluating results by person and by deal is the purpose — a lever without the deal it names cannot be pulled. Strategist does not mask or pseudonymise them. The protection sits in the agreement with the language model provider, not in the payload. Access is limited by the per-user authorisation.

One narrow exception

A session recognised as a sales rep does not receive its colleagues' names from one tool (conversion quartiles). That is one of eleven tools that carry rep names — not a general property.

Einstein Trust Layer

Applies to Core through Agentforce. It does not apply to Strategist.

Screenshots

Product screenshots on this website are anonymised under a fixed rule: people, opportunities and model names are replaced.

Compliance with GDPR or CCPA is a property of your configuration, roles and agreements — not something a product can guarantee on its own.

The leg through our service requires a data-processing agreement, which we are preparing. The AppExchange security review is planned after the beta phase.

Website and Discover

No tracking cookies. Privacy by default.
Sign-in to Discover is intended to work by one-time code sent by email — no password.
Three separate consents are intended; marketing emails only after a double opt-in confirmation.
Discover is meant to show a demo org with fictitious data only. It is not a Strategist trial and does not connect to your Salesforce.
For its AI features, Discover is intended to send the fictitious demo data and the figures you enter — no name, email address or company — to Anthropic as AI provider.

Questions from security reviews

Short answers, in the same wording as above. For anything beyond them, ask us directly.

Does our data leave Salesforce?

With Core alone, the analysis runs inside your tenant. Its AI commentary is off by default; when enabled, the package masks personal values before the callout leaves. With Strategist, yes: the context passes through the service we operate to the language model you have chosen. “Data never leaves Salesforce” is therefore not a statement we make about the product as a whole.

Is there a data-processing agreement for Strategist?

The leg through our service requires a data-processing agreement, which we are preparing.

Are personal data masked before they reach the language model?

In Core, yes: the package replaces personal values with placeholders before the callout and restores them inside your org. In Strategist, no: the names of sales staff and opportunities reach the model by design, because evaluating results by person and by deal is the purpose. This is a decision, not a gap.

Can our administrator switch Strategist off?

Per user, yes: Strategist requires a per-user authorisation in your Salesforce that your administrator controls and can revoke. The package's AI switch and permission set do not stop Strategist yet — the server-side check is built but not in force. A central switch-off in Funnel Economics does not exist yet.

Does Funnel Economics store our data?

The service keeps no result store of its own. Inside your org, Core's aggregates carry the names of sales staff and opportunities — a figure per rep is impossible without them.

Does the Einstein Trust Layer protect Strategist?

No. It applies to Core through Agentforce. Strategist has no additional layer; the model is governed by your own agreement with the language model provider.

Is Funnel Economics GDPR and CCPA compliant?

Compliance is a property of your configuration, roles and agreements — no product can guarantee it on its own. This page states per path which data go where, so that you can assess it for your setup.

Is Discover a test of Strategist with our data?

No. Discover is built for a demo org with fictitious data and does not connect to your Salesforce.

Ask About the Architecture

Questions from your security review, answered by the people who built it.

Ask About the Architecture

kontakt@funneleconomics.ai

Funnel Economics is entering its beta phase. During the beta, you start by running it against a Salesforce sandbox.

Request a place in the beta

One e-mail is enough. No form, no sign-up.