Documentation

AI Gateway · Guides

Use cases

Common ways teams use the intSignal AI Gateway — safe customer-support assistants, developer copilots on real data, document capture into a CRM (scan-to-text, scan-to-form), and a compliant proxy in front of ChatGPT or Copilot.

The Gateway exists so you can point AI at the data that actually runs your business — tickets, emails, contracts, records — without that data leaking. Here are the patterns teams adopt most often; each one is the same private AI layer for regulated business data pointed at a different workflow.

Customer-support assistants

Support tools are full of PII: names, account numbers, card details, health notes. Put the Gateway between your assistant and the model so agents get AI summaries, drafted replies, and next-best-actions while the customer's raw data is redacted before it's sent. Reversible tokens keep the answer specific to the right customer.

Copilots on real business data

Internal copilots are only useful if they can read real records. The Gateway lets a copilot query orders, HR data, or code while stripping secrets, credentials, and regulated fields — so a helpful answer never comes at the cost of a leak.

Document capture — go paperless

Point the Gateway at a scan, PDF, or photo and it turns paper into structured data.

Scan to text (OCR)

The Gateway reads the document and extracts clean, searchable text from scans, PDFs, and photos.

Scan to form

It maps the extracted values into fields — company, amount, dates, line items — ready for a CRM record or a form.

Redact on capture

Sensitive fields are masked or redacted as the document is ingested, so nothing sensitive is stored or sent onward in the clear.

File it automatically

The structured result is pushed into your CRM, ticketing, or ERP — intake that used to be manual becomes automatic and paperless.

A compliant proxy in front of existing AI

Teams already using ChatGPT Enterprise or Microsoft Copilot can keep those tools and route them through the Gateway. Every prompt is sanitized and every response checked, with a central audit trail — you get governed AI without changing the tools people like. See Integrations.

Building AI into a product

Software teams embed the Gateway so their own product ships privacy-safe AI features. Developers get an SDK and REST API; integrators run a managed, multi-tenant proxy for the clients they build for — the same protection whether you're one team or a hundred tenants.

By industry

IndustryWhat the Gateway protectsFrameworks
HealthcarePatient records, clinical notes (PHI)HIPAA
Finance & retailCard and account data (PAN)PCI DSS
Legal & professional servicesContracts, client PIIGDPR, SOC 2
Any regulated businessNames, SSNs, secrets, credentialsGDPR, SOC 2

Tip

Not sure which entities appear in your data? Run a project in Automatic policy setting for a day — the Gateway reports what sensitive types it saw and proposes rules to cover them.

Need a hand with AI Gateway?Talk to our team →