Trust · Responsible AI

TextGopher™ prepares. You decide.

Last updated: July 2026

TextGopher™ uses AI to understand a caller and organize the request. It never speaks for your business without guardrails, and it never makes the final call. A person at your business reviews and decides every outcome.

How we use AI

When a business misses a call, TextGopher™ texts the caller back and holds a short, natural conversation to understand what they need: the problem, how urgent it is, and when works best. The AI reads those replies and turns them into a structured opportunity you can act on. That is the whole job: understand and organize. It is not a chatbot pretending to be your business, and it is not making commitments on your behalf.

Human in the loop

Every outcome is reviewed and decided by a person at your business. The AI recommends a next action (call now, confirm a time) but it doesn’t take it. You can take over the conversation at any point and text the customer yourself, and you decide whether an opportunity is worth pursuing, editing, or closing.

Guardrails on every reply

Before any AI reply is sent, it passes through deterministic guards that can block or rewrite it. These are compliance controls, not suggestions — and they fail closed: if a guard can’t confirm a reply is safe, the reply doesn’t go out.

• The Commitment Firewall. Blocks any reply that would state a price, promise a time, confirm a booking, or name a specific technician. TextGopher™ never commits your business.
• Price & fact guards. A dollar figure the customer didn’t state, or a detail the system can’t back up, is blocked or scrubbed rather than sent.
• Anti-stall guard. A false “let me look that up” is replaced with an honest hand-off to a person.

Fair Housing

For the real-estate vertical, a Fair Housing content filter screens every reply for steering, discriminatory preference, or unlawful language. When it detects a serious issue it hands the conversation to a human agent instead of sending anything; lesser issues are redacted. This filter stays on for real estate — it can’t be switched off.

Safe by default

If the AI model is slow, errors, or returns anything the guards reject, TextGopher™ falls back to a plain, deterministic message or an honest hand-off. A caller always gets a safe, sensible reply — the AI layer never fails toward the customer.

What we record

Every decision a guard makes — every block, redaction, and hand-off — is written to an append-only ledger that can’t be edited or deleted. That gives us a durable record of how the controls behaved. We log control decisions and identifiers, not raw personal content, in keeping with our security practices.

Limits & your controls

AI can misread a message — which is exactly why a person decides every outcome. You approve the questions and rules TextGopher™ uses, you can take over any conversation, and you can edit or close any opportunity. TextGopher™ does not provide legal, financial, medical, or other professional advice.

Related: Security · Acceptable Use · Privacy