Trust & safety · Updated June 2026
Is it safe to let an AI negotiate my bills?
Yes — when you stay in control of what the agent can actually do. Dip can ask your provider for a lower price; it can’t move money, change your payment method, or commit anything without your approval. You authorize it once, every call is recorded so you can hear what was said, and you approve every change before it takes effect. The thing people worry about — an agent doing something irreversible on its own — is designed out.
Six ways Dip keeps you in control
It can ask, not act
Dip can request a lower price. It can't move money, change autopay, or alter your payment method — there's no path for it to do something financial behind your back.
You authorize once, revoke anytime
A one-time letter of authorization names the accounts Dip may act on. It's the same instrument a human negotiator uses, and you can withdraw it whenever you want.
You approve every change
Nothing commits at the provider until you see the offer and tap approve. Don't like it? Your old rate stays, full stop.
Every call is recorded
You can listen back to exactly what was said on your behalf. No black boxes — if Dip speaks to your provider as you, you can verify it.
Read-only bill data
Dip reads statements and recurring charges (via Plaid, read-only). It never gets your bank login or debit authority.
Honest AI disclosure
Dip identifies itself as an AI agent acting on your behalf and answers honestly if asked — transparency with the rep, not deception.
Frequently asked
Is it safe to let an AI negotiate my bills?
It's safe when you stay in control of what the agent can actually do. With Dip, the agent can request a lower price on your behalf — it cannot move money, change your payment method, or commit a change without your approval. It reads your bill details (and recurring charges if you connect Plaid) but never gets your bank login or debit authority. Every call is recorded and transcribed so you can hear exactly what was said, and every change is shown to you for approval before it takes effect at the provider. The risk you're guarding against — an agent doing something irreversible behind your back — is designed out: nothing commits without your tap.
Is it legal to have someone (or an AI) call my provider on my behalf?
Authorizing a third party to contact your provider and negotiate is a long-standing, ordinary arrangement — it's how human bill-negotiation services and account representatives have always worked. Dip uses the same instrument: a letter of authorization you sign once that names which accounts it may act on and what it's allowed to do, and that you can revoke anytime. On the call, Dip identifies itself as an agent acting on your behalf rather than impersonating you. (This page describes how Dip is designed to work, not legal advice — your provider's terms govern your account.)
Can the AI access my bank account or move money?
No. Dip reads bill information — statements, account details, and recurring charges if you choose to connect Plaid (a read-only connection). It never receives your bank credentials and has no authority to move money, change autopay, or alter your payment method. The most it can do is ask a provider for a lower rate; you approve any resulting change before it commits.
Will the company know they're talking to an AI?
Dip is built to disclose that it's an AI assistant acting on your behalf, and to answer honestly if asked. It's designed to be transparent with the rep, not to deceive them — that's both the right thing to do and the durable way to operate as the rules around AI disclosure tighten.
What if I don't like the deal the AI negotiates?
Nothing happens until you approve it. Dip brings back the offer — the new rate, the terms, what changes — and you decide. If you pass, your existing rate stays exactly as it was. You're never committed to an outcome you didn't choose, which is the core of why handing off the call is low-risk.
Can I hear what the AI actually said on my behalf?
Yes. Every call Dip places is recorded and transcribed, and you can listen back. That's a deliberate design choice — "no black boxes." If an agent is going to speak to your provider as you, you should be able to verify exactly what it said.