Home · Business · Customer Service · Tier-1 support & chatbots
As of 13 August 2026, AI can only partly help customers reset their passwords.
This still needs a person who signs their name to it.
Can you do it?
15 minutesto a draft.
1 hourto something you’d act on.
Cost, all in£0/month
Skill neededpower-user
Who has to check ita colleague
What the alternative costsA human support team remains the alternative for account-specific cases; the supplied tool data gives no price for that service.
If this goes wrong: a customer is sent through an unsafe or unusable flow, or an unauthorised person gains access to an account.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Use a tool built for this
Second choiceDo it yourself
The distant thirdA chat interface, power-user skill, and roughly 1 hour until you can act on the result.
How to actually do it
- Open your current account-help page and identity-provider documentation, then write down the exact reset-link process, expiry rules and approved escalation route.
- Remove passwords, authentication codes, payment details and other secrets from the material before pasting the process and policy into the prompt.
- Paste the process, security policy, escalation details and approved wording into the prompt, then ask the AI to produce the customer flow and test cases.
- Compare every instruction in the draft against the live reset journey, including the sender address, link destination, expiry behaviour and hand-off route.
- Build the conversation in a support platform such as Botpress, connect only the approved account and escalation integrations, and prevent the bot from collecting secrets.
- Run the listed test cases with test accounts, including an expired link, an inaccessible email address, a locked account and a suspected account takeover, then send failures to a colleague responsible for security before release.
Prompt
Design a safe tier-1 customer-support flow for password resets for [BUSINESS TYPE]. Use only the policy and process details I provide below. Do not ask customers for passwords, full payment-card details, security answers, authentication codes or other secrets. Do not claim to verify identity yourself. Route identity verification and the actual password change through the existing secure account system. Include: 1) a short opening message, 2) the non-sensitive information the customer may provide, 3) numbered instructions for the current reset process, 4) messages for expired links, missing emails, locked accounts, suspected compromise and customers who cannot access their registered email, 5) the exact point at which to hand over to a human, 6) a compact list of data the chatbot must not collect, and 7) test cases for a colleague to run. Mark every item that depends on an integration or a business policy. Invent nothing and leave [UNKNOWN] where information is missing. Current reset process: [PASTE PROCESS] Identity and security policy: [PASTE POLICY] Human escalation route and hours: [PASTE ESCALATION DETAILS] Approved customer-facing wording or help-centre content: [PASTE CONTENT]
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
What it gets wrong
- The AI cannot identify a customer or change an account without a secure connection to your identity system.
- It cannot safely replace identity verification, especially when the customer has lost access to their registered email or suspects compromise.
- It cannot know whether your live reset links, policies and integrations still work unless someone tests them against the real system.
- It cannot carry the security and data-protection responsibility for an incorrect reset flow.
- It cannot resolve unusual ownership disputes or decide whether an exception to the normal process is safe.
What caps this at PARTLY: private data access, consent and privacy and verification cost.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 1 |
| Verification | 1 |
| Liability | 1 |
| Effort delta | 1 |
| Total | 6 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can a chatbot reset a customer's password?
- It can guide the customer to your secure reset process and may start that process through an approved integration. It should not ask for the customer's password or authentication code, and a human or secure identity system must handle exceptions.
- Can AI verify a customer's identity for a password reset?
- Not safely from the conversation alone. Use your existing identity and authentication controls, and hand over cases where the customer cannot use them or reports suspected account compromise.
- What should a password-reset chatbot ask for?
- Ask only for the minimum non-sensitive information needed to direct the customer to the right help, such as the product area or whether they can access their registered email. Never ask for passwords, authentication codes, full payment-card details or security answers.
- Is it safe to use AI for password resets?
- It can be safe for instructions and routing when it is connected to a properly secured reset flow and blocked from collecting secrets. You must test the live journey and keep a human route for failed verification, locked accounts and suspected takeover.
Nearby answers
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.