New WhatsApp API policies

WhatsApp recently announced an update to the WhatsApp Business Policy on how automation can be used to respond to users on WhatsApp. The updated policy now requires businesses that use automation to respond to user-initiated messages to provide a direct human escalation path within the conversation thread. Examples include handover to a human agent, a phone number, email, a web support form and in-store visits.

From 14 November, WhatsApp may reactively review bot experiences directly on the platform and, as a result, keep a record of messages between WhatsApp and the bot for quality monitoring purposes. During these initial reviews, WhatsApp will be transparent when it is engaging in a conversation with a bot. Businesses without a human escalation path may see limits on their ability to send messages through the API. And as WhatsApp continues to develop review processes, it may introduce more proactive review systems in the future.

FAQ:

Q: How will I know if WhatsApp is reviewing my bot?

A: When they start a conversation with the bot, WhatsApp will make it clear in the conversation thread that it is WhatsApp and that messages with the bot may be recorded for quality monitoring purposes. For example, the introductory message may say: “Hello, this message is from WhatsApp - we are starting this conversation to understand the quality of your bot experience. This conversation may be recorded for quality and testing purposes. Please do not share any sensitive or confidential information.”

Q: How will WhatsApp enforce this updated policy?

A: WhatsApp may notify businesses if it cannot detect a direct human escalation path. Businesses without a human escalation path may see limits on their ability to send messages through the API.

Q: Will I incur high costs for messages generated by this review?

A: No, the WhatsApp team will send a negligible number of messages during the review.

Q: What are acceptable direct human escalation methods if I use automation in user-initiated messages?

A: Examples of supported direct channels: handover to a human agent in the chat, phone number, email, web support, support form and in-store visits. Escalation channels to social media accounts (Facebook, Messenger, Twitter, Telegram, etc.) are not supported because they require the user to have an ID and an account on another service. In addition, indirect channels such as a help-centre web page or a link to an app do not meet this requirement.

Q: With this new user-initiated policy, can I still offer brand and user engagement experiences that encourage people to contact my business using user-initiated messages?

A: Businesses can use user-initiated messages for information that the business's human agents could also handle. For brand experiences, such as engaging with a story or business content, the information shared on WhatsApp must also be information for which users can get human support. Businesses that drive large volumes of users to connect with them on WhatsApp must keep this requirement in mind to manage expectations for any human agent follow-up with users.

The Sparrow team