Skip to content

Format chatbot answers for scannable mobile reading and quick replies

Stop burying answers inside 350-word desktop essays that overflow a 380px chat widget—lead with the bottom line and layer progressive detail.

Copy the prompt and paste it into the AI assistant you already use.

Formatting rules for small-viewport chat widgets

  1. 01

    Put the direct answer in sentence one

    Answer yes/no, exact price, or root cause in the first 20 words before adding supporting conditions or setup instructions.

  2. 02

    Cap replies at 90–130 words by default

    Inside a 380px widget, 120 words fills an entire screen; summarize concisely and ask if the user wants deeper edge-case details.

  3. 03

    Use max 4 short bullets, never wide tables

    Markdown tables wrap horribly inside narrow chat bubbles—use compact bolded bullets (`• **Growth:** 50 seats`) instead.

  4. 04

    Adapt Markdown syntax per target channel

    Convert links, bold markers, and lists cleanly to channel-native formatting for WhatsApp (`*bold*`), Slack (`mrkdwn`), and plain SMS.

Questions and answers

How do we enforce word-count limits when LLMs naturally write verbose prose?

Combine explicit word caps ('under 110 words unless code is requested') with 3 ultra-concise few-shot examples in the system prompt.

What if a developer explicitly asks for a complete code snippet or curl example?

Carve out an explicit exception allowing complete, runnable code blocks while keeping surrounding prose explanation under three sentences.

Should chatbots split long replies into three consecutive rapid-fire chat bubbles?

At most split an answer from a follow-up question bubble; firing four rapid notifications buzzes phones excessively and breaks screen-reader flow.

Learn more

Related guides and background articles for this task