How do I create a hot content bot?

Also available in:Français

A hot content bot distributes articles that are still recent but do not require the cadence of a breaking-news alert. Give it a dedicated theme, a short freshness window, and a steady publishing rhythm.

1. Prepare a hot content theme#

In Intelligence, create or duplicate a theme dedicated to recent content.

  1. Select the section containing your publishable articles.
  2. Filter for articles published in the last 24 hours. If your volume is low, widen the period gradually.
  3. Exclude the tag reserved for breaking news, for example with NOT tag:"breaking news", so the same article cannot be selected by both bots.
  4. Sort results according to your editorial priority: publication date for freshness, views for performance, or a saved editorial order.
  5. Review about ten results manually. Every result should be safe to republish without correction.

For more precise filters, see How do I create a Boolean query?.

2. Create the bot from this theme#

In Bots, select Create a bot, choose From a theme, then Publish to your social accounts. Give the bot a clear name, select the brand, hot content theme, theme section, and only the relevant social accounts. The bot form shows the theme's name, but not its filter settings: check in Intelligence that the freshness window matches the period you chose in step 1. Then set the bot's timing as shown below.

Nonli bot schedule showing every day from 06:00 to 23:00 in Europe/Paris, consultation every 50 to 70 minutes, and immediate publishing

A practical starting point for a news feed is:

  • active every day from 06:00 to 23:00 in the newsroom timezone;
  • consult the theme every 50 to 70 minutes;
  • publish immediately after the bot consults the theme.

These values are starting points, not universal rules. Run the bot more often when the theme contains many priority articles, or less often when the destination accounts are already busy.

3. Configure content and safeguards#

Use these content sources as a starting point for posts in any language:

  • Message's source: Template: [public] Message from first standfirst sentence, so the post starts with the first sentence of the article's standfirst;
  • Title's source: Title;
  • Description's source: No description;
  • Image's source: Image, the article's main image;
  • Comment's source: Template: [public] 1st comment Optim Clic Multilingue for the first comment. It uses the same shortlink rule across languages and adapts the Instagram link-in-bio prompt to the article language;
  • Main image alt text's source: Template: [public] Alt text.

Nonli bot content-source settings with the truncated standfirst message template name, title, no description, main image, multilingual Optim Clic first comment, and Alt text template

The form shortens the message template name in the dropdown. Its full name is Template: [public] Message from first standfirst sentence, as listed above.

The two templates coordinate link placement. Nonli replaces the [SHORTLINK] marker with the article's shortened URL when the post is published:

  • Shortlink in the first comment: on an eligible Facebook or LinkedIn photo post with an article link, 1st comment Optim Clic Multilingue may put [SHORTLINK] in the comment. The message template detects it with hasFirstCommentShortlink() and does not repeat the URL in the main post.
  • No shortlink in the first comment: when the post has a link and nonempty opening text, and its format and network allow a URL in the message, Message from first standfirst sentence adds the shortlink after that text. The URL appears only once.
  • Native link post or article without a URL: a link post uses its native link; without an associated URL, neither template adds [SHORTLINK]. On Instagram, the comment template points readers to the link in bio in the article language when available, with English as a fallback.

Optim Clic Multilingue does not place a shortlink in the first comment on every post. For eligible photo posts, it uses shouldAddComment('click'), a click-oriented comment sampling rule, to decide whether the link goes in the comment. Check previews with and without a comment, then compare clicks measured in Nonli; this setting does not guarantee more clicks. For more detail, see how Smart Messages work and how first comments are published.

3.2. How does automatic click optimization work?#

Under Post type selection mode, choose Auto - Photo/Link only, then set Performance metric to Click. The bot can then choose between photo and link posts, favoring the format that performs better on clicks. This is an automatic choice rather than a fixed photo/link ratio; it does not guarantee more clicks for every post.

This setting chooses the post format. Optim Clic Multilingue separately decides whether an eligible photo post puts its shortlink in the first comment. The chosen format and comment determine where the link appears as described above.

Nonli bot settings with Auto - Photo/Link only selected as the post type mode and Click selected as the performance metric

The message template must have a usable standfirst to draw from. Review a few previews and adapt the source if your articles do not have one. Set a global daily limit and a per-account limit, then start with the review queue enabled so you can inspect the first results before switching to automatic publishing.

Pay particular attention to these settings:

  • Maximum number of reposts: keep 0 when each article should be used only once;
  • Daily limits: prevent an editorial spike from filling the entire social calendar;
  • Excluded terms: remove sensitive topics, finished live coverage, or incompatible formats;
  • Post type: review the photo/link mix and clicks after the bot has published enough posts;
  • PublishGuard: enable it after the brand rules have been configured and tested.

4. Test before enabling automation#

  1. Keep the bot in the review queue for at least one complete editorial cycle.
  2. Check freshness, image, copy, link, destination account, duplicates, and that each run proposes no more than one article. If it proposes more, keep the review queue active and adjust the bot before enabling automatic publishing.
  3. Compare the number of eligible articles with the selected daily limit.
  4. Switch to automatic publishing only when results are stable.
  5. After a few days, adjust the freshness window and frequency using the actual published volume.

If an expected article is not published, follow Why did my bot not publish this article?.

Was this article helpful?

Book a personalized demo with our team.