Ship spam detection faster: Three guided steps with Illuminate
Spam kills momentum.
A few actors flooding a channel or blasting the same line across rooms can drown real conversation and open the door to scams, shady links, and fake “support.” If you do not catch it early, it can degrade the chat experience, engagement drops, and your users take the hit.
Some teams reach for AI first. For classic patterns like flooding and cross-posting, behavioral rules often get you there sooner, with clearer thresholds and less guesswork.
In our previous posts, we showed how to spot flooding and cross-posting and how to configure PubNub Illuminate to respond automatically. That path worked, but it took more configuration than most product and development teams want to own day to day.
Illuminate now simplifies this process in as little as three simple steps.
Our new Query Builder starts with predefined spam detection use cases that automatically configure most of the setup for you. Instead of building everything manually, you simply select the use case, choose your data source, and adjust a few values.
Here's how it works.
Step 1: Select a predefined spam detection use case
Start by opening Query Builder in Illuminate and selecting one of the predefined spam detection options.
Two common chat moderation use cases are available:
- Cross-Posting Spam: Detect users sending duplicate messages across multiple channels.
- Chat Flooding Spam: Detect users posting excessive or repetitive messages within a single channel.
After selecting a use case, choose the Business Object that contains your chat data.
Illuminate automatically selects the relevant fields from the data source. These typically include the user identifier, the channel or room, and the message text. You can even make sure that the message text has a minimum of 10 characters to reduce false positives, like Hello, Welcome to the team, etc.
Because these fields are automatically mapped when the Business Object was created, you don’t need to configure them manually.
Step 2: Configure detection thresholds
Next, configure the values that determine when activity should be considered spam.
For example, you might detect:
- users sending more than 10 messages within one minute in a single channel
- users posting the same message across multiple channels within a short time window
You can also configure the aggregation window, such as the last minute or the last hour. Once the values match the behavior you want to detect, save the configuration.
At this point, Illuminate can identify suspicious activity based on the thresholds you defined.
Step 3: Activate the decision to automate actions
After saving the configuration, you can create a Decision directly from it.
When you choose the predefined decision option, Illuminate automatically creates a decision table with rules and actions already configured.
You only need to adjust the rule values and activate the decision.
Typical actions include:
- notifying moderators
- flagging suspicious users
- muting users temporarily
- banning repeat offenders
Once activated, the decision continuously monitors incoming chat activity and triggers actions whenever the configured thresholds are reached.
Respond to spam faster
Automated spam detection keeps your chat healthy without standing up a complex bespoke moderation stack.
With Illuminate’s predefined spam paths, you can spot flooding and cross-posting fast and act automatically when those patterns show up.
Thresholds stay yours to tune as your chat experience evolves—so product and engineering can iterate without a rewrite every time traffic shifts.
Try it out now! Start a free two-week trial: open Illuminate in the PubNub Admin Portal and explore.
