How filters work in Tender Navigator
A filter in Tender Navigator is a saved tender search with an AI prompt and an email schedule. While it is active, Tender Navigator checks new tenders against its keywords, countries, classification codes and budget rule every day, evaluates each matching tender with AI against the prompt, and delivers the verdicts to the Results page and a digest email.
Last updated
A filter is a saved search that keeps running
The Filters page defines it in one sentence: a filter is a saved tender search that keeps running. Where the Pre-search tool answers one question once, a filter asks the same question of every new tender, every day, and adds a judgement: for each tender it picks up, an AI model reads the notice against the filter's prompt and records Matched or Unmatched with a written reasoning.
What a filter contains
| Part | Role |
|---|---|
| Name | How the filter appears in lists, on results and in email subjects. |
| Keywords, countries, CPV / NAICS / UNSPSC codes | The search criteria. They decide which tenders the filter picks up. Same semantics as the Pre-search tool: each list widens the search. |
| Deadline offset | Only include tenders whose deadline is at least this many days away. |
| Translate keywords | Optionally translate your keywords into the languages of the selected countries. |
| Tender budget | An optional range and a rule for tenders that publish no budget. See Filtering tenders by budget. |
| AI prompt and LLM model | What every picked-up tender is judged against, and which model does the judging. |
| Notifications & schedule | Email subject, daily or weekly frequency, day of week, and time. |
| Status | Active or paused. |
Full field-by-field guidance is in Creating and editing a filter.
What a filter produces
- Results. Every evaluated tender becomes a row on the Results page with the verdict, the reasoning, the tender's details and a snapshot of the filter as it was at evaluation time. See Understanding Matched and Unmatched results.
- Digest emails. On the filter's schedule, an email with the recent results. See How filter digest emails work.
- Usage counters. Evaluations performed and stuck tenders this month, shown on the filter and on the Filters page. See Filter usage and limits.
Active and paused
Only an active filter searches and evaluates. A paused filter is stored, editable and free, and does nothing else. New filters start paused whether you create them by hand, from a search or from the Business Profile. Each plan allows a number of active filters at a time; pausing one frees a slot. See Activating, pausing, duplicating and deleting filters.
What happens each day
- Tender Navigator takes the new tenders ingested since the last run.
- For each active filter, it keeps the tenders that satisfy the filter's keywords, countries, codes, deadline offset and budget rule.
- Each kept tender is sent to the filter's AI model with the filter's prompt. The model returns a verdict and a reasoning.
- The verdict is stored as a result and counted against the plan's evaluations for the cycle. When the cycle's evaluations, or the filter's own evaluation limit, are used up, further tenders are recorded as stuck and not evaluated.
- On the filter's schedule, the digest email goes out.
Filtering and evaluation happen daily on every plan; the schedule only decides when the email is sent.
Where to manage filters
Filters in the dashboard sidebar lists every filter, grouped as Active and Paused, with the number of evaluations performed, tenders stuck and the evaluation limit. From there you can open, activate, pause, duplicate or delete a filter, search by name and sort by status, name or evaluations. Filters is available only with an active plan.
Related articles
- Creating and editing a filterEvery field of the Tender Navigator filter editor explained: Basics, Search criteria, AI evaluation, Notifications & schedule, Usage, Tender budget, the setup checklist, and the three ways to start a filter.
- Writing the AI prompt for a filterHow to write the business criteria that Tender Navigator's AI evaluation judges every tender against: what to include, what to leave out, worked structure, and how to sharpen a prompt when verdicts look wrong.
- How AI tender evaluation works in Tender NavigatorHow Tender Navigator's AI evaluation judges each tender a filter picks up against the filter's prompt, what the model receives, what it returns (Matched or Unmatched with reasoning), and when it runs.
- How filter digest emails workThe digest email each Tender Navigator filter sends: what it contains, the Daily or Weekly frequency, day of week and time (UTC), the custom subject line, where the links lead, and why the schedule does not change when tenders are evaluated.
- Filter usage, stuck tenders and evaluation limitsWhat the usage counters on a Tender Navigator filter mean: Evaluations performed, Stuck tenders, the per-filter Evaluations limit, and how they relate to the plan's evaluations per billing cycle.