Categories and labels give your team a shared way to describe what each piece of knowledge is about - whether it lives in the Bid Library as a document, or in the Q&A Store as a saved answer.
When you generate a draft, BidScript can use those labels, alongside the libraries you select, to ground the draft in the right company knowledge and the most similar past answers - rather than relying on search alone.
Think of it as a lightweight taxonomy. Categories are the buckets - Product line, Market segment, Compliance. Labels are the specific tags inside each bucket - Enterprise hardware, Retail, ISO 9001.
Two scopes
Global | Team-wide categories and labels, not tied to one library. Use for standards many libraries share - company-wide product families, or quality certifications appearing across tender types. |
Library | Categories and labels belonging to one Bid Library. Use where tagging is only about that library's content - a library for a single product range with its own sub-categories. |
When you choose grounding libraries for a draft, library-scoped labels for libraries you did not select are hidden from the picker, so what you see stays aligned with what you are searching.
Create categories and labels
Administrators set the taxonomy up.
Open the Bid Library and select the library to configure.
In library settings, open Manage labels.
Create your categories first, naming them for how you want content grouped - Product line, Target market, Technical standard.
Add labels under each category. These are the tags you will assign to documents and Q&A items - On-premise software, SME, CE marking.
Apply labels to content
Documents and folders
Open Manage labels on the document or folder.
Check labels to add them, uncheck to remove.
Labels inherited from a folder may appear as read-only context. Only the ones you toggled directly on the item are its own.
You can bulk-assign and bulk-remove labels across a multiple selection - useful when onboarding a new product range and tagging all its documents at once.
Q&A Store items
Q&A items use the same taxonomy. Tag each entry as you would a library document, so draft generation can favour answers matching the product or market segment you care about.
A saved answer about your software's uptime SLA, for example, should carry the relevant product line and market labels so it surfaces when drafting in that space.
Use labels when generating a draft
When you start draft generation from the questions workflow, you configure how BidScript should retrieve knowledge.
Grounding labels
Optional. Leave labels empty and retrieval is not limited to specific tags.
Expand Grounding Labels to see categories grouped under Global or the library name.
Use Search to filter by scope, library, category or label name.
Use Select all or Clear all on a category to toggle every label in it.
A counter shows how many labels are selected.
Matching multiple labels: Any or All
When you select more than one label, you control how they combine. This applies anywhere labels are used to ground an AI service, including draft generation.
Any | Content matching either label is eligible. Selecting "Retail" and "ISO 9001" returns content tagged with one or the other, or both. |
All | A strict search. Content must carry every label you selected. Selecting "Retail" and "ISO 9001" returns only content tagged with both. |
Any broadens; All narrows, often sharply. Start on Any and switch to All when you need a tightly defined intersection - and expect fewer results, because content has to have been tagged thoroughly for All to find it at all.
Include untagged shared content
On | Even with specific labels selected, unlabelled content can still be used where relevant. |
Off | Strict label-only grounding. Untagged items are excluded entirely. |
Turn it off when a draft should reflect a tightly defined slice of your knowledge - limiting it to a single product line without pulling in other ranges.
Why this affects draft quality
Library labels help BidScript find the right specifications, compliance documents and reusable commercial prose. Q&A labels surface past answers from the same product areas or market segments.
Together they give draft generation a clearer map of what to trust - which is what reduces generic answers.
Quick reference
Category | A named group of labels, such as product line or target market. |
Label | A specific tag within a category, assigned to library documents and Q&A items. |
Global vs library scope | Team-wide taxonomy, versus per-library taxonomy. |
Grounding libraries | Which Bid Libraries to search when generating. |
Grounding labels | Optional filter so retrieval favours tagged content. |
Any / All | Whether content must match one of your selected labels, or every one. |
Include untagged shared content | Whether unlabelled items can still be used when labels are selected. |
Rolling this out
Agree a small set of categories first. Three to eight is usually enough — Product line, Target market, Technical standard and Document type is a good starting point for product businesses.
Use labels that reflect how your team actually talks about products in tenders: model names, product families, sectors you sell into.
Tag your high-value content first — product datasheets, case studies, specification responses — before relying on strict grounding.
On a first pass, leave Include untagged shared content on and matching set to Any, so older untagged material stays visible while you build the taxonomy out.
Related articles
Generating drafts — where grounding labels are applied.
BidBot in the Answer Workspace — selecting the libraries labels sit within.
Smart Search — finding content by meaning rather than by tag.
Creating & Managing Libraries — the libraries labels are scoped to.