Settings and personalization
Manage GovTribe AI personalization settings, memories, full conversation recall, GovTribe data access, verbosity, reasoning effort, and custom instructions.
GovTribe AI personalization controls determine what user context the assistant can use and how it should shape responses. Open the GovTribe AI settings menu to manage personalization, custom instructions, connector settings when enabled, and the Memories page when memories are available for your account.
Open GovTribe AI settings
Use the settings menu in GovTribe AI to open these controls.
| Menu item | What it opens |
|---|---|
| Settings | The main GovTribe AI settings menu. |
| Personalization | Controls for memories, conversation recall, GovTribe data access, verbosity, and reasoning effort. |
| Custom Instructions | Fields for stable information about you and instructions for how GovTribe AI should respond. |
| Memories | The page where saved AI memories can be reviewed and deleted when memories are enabled. |
| Connectors | Opens connector settings when external connector access is enabled for your account. See Manage connected tools. |
Manage personalization
The Personalization drawer controls how GovTribe AI can use durable context.
| Setting | What it does |
|---|---|
| Memories | Allows GovTribe AI to save and reuse durable preferences, facts, and work context. Turning this off deletes previously saved memories and prevents new memories from being created. |
| Full Conversation Recall | Allows GovTribe AI to search prior AI conversations when useful. New conversations do not automatically include old conversation snippets. |
| Access GovTribe Data | Allows GovTribe AI to use permitted GovTribe activity and profile context for personalization. |
| Verbosity | Controls the default response length and level of detail. Choices are Low, Medium, and High. |
| Reasoning Effort | Controls the default reasoning effort for supported AI work. Choices are Low, Medium, and High. |
Use lower verbosity when you want shorter answers. Use higher verbosity when you want fuller explanations, more structure, or more complete research notes. Use higher reasoning effort for complex analysis and lower reasoning effort for simpler requests where speed is more important.
Use custom instructions
Custom instructions are stable guidance GovTribe AI can include in each conversation.
| Field | What to include |
|---|---|
| About You | Stable facts about your role, company, pursuits, agencies, vendors, terminology, or responsibilities that GovTribe AI should know. |
| Response Style | Preferences for tone, length, structure, formatting, language, or working style. |
Each custom-instruction field is limited to 1,000 characters. GovTribe AI follows your response-style instructions unless they conflict with system behavior, safety requirements, or the specific prompt you send.
Example About You instructions
Use About You for stable context GovTribe AI should remember across conversations. Good instructions describe your role, market focus, company capabilities, preferred record types, pursuit posture, and recurring constraints.
Use the copy button on any example, then adjust the wording for your company, markets, and responsibilities.
Capture manager
I manage capture for a mid-sized federal IT services firm. We focus on cloud migration, cybersecurity, help desk, and data modernization work for civilian agencies. Prioritize recompetes, incumbent context, contract vehicles, set-aside posture, likely competitors, and near-term next actions.Business development lead
I lead business development for a small business contractor with strong past performance in professional services and program support. We prefer prime opportunities under $50M and teaming opportunities on larger IDIQs. Highlight agency fit, relationship signals, and partner angles.Proposal manager
I support proposal planning and compliance. When I ask about an opportunity, help me find evaluation factors, instructions, submission deadlines, required attachments, compliance risks, and draft outline sections that map to the solicitation language.Pricing analyst
I work on pricing strategy. I care about labor categories, place of performance, wage geography, ceiling rates, awarded line-item evidence, incumbent pricing clues, and assumptions that should be reviewed before leadership sees the estimate.Agency account lead
I manage our account strategy for selected federal buyers. Prioritize buying patterns, expiring awards, active vehicles, key offices, likely recompetes, and how new opportunities relate to our existing agency plan.State and local seller
I focus on state and local technology opportunities. Give extra attention to jurisdiction context, buyer history, related awards, line-item pricing evidence, incumbent vendors, and whether the opportunity looks realistic for a vendor like us.Do not use custom instructions for one-time details that only apply to the current prompt, such as a single opportunity, a temporary deadline, or a specific document you only need once. Put those details in the conversation prompt or project instead.
Example Response Style instructions
Use Response Style for how GovTribe AI should structure answers.
Executive summary first
Start with the recommendation and the top three reasons. Put supporting evidence in a table after the summary.Capture scoring
When reviewing opportunities, organize the answer by Fit, Competition, Timing, Vehicle or Set-Aside, Risks, and Recommended Next Action.Evidence discipline
Separate confirmed GovTribe evidence from assumptions. If evidence is missing, say what is missing and how to verify it.Proposal-ready output
Use proposal and capture language. Avoid generic sales copy. When drafting win themes, tie each theme to a customer need, proof point, and likely evaluation factor.Team handoff
End with action items for BD, capture, proposal, and pricing teams when those roles are relevant.Concise research notes
Use short bullets and tables. Avoid long narrative unless I ask for a memo or briefing.How custom instructions shape responses
Custom instructions work best when they name the structure you want GovTribe AI to use repeatedly. These examples show how a reusable instruction can change the shape of future answers.
Capture decision reviews
Instruction:
When reviewing opportunities, start with a pursue / watch / no-bid recommendation, then show Fit, Timing, Competition, Vehicle or Set-Aside, Risks, and Next Action.Example response shape:
Recommendation: Watch
Top reasons:
- Fit is strong for cloud and help desk work.
- Timing is early, with no solicitation package yet.
- Competition risk is unclear because incumbent evidence is incomplete.
Evidence table:
| Area | What GovTribe found | What it means |
| Fit | The requirement includes cloud operations support. | Strong alignment with our delivery history. |
| Timing | The notice is pre-solicitation. | Start capture research, but wait for full requirements before a bid decision. |
| Competition | Incumbent evidence is incomplete. | Confirm predecessor award and likely competitors. |Evidence-first research
Instruction:
Separate confirmed GovTribe evidence from assumptions. If the record does not support a claim, put it under Open questions instead of stating it as fact.Example response shape:
Confirmed evidence:
- The buyer is the Department of Veterans Affairs.
- The notice is tied to facilities support.
- The due date is listed in the opportunity record.
Assumptions:
- This may connect to prior sustainment work, but the record does not confirm the incumbent.
Open questions:
- Confirm whether the current incumbent is still performing the predecessor work.
- Check whether the buyer has released attachments that change the scope or evaluation criteria.Proposal team handoffs
Instruction:
When I ask about a solicitation, organize the answer for a proposal team: compliance items, evaluation factors, submission controls, pricing items, attachments, risks, and writer actions.Example response shape:
Proposal handoff:
| Area | Action | Owner |
| Compliance | Confirm page limits and required forms. | Proposal manager |
| Pricing | Check labor categories and place of performance. | Pricing lead |
| Technical | Map requirements to draft sections. | Technical lead |
| Attachments | Review all amendments and file changes. | Proposal coordinator |
Risks:
- Submission instructions may differ between the notice text and attached files.
- Pricing assumptions need review before leadership approval.Review memories
When memories are enabled, you can ask GovTribe AI to save stable facts, preferences, and long-term working context for future conversations. Use direct wording such as remember that..., save this as a memory..., or do not forget that....
Good memory requests include:
Remember that I work for XYZ Corp.Remember that our strongest past performance is in federal cybersecurity, cloud migration, and managed help desk work.Remember that I prefer opportunity reviews to include incumbent, likely competitors, set-aside posture, timeline risk, and a bid / no-bid recommendation.Remember that we usually avoid opportunities under $5M unless they are strategic agency entry points.GovTribe AI can also update or remove memories when they are no longer accurate:
Update my memory: I now work for ABC GovTech, not XYZ Corp.Forget the memory about avoiding opportunities under $5M.Memories are best for durable context you want GovTribe AI to reuse across future work. Do not use memories for one-time details that only apply to the current prompt, temporary deadlines, sensitive personal details, or information you would not want reused later.
The Memories menu item opens the GovTribe AI Memories page. The page lists saved memories with their created date and an action to delete individual memories. Use the Memories page to remove saved context that is outdated, no longer useful, or should not affect future AI responses.
Use uploaded user files
GovTribe AI can work with User Files that are available to your workspace and visible to your account. These can include files your team uploaded, files associated with pursuits, and files generated or saved from prior GovTribe AI work.
This lets you reuse past work instead of starting every prompt from scratch. For example, you can ask GovTribe AI to find a past capability statement, proposal outline, debrief, pricing note, market analysis, or pursuit document, then adapt that material for a current opportunity.
| File source | How GovTribe AI can use it |
|---|---|
| Uploaded workspace files | Search file names, descriptions, summaries, and extracted text when available. |
| Pursuit-linked files | Bring prior pursuit context into a new opportunity review, bid / no-bid discussion, or proposal workflow. |
| Generated AI files | Reuse documents, spreadsheets, charts, or summaries created in earlier GovTribe AI conversations. |
| Saved research files | Compare older market research, customer notes, or capture material against a new record. |
Strong prompts include:
Find our past capability statements for cybersecurity work and adapt the strongest language for this opportunity.Use prior pursuit documents about this agency to identify likely win themes for the current solicitation.Compare this opportunity against our uploaded past-performance files and flag the best matching examples.Find the proposal outline we created last quarter and update it for this new requirement.User-file access follows GovTribe workspace permissions. If a file is in another workspace, hidden from user-file discovery, deleted, not indexed, or not accessible to your account, GovTribe AI may not be able to find or use it. For direct file management, use User files.
Upload company assets for AI reuse
Upload the asset. Add the file from User Files, or save a generated GovTribe AI file to the workspace when that option is available.
Add useful context. Use clear file names, descriptions, tags, pursuit links, or project context so the asset is easier to find later.
Ask GovTribe AI to use it. Name the file type, topic, pursuit, agency, or tag in the prompt, then ask for the specific output you need.
Review the source. Check the uploaded file and the generated answer before reusing language in a proposal, customer communication, or internal decision.
For the file-management side of this workflow, see User files. For capability boundaries and verification guidance, see GovTribe AI capabilities and limits.
How personalization affects GovTribe AI
Personalization can affect multiple parts of GovTribe AI.
| Area | Possible effect |
|---|---|
| New conversations | GovTribe AI can use allowed memories, custom instructions, user files, and GovTribe activity context to shape responses. When Full Conversation Recall is enabled, it can search prior AI conversations on demand when useful. |
| Recommended prompts | Starter prompt cards can be generated from user and workspace context. |
| Daily Briefing | The briefing depends on opt-in personalization and can use allowed user, workspace, capture, and opportunity context. |
| Automations | Automation responses can reflect the selected prompt, skill, connected-tool configuration, project context, relevant user files, and available personalization settings. |
Personalization does not remove the need to verify important information. Review AI output against the underlying GovTribe records before acting on critical details.
Related articles
- Create a conversation: Start chats that use your personalization settings.
- Manage connected tools: Connect official tools, control chat tool selection, and request a custom connector.
- User files: Upload, search, preview, download, edit, and manage workspace files.
- Daily Briefing: Review the Daily Briefing page that depends on personalization.
- Automations: Configure recurring AI work that may use personalization context.
- GovTribe AI capabilities and limits: Understand current AI capabilities, practical limits, generated files, uploaded user files, automations, and verification guidance.
Automations (beta)
Create and manage GovTribe AI automations that run on schedules or supported workspace events.
GovTribe AI capabilities and limits
Understand current GovTribe AI capabilities, FastTrack priority processing, practical limits, generated files, uploaded user files, automations, and when human review is required.