AI Context Playbook

AI Context Playbook

AI Context helps your organization define shared business language so responses are more consistent and actionable.

It is shared at the organization level, so the right entries can improve chat, widgets, and scheduled outputs across the team.

What to Put in AI Context

Add durable, high-reuse information such as:

• KPI definitions used in your business
• Line, area, and shift naming conventions
• Product-family and SKU grouping rules
• Quality defect taxonomy and reason codes
• Escalation thresholds used in meetings

Think of AI Context as your operating glossary, not a daily log.

Best Content Types

You can add:

• Text entries for definitions, rules, and standards
• File-based references for stable lookup material
• Image or PDF references where visual context matters
• Voice notes transcribed into reusable text context

Choose the format that makes upkeep easiest for your team.

In the product, these typically map to:

• Add Text for rules and definitions
• Upload Document for searchable source material such as SOPs, exports, trackers, and reference docs
• Add Visual for screenshots, diagrams, PDFs, and visual references
• Record Voice for quick spoken notes that should become reusable context

Writing High-Quality Context Entries

Use this structure:

1. Term or policy name
2. Exact business definition
3. Scope (which plant/line/team it applies to)
4. Exceptions or edge cases
5. Last review owner

Short, explicit entries outperform long narrative notes.

When uploading a file or visual, add a short description that says why it matters. The best description explains the business meaning, not just the filename.

Voice entries work best the same way. Treat the recorded note like a draft for a clean operating definition, not like an unstructured memo dump.

Example Entries

• "Planned downtime excludes sanitation windows between 2:00-2:30 AM."
• "Line naming standard: PlantCode-LineNumber (P3-L2, P3-L3)."
• "First-pass yield is calculated before any rework."
• "Use the shortage tracker for customer-service risk, not for historical production totals."
• "This packaging diagram shows the difference between club packs and retail packs."

Best Documents and Visuals to Add

Good candidates:

• KPI glossaries
• shift or line naming references
• recurring shortage or demand trackers
• policy documents that change slowly
• annotated process screenshots
• diagrams that explain plant layout, flow, or packaging logic
• short voice captures that should become durable operating guidance

Avoid:

• temporary updates better handled in chat
• files no one will maintain
• documents with unclear ownership
• duplicate versions of the same reference

Editing and Ownership

Treat AI Context like shared operating infrastructure.

Use it so:

• the original author can keep entries current
• admins can clean up outdated or conflicting entries
• the best definitions survive even when teams change owners

If an entry is worth using repeatedly, it is worth naming clearly and reviewing periodically.

Governance Model

Use lightweight ownership:

• Assign one owner per major context domain (operations, quality, finance)
• Review high-impact entries monthly
• Remove duplicates and outdated definitions quickly

Consistency matters more than volume.

Common Mistakes to Avoid

• Adding temporary updates that expire quickly
• Mixing policy with opinion in the same entry
• Keeping multiple conflicting definitions for one KPI
• Leaving context unowned after initial setup
• Uploading files with no description or business framing
• Treating AI Context as a dumping ground for every export

Health Check (Monthly)

Run this quick check:

1. Spot-check 5 critical KPI definitions
2. Verify line/shift names match current operations
3. Remove obsolete files or duplicate entries
4. Confirm each section has an active owner

Next Guide

Continue with Mobile and Voice Usage.