Skip to main content

​
Overview

Even the best models or agentic systems will fail without the right context. If the system doesn’t know what active users means in your organization, it cannot produce the right result. That’s why context is a first-class concept in Bag of words: every Agent run begins by constructing a context block that contains definitions, rules, and prior learnings. Bag of words includes features for creating, managing, and monitoring context and context performance, including support for complex data structures. For best results, data teams should ensure that context coverage is thorough and use the available tools to track and improve context quality.

​
What goes into context

Context block diagram Every Agent run begins by constructing a context block. This block combines static context (persistent organizational knowledge) with warm context (dynamic signals from the current run). Together, they ensure the Agent reasons with your business semantics, not guesses.

​
Static context

Static context is the foundation: information that rarely changes and encodes the organization’s definitions. It is loaded at the start of each run.
  • Instructions — scoped rules and guardrails such as KPI definitions, preferred joins, exclusions, or conventions (e.g., “use fiscal calendar,” “exclude PII columns”).
  • Schemas & tables — metadata describing sources, tables, columns, relationships, and types.
  • dbt/LookML, Git repos, Agents.md — canonical definitions of metrics, joins, and code that encode business logic.
  • Learnings — knowledge accumulated across runs: query history, user feedback, validation outcomes, lineage down to the table/column level.

​
Warm context

Warm context is transient and grows throughout a single Agent run. It captures everything the system learns or produces while satisfying the request.
  • Previous messages — the conversation history for the request.
  • Tool calls — queries, dashboards, and their intermediate results.
  • Clarifications — user inputs captured when the Agent asks for missing info.
  • Observations — errors, validation outcomes, and reflections from the loop.
This evolving layer allows the Agent to refine its reasoning with every step.

​
Monitoring & Context Quality

Because context is central to the effectiveness of AI systems like Bag of words, we offer tools to monitor its quality and impact on system accuracy. You can view high-level metrics on context performance and drill down into individual agent runs for detailed analysis. These capabilities help ensure your results are reliable and your analyses are accurate.

​
Instructions Effectiveness

Each time a user prompt is processed, the system evaluates how well the provided instructions (business rules, KPIs, definitions) contributed to producing the result. This produces an Instruction Effectiveness Score—a measure of whether the right definitions were applied at the right time.

​
Context Inspection for Every Prompt

Admins can review the full context used in each Agent run to verify decisions and identify areas for improvement. More on monitoring in the Monitoring section.