Train with Queries & Protocols

Teaching it Queries and Protocols

Training Patterns AI is about turning it into an indispensable financial analyst for your business. The goal is to equip Patterns with the knowledge it needs—from SQL queries to sophisticated protocols—to provide meaningful, actionable insights. This guide compresses the training process into two key elements: Queries and Protocols.

Building a Query Library

Queries are the foundation of your agent’s training. They determine how well Patterns AI understands your data and generates accurate analyses.

Defining Metrics and Designing KPIs

Key Metrics: Start by defining core metrics—such as gross margin, operating expenses, or customer acquisition costs. These metrics form the basis of future analysis, giving Patterns a foundational understanding of your data.
Metrics Graphs: Create metrics graphs—interrelated SQL queries that establish a network of relationships between key metrics. For instance, a profitability graph could connect revenue, cost of goods sold, and other expenses, allowing Patterns to develop a layered understanding of financial performance.

Standardizing Joins and Query Patterns

Automatic Joins: Patterns automatically uses FK and PK relationships for joins, but you can supplement this with explicit instructions on how different tables relate. This makes Patterns even more effective at understanding your unique data landscape.
Reusable Query Patterns: Train Patterns on common query patterns, such as joins, aggregations, and subqueries that are integral to your business logic. By building a query library, you help the AI perform consistently across different types of analyses.

Designing Protocols

Protocols are instructions that give your agent strategic direction—defining how it should act, which analyses to perform, and how to communicate findings.

Crafting Multi-Step Workflows

Example Workflow - Profitability Analysis: Start with a few key queries to calculate revenue, cost of goods sold, and operating expenses. Then, instruct Patterns to compile these results into a single profitability analysis. By creating workflows that integrate multiple queries, Patterns can produce reports that provide a comprehensive view of your business's health.
Weekly Sales Report: Define a protocol for a "Weekly Sales Report" that involves generating charts of sales trends, highlighting best-selling products, and flagging any anomalies. This workflow can run automatically, ensuring your team always has up-to-date insights.

Business Rules and Conditional Logic

If-Then Rules: Create business rules to handle conditional logic. For example, if weekly sales dip below a certain threshold, Patterns could trigger an analysis into cost variances or operational inefficiencies. These protocols ensure Patterns stays proactive, offering insights before problems escalate.
Behavioral Protocols: Define how Patterns should communicate its findings. Whether it’s summarizing the key takeaways from a report or suggesting next steps after identifying an issue, these behavioral protocols help Patterns deliver consistent, meaningful analysis.

Example Workflow - Reconciliation Tasks

Reconciliation Workflows: Create workflows that reconcile internal financial records with vendor invoices. These workflows can run automatically and alert your team to discrepancies, significantly reducing the burden of manual reconciliation.
Alerting and ERP Updates: Set up workflows that not only identify issues—like inconsistencies in financial reporting—but also prompt adjustments in your ERP system. For instance, if Patterns identifies a recurring discrepancy in purchase orders, it can alert your accounting team and suggest adjustments in the ERP to prevent future issues.

Bringing It All Together

By training Patterns AI with well-crafted queries and strategic protocols, you turn it into a proactive analyst rather than a passive tool. A properly trained agent doesn’t just answer your questions; it helps you discover the questions you should be asking.

Automated Insights: By combining queries and protocols, Patterns produces not just data but insightful, automated analyses.
Scalable Workflows: With protocols that manage multi-step analyses and alerting workflows, Patterns scales alongside your business, providing consistent value without the need for constant manual adjustments.

Ready to take your AI to the next level? Start building a robust query library and crafting intelligent protocols to make Patterns AI an irreplaceable part of your business strategy.

PreviousRequest an Analysis NextCollaboration

Last updated 2 days ago

Building a Query Library

Queries are the foundation of your agent’s training. They determine how well Patterns AI understands your data and generates accurate analyses.

Defining Metrics and Designing KPIs

Key Metrics: Start by defining core metrics—such as gross margin, operating expenses, or customer acquisition costs. These metrics form the basis of future analysis, giving Patterns a foundational understanding of your data.

Metrics Graphs: Create metrics graphs—interrelated SQL queries that establish a network of relationships between key metrics. For instance, a profitability graph could connect revenue, cost of goods sold, and other expenses, allowing Patterns to develop a layered understanding of financial performance.

Standardizing Joins and Query Patterns

Automatic Joins: Patterns automatically uses FK and PK relationships for joins, but you can supplement this with explicit instructions on how different tables relate. This makes Patterns even more effective at understanding your unique data landscape.

Reusable Query Patterns: Train Patterns on common query patterns, such as joins, aggregations, and subqueries that are integral to your business logic. By building a query library, you help the AI perform consistently across different types of analyses.

Designing Protocols

Protocols are instructions that give your agent strategic direction—defining how it should act, which analyses to perform, and how to communicate findings.

Crafting Multi-Step Workflows

Example Workflow - Profitability Analysis: Start with a few key queries to calculate revenue, cost of goods sold, and operating expenses. Then, instruct Patterns to compile these results into a single profitability analysis. By creating workflows that integrate multiple queries, Patterns can produce reports that provide a comprehensive view of your business's health.

Weekly Sales Report: Define a protocol for a "Weekly Sales Report" that involves generating charts of sales trends, highlighting best-selling products, and flagging any anomalies. This workflow can run automatically, ensuring your team always has up-to-date insights.

Business Rules and Conditional Logic

If-Then Rules: Create business rules to handle conditional logic. For example, if weekly sales dip below a certain threshold, Patterns could trigger an analysis into cost variances or operational inefficiencies. These protocols ensure Patterns stays proactive, offering insights before problems escalate.

Behavioral Protocols: Define how Patterns should communicate its findings. Whether it’s summarizing the key takeaways from a report or suggesting next steps after identifying an issue, these behavioral protocols help Patterns deliver consistent, meaningful analysis.

Example Workflow - Reconciliation Tasks

Reconciliation Workflows: Create workflows that reconcile internal financial records with vendor invoices. These workflows can run automatically and alert your team to discrepancies, significantly reducing the burden of manual reconciliation.

Alerting and ERP Updates: Set up workflows that not only identify issues—like inconsistencies in financial reporting—but also prompt adjustments in your ERP system. For instance, if Patterns identifies a recurring discrepancy in purchase orders, it can alert your accounting team and suggest adjustments in the ERP to prevent future issues.

Bringing It All Together

Automated Insights: By combining queries and protocols, Patterns produces not just data but insightful, automated analyses.

Scalable Workflows: With protocols that manage multi-step analyses and alerting workflows, Patterns scales alongside your business, providing consistent value without the need for constant manual adjustments.

Ready to take your AI to the next level? Start building a robust query library and crafting intelligent protocols to make Patterns AI an irreplaceable part of your business strategy.