Resources

Library

Explore Perstack and the ideas behind AI-native business through our guides and publications.

Proposed workflows

Project design & estimation

Work the agents take on
Read past designs and project documents alongside the new requirements. Draft a design, risk assessment, and effort estimate for review.
Rules and capabilities behind the work
Scope access to project material and shared knowledge. Record references so reviewers can trace the basis of each proposal.
Where AI finishes and people take over
AI handles research and drafting. A reviewer evaluates the design and makes the final estimation decisions.

Construction estimates

Work the agents take on
Read drawings and specifications to identify quantities and line items. Use the rate table to draft an estimate with supporting references.
Rules and capabilities behind the work
Limit available rates to the defined rate table. Check margin conditions during operations and route exceptions for approval.
Where AI finishes and people take over
AI prepares the estimate. People review drawing interpretation, quantities, and the final commercial terms before approval.

Skills & development plans

Work the agents take on
Adapt interview questions to each employee’s answers. Prepare achievements, a skills assessment, and a development plan for their manager.
Rules and capabilities behind the work
Carry answers and conversation history forward to guide follow-up questions. Require manager approval before publishing or updating assessments.
Where AI finishes and people take over
AI conducts the interview and drafts the assessment; a manager reviews it. Our published customer story covers a related engineering skills management deployment.
Read the SES customer story

Continuous knowledge curation

Work the agents take on
Crawl connected wikis and drives on a schedule to collect changed documents. Refresh the search index and generate reports of stale documents and possible duplicates.
Rules and capabilities behind the work
Configure sources, freshness criteria, and the run schedule in advance. Update the index and review flags while leaving source documents intact.
Where AI finishes and people take over
This example delegates collection, index updates, and reporting to AI during normal runs after setup. People revise source documents and handle exceptions such as connection failures.

Every expense reviewed

Work the agents take on
Read the expense policy alongside pre-approvals and receipts. Review every claim in scope and submit findings and reasons for attention.
Rules and capabilities behind the work
Check defined limits, pre-approvals, and aggregation rules for split claims. Require human approval for payment operations.
Where AI finishes and people take over
AI reviews claims and prepares findings. People resolve items requiring judgment and approve payments.

System activity audits

Work the agents take on
Read activity logs to examine permission changes and privileged operations. Draft an audit report with operations that need investigation and supporting evidence.
Rules and capabilities behind the work
Define the scope and criteria, and record which operations were examined. Keep execution records append-only and require owner approval to finalize the report.
Where AI finishes and people take over
AI inspects logs and drafts the report. The responsible person reviews the findings and finalizes it.