Choose
Decide whether AI belongs in the task before opening a tool.
Practical AI literacy for office work
Self-paced courses for responsible AI use, role-specific workflows, spreadsheet automation, applied analysis and data stories that a colleague can actually check.
Start with a decision
One repeated task. One controlled assisted step. One named reviewer.
Scroll the timeline. Every course uses this operating rhythm, then adapts it to a specific skill or role.
Decide whether AI belongs in the task before opening a tool.
Set the audience, evidence, exclusions and owner of the final work.
Use the narrowest useful AI or automation step.
Verify facts, calculations, bias, permissions and missing context.
Record the method and communicate only what the evidence supports.
Role map
Operations workflow
Define the repeated task, source files and owner.
Use AI or a script for the narrow repeatable step.
Validate totals, exceptions and permissions.
Record the result and keep a manual fallback.
HR workflow
Separate drafting support from employment decisions.
Remove confidential and unnecessary personal data.
Check bias, accuracy, tone and accessibility.
Keep a named human responsible for the final use.
Support workflow
Identify intent without exposing sensitive records.
Generate a response from approved source material.
Check facts, policy boundaries and escalation rules.
A person owns the final response and case note.
Design workflow
Define audience, constraints and original source rights.
Use AI for options, not an unreviewed final asset.
Check consistency, accessibility and provenance.
Make and document the final human design decisions.
Seven-course route
Begin with a compact guide or follow the complete progression. Every product has a permanent detail page with exact files, access, requirements, support and limitations before payment.
A compact starting system for deciding when AI belongs in an office task and how to review the output.
A practice-led workbook for writing bounded requests, testing outputs and leaving a useful audit trail.
A practical route from repeated spreadsheet clean-up to a documented, testable automation.
A complete foundation course for using everyday AI tools with calibrated trust, privacy awareness and human responsibility.
Four role-based project tracks for turning a task into a bounded, reviewable AI-assisted workflow.
A practical path from a business question and imperfect spreadsheet to a checked analysis and concise visual story.
The full personal-learning pathway across AI literacy, role workflows, spreadsheet automation, analysis and data communication.
Human-check method
The material separates drafting, transformation and exploration from the moments that require accountability: approving facts, interpreting people, protecting data and deciding what happens next.
Read the quality policyDefine purpose, owner, source material and the risk of getting it wrong.
Use only information that is necessary, permitted and suitable for the chosen tool.
Check evidence, calculations, assumptions, bias, tone, accessibility and omissions.
A named person approves the final work, records limitations and keeps an escalation route.
Public reference layer
These sources inform the subject research. They do not endorse Polina or certify the courses.
No. Examples may refer to common workplace tools, but the core method is task-first: define the work, protect the input, review the output and retain human ownership. Tool availability and features still depend on your own account and employer settings.
Use only the supplied fictional materials or non-sensitive data you are authorised to use. Do not enter personal, confidential, regulated or commercially sensitive information into an AI service merely to complete an exercise.
No. They are educational materials, not legal advice, accreditation, an audit or a compliance certification. Organisations remain responsible for policy, data protection, procurement, accessibility and professional review.
Not for the entry material. The automation course introduces the choice between formulas, recorded actions, Power Query and lightweight scripts. Script exercises are guided and use fictional data, but permissions depend on your Microsoft or Google account.
No live cohort, project grading or consulting is included in the current product definitions. Each detail page states the exact delivery format, access period and support boundary before ordering.