Tax preparation grunt work, done by AI agents

External agents can help collect and organize information that taxpayers provide to the firm outside of CRA systems.

This can include uploaded PDFs, scanned slips, photos of receipts, spreadsheets, working papers, email attachments, client notes, questionnaire responses, and preparer interview notes. The agent can use its own LLM or document-processing tools to interpret the material, then use Armada T1 to determine where the information belongs in the tax file.

This is especially useful for information that may not be available from CRA AFR, such as donation receipts, medical expenses, self-employment income and expenses, business-use-of-home details, vehicle expenses, rental-property details, childcare information, or supporting documents for claims.

Result: Client-provided tax information becomes structured, reviewable tax-file data instead of remaining scattered across emails, folders, PDFs, and spreadsheets.

Agents do not need to guess how a Canadian tax file is structured. They can ask Armada T1 which sections exist for a given tax file and what each section requires.

For example, an agent can discover sections such as T4 slips, T5 slips, RRSP contributions, donations, medical expenses, child care expenses, self-employment/T2125, rental income, tuition, dependants, pension income, and other relevant areas.

For each section, Armada T1 can expose the available fields, required fields, field types, allowed values, validation rules, whether multiple records are supported, and whether the section requires professional review.

Result: The agent works from Armada T1’s valid tax-file structure instead of inventing fields, categories, or workflows.

Once an agent understands the relevant section and required fields, it can create structured records in Armada T1.

For example, if a taxpayer uploads three donation receipts, the agent can create three donation records. If a taxpayer uploads a T4, the agent can create a T4 instance. If the taxpayer provides a self-employment working paper, the agent can create or update the relevant self-employment/T2125 records.

These records can include source references, validation status, and review status, so the firm can trace where the information came from and decide whether it is ready to use.

The important point is that the agent is not merely summarizing documents. It is converting taxpayer-provided information into structured, reviewable tax-file data.

Result: Staff review organized tax records instead of re-entering information from loose notes, PDFs, spreadsheets, and emails.

Armada T1 can help agents identify incomplete, inconsistent, or questionable tax-file sections and turn those issues into actionable follow-up tasks.

For example, if a donation receipt is missing the eligible amount, the agent can create a task to request clarification. If a self-employment working paper indicates business-use-of-home expenses but lacks square footage details, the agent can create a follow-up task. If the client mentions moving during the year but does not confirm province of residence on December 31, the agent can create a task for staff or the client.

Tasks can be assigned, tracked, and tied to the specific section or record that caused the issue.

Result: Missing information becomes a managed workflow instead of a last-minute surprise buried in an email thread.

Agents can use Armada T1’s structured tax-file data to prepare concise summaries for preparers, reviewers, or managers.

A reviewer-ready summary can show what documents were received, what records were created, which fields were changed, which items came from taxpayer-provided documents, which items still need review, what validation warnings exist, what tasks remain open, and whether any duplicate or unusual items were flagged.

This is especially useful before senior review, because the reviewer does not have to reconstruct the file history manually.

Result: The file reaches the reviewer with a clear summary of what changed, what is missing, what needs judgment, and what is ready to proceed.

There is no need for firms to abandon their current EFILE software; Armada T1 acts as the structured front-end data gathering and task automation layer.

After the agent gathers data, creates records, validates sections, and resolves or tracks missing items, it can generate a handoff package for the firm’s existing tax software workflow.

That package can include section-by-section data tables, document indexes, source references, validation warnings, open tasks, reviewer notes, and summaries that help staff enter or import the data into their existing EFILE software.

This allows the firm to benefit from agentic data gathering without replacing the software it already trusts for final preparation and filing.

Result: Firms can keep their existing EFILE software while using Armada T1 to organize, validate, and prepare the data before the return gets there.