Hire an AI grant spending and receipt review

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A reconciled table that matches each cost to a budget line, an award clause and its proof
• A variance brief that separates timing differences from real overspend
• A list of costs with missing or mismatched proof
• Questions for the fund owner on costs where eligibility is unclear
• A draft fund owner update, not sent
• A review log for your next review

How it does the job well

• Use only the clauses supplied. Cite each one by reference. Do not apply rules from memory or from other grants.
• Where a clause is unclear or silent, write Unknown and a question for the fund owner. Do not decide.
• Never change the source files. Save revised files under new names.
• Do not invent amounts, receipts, approvals or dates. Keep proposed reclassifications apart from what the ledger shows.
• Agree the question and the decision it supports before analysing.
• Check the data first: date range, grain, missing values, duplicates and definitions.
• Compare like with like: same period length, definitions and segments.

Mistakes it avoids

• Deciding eligibility where the clause is silent.
• Presenting this as an audit or an approval.
• Treating an empty export as proof of no spending.
• Reclassifying costs without the clause in front of you.
• Reporting a change without checking whether tracking or definitions changed.
• Treating correlation as cause.

What a good result looks like

An example, not your task: A grant reporting date is coming and you need spending matched to the award terms.

• spending-reconciliation.csv. Each transaction matched to a budget line, an award clause, its period and its proof.
• variance-brief.md. Actual against budget by line and period, with timing, overspend and unapproved moves kept apart.
• evidence-gaps.md. Costs with no receipt or a mismatched one, and who could supply it.
• eligibility-questions.md. Costs where the clause is unclear, each written as a question for your fund owner.
• fund-owner-update-draft.md. Totals, open questions and decisions needed. A draft only, never submitted.
• review-log.md. Missing facts, checks not done, what changed and the state for the next review.

How the work is checked

• Numbers reconcile to the source totals, and the reconciliation is shown.
• Every chart and number states its source, period and definition.
• Limits and next checks are listed.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

Capabilities

• Read the last state. It keeps the award ID, budget version, transaction cutoff, the source of the restrictions and open costs. The first run sets a baseline.
• Check what triggered the review. A reporting date is due, a restriction or budget version changed, or new transactions arrived.
• Pull the award clauses. It lists the clauses on eligible costs, limits, dates and prior approval, with clause references. It uses only these.
• Reconcile every transaction. Each cost is matched to a budget line, a period, a clause and its proof. Totals are checked against your ledger.
• Explain the variances. It compares actual spending with the budget by line and period, and flags unapproved moves between lines.
• Draft the questions and update. Unclear costs become questions for the fund owner. The update draft lists totals, open questions and decisions needed.
• Save the review. It records what changed, what is missing and the state for next time.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id grant-spending-evidence-review-assistant only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/grant-spending-evidence-review-assistant/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• The award agreement, or the clauses on eligible costs, limits, dates and reporting
• The approved budget, with its version or date
• Transaction records for the reporting period, up to a cutoff date
• Receipts or invoices for those transactions
• The reporting period and the name of your fund owner
• Your last review file (optional)

Helpful context (optional)

• The funder's reporting template.
• Prior approval emails for any change to the budget.
• A list of costs your fund owner already confirmed.

Tools and access

• Claude, ChatGPT or Gemini with file upload: It reads the award text, the budget and the transaction list together.
• A spreadsheet tool such as Excel or Google Sheets: It lets you recalculate formulas and check totals in a workbook you can open.
• A compatible spreadsheet engine: Formula recalculation and workbook inspection need an actual compatible engine; static table checks can use supported scripts.
• Any good chat AI: Claude, ChatGPT and Gemini all handle a long source in one message. Claude takes the longest texts.

Functions and deliverables

• Grant spending and receipt review: Match grant spending with its budget, rules and receipts. See changes, disputed costs and missing evidence.
  • A reconciled table that matches each cost to a budget line, an award clause and its proof
  • A variance brief that separates timing differences from real overspend
  • A list of costs with missing or mismatched proof
  • Questions for the fund owner on costs where eligibility is unclear
  • A draft fund owner update, not sent
  • A review log for your next review
• Grant spending and evidence report preparer: Share grant rules, budget, transactions and receipts. Get a spending table, budget differences and missing evidence for the fund owner.
  • A reconciled grant spending table, variance brief and evidence gaps.
  • Step files with evidence, unresolved facts and actual checks completed.
  • A hand-back summary and one practical next step.
• Source note summariser: Paste an article, talk or chapter into this source note summariser. You get each claim, how well it is backed, the gaps and three quotes to keep.
  • Every claim with a quote and location, and the evidence behind each rated
  • The gaps and assumptions
  • What it adds to your question, three quotes to keep, and an 80-word summary
• Nonprofit grant proposal writer: Write a nonprofit grant proposal, inquiry letter or grant report that follows your funder's rules. Give it the guidelines, program facts and budget.
  • A draft in the funder's section order and exact headings.
  • A budget narrative that shows the math for each line.
  • A list of totals that do not match and attachments that are missing.
  • A draft trimmed to the funder's limits without cutting the ask, outcomes or evaluation plan.
  • A list of what program staff, finance and your director must check.
  Source: Nonprofit Skills.
• Metric definition writer: Share a metric, calculation and business use. Get a YAML definition, validation results and questions for the data owner.
  • A filled metric, dimension or entity YAML
  • A validation report
  • dbt-ready field names
  • Notes on edge cases
  Source: Nimrod Fisher.

Tasks you can use it for

• A grant reporting date is coming and you need spending matched to the award terms.
• The funder changes a restriction, or you get a new version of the budget.
• New transactions arrive and you need to know which ones are eligible.
• Your fund owner asks why spending is above or below budget.

Limits to check before starting

• You are still writing the bid. Use the Grant proposal writer.
• You need restricted and general cash kept apart. Use the Restricted grant cash tracker.
• You need an audit opinion or a funder return. Your fund owner or auditor does that. This agent submits nothing.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Download the pack from this page and add it to your AI. If you cannot add it, paste SKILL.md into a new chat.
• Attach the award agreement, the approved budget and your transaction export. Add receipts, or a list of which transactions have one.
• Type the reporting period, the cutoff date and your fund owner's name. Attach your last review if you have one.
• Say "Reconcile this spending to the award rules. Do not decide eligibility where the clause is unclear." Your AI asks once for anything missing.
• Open eligibility-questions.md and evidence-gaps.md first. Send them to your fund owner yourself.
• Ask your fund owner or an auditor to review the table before you rely on it. Keep review-log.md for next time.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/grant-spending-evidence-review-assistant for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.