Review a MEL Plan Draft

Review the supplied MEL plan draft for evidence-based completeness, feasible operations and conditions for the next review.

On this page
What You Need
Paste the full MEL plan and appendices, indicator sheets, data flow, staffing and budget, plus applicable submission rules.
Choose This Prompt When
Use this to review an existing MEL plan. Choose Draft a Data Collection Plan to develop methods, or Review an Evaluation Report Draft for completed-study reporting.

Review the actual MEL plan draft for its purpose and stage. Request missing text and relevant appendices, indicator sheets, instruments, role commitments, budget and submission criteria; filenames do not establish content.

Assess six dimensions: completeness across ToC links, indicators, sources/methods, data flow, responsibilities, quality, ethics/protection, learning, budget and timing where relevant; concrete source/method/frequency/owner for each indicator; coherence with intended program change rather than activity completion alone; staffing/budget realism; meaningful disaggregation and visible equity; and how data feed review and decisions. Flag dependence on undefined systems.

For each issue, cite the actual section or row and distinguish a documented applicable gap, justified N/A and unavailable evidence. Continue review of supplied components, leaving unseen appendices and unsupported operations Not Assessable. Supporting records can inform a judgment but do not make an undocumented plan section complete. Do not credit safeguards the draft never describes or infer a procedure never happened solely from missing documentation.

Return an overall assessment, a six-dimension evidence/judgment/action table, must-fix/should-fix revisions and conditional internal-review/submission advice. Explain ratings; numerical scores need a supplied anchored scale. Planned systems are not implemented or donor-approved, and actual approvals remain with the responsible owner.

Loading saved answers...
Did this page help?

Worked Example

Illustration only. The setting and figures are invented to show the pattern.

Scenario: A five-year agriculture program is sending its MEL plan to internal review. The plan has an indicator table, but staffing and budget sit in other files.

What You Give the AI: The full plan, the indicator reference sheets, the staffing plan, the M&E budget lines and the submission checklist. The plan also relies on a referral tracking database that has not been built.

What the AI Might Return: A six-dimension table. One row reads: dimension 'Coherence', location 'section 4, indicator 3', finding 'counts training sessions delivered, which shows activity completion but says nothing about change in practice', judgment 'Needs Revision', action 'add an adoption indicator'. The conditional advice lists the unbuilt database as a dependency and leaves budget realism Not Assessable until the budget lines are added.

When to Choose This Prompt

  • A MEL plan exists and you want it checked before internal review or submission.
  • You want gaps found by dimension and located to a section or row.
  • You want conditional readiness advice that leaves approval to the people responsible. Pick this over Review My MEL Plan when you need the six-dimension table, and pick that one for a quicker list of the main gaps.

When Not to Choose It

Checking the Output

  • Open several of the sections or rows the review cites. Each should exist in your plan and say what the review claims.
  • For each indicator, compare the AI's statement of source, method, frequency and owner with your reference sheets. Gaps it lists should be real.
  • Whatever sits under Not Assessable should be the appendices and budget lines you did not paste.
  • Disaggregation comments should name the categories in your plan.
  • Credit for safeguards should go only to measures the draft describes.
  • Submission advice should stay conditional. A statement that the plan is donor-ready has no support.

Common Problems and Fixes

  • Safeguards the draft never mentions get credit. Fix: Tell the AI to credit only what the text describes and to list the rest as evidence gaps.
  • A planned system is treated as working. Fix: Flag every dependency on a system that does not yet exist.
  • A section is called missing because the appendix was not pasted. Fix: Paste all parts, or have those items marked Not Assessable.
  • Numerical scores appear. Fix: Explain ratings in words unless you supply a scale with anchors.

Adapting It to Your Context

  • Submission rules apply: Paste the criteria and ask for gaps to be listed against them as a separate pass.
  • Early draft: State the stage and ask for must-fix items only, with polish left for later.
  • Plan depends on partner systems: List the partners and the systems they run, so dependence is judged on facts you supplied.

Next Step

MEL Plan Review