Mentors review the file, in a fixed order, then expect a second version.

The review order GVI asks mentors to use

Question quality. Source quality. Analytical logic, including sensitivity. Consumer or stakeholder trade-offs. Clarity of the recommendation. Realism of next steps. Communication in the deck and the talk. Time management as evidenced by drafts. Integrity if something looks un-owned.

This order stops mentors from spending the scarce coaching call on colors. It also tells students where to spend the last twenty-four hours: on the joint that actually fails.

How to arrive

Send the file. Send three questions that show forks, not “is this good?” Send a paragraph of what changed since last time. If you cannot send a file, cancel and reschedule. The brochure’s two-call model is too small to waste on logistics.

What happens live

You restate the recommendation. The mentor pressure-tests one weak joint. You agree two revisions and a filename. Career advice, if any, happens after the file is honest. Mentors who rewrite your slides in the call are failing the standard even if it feels luxurious.

Written comments

They should cite locations: page, slide, cell. They should be readable by a parent without humiliation. They should expect a dated revision. Students should not argue first; they should revise, then explain if they disagree with a comment using evidence.

When mentors suspect the work is not yours

They will ask you to walk through the model, to show the source log, to explain a sentence that does not sound like you. This is not an accusation of character in advance. It is how a virtual internship remains educational. See standards.

How this connects to recommendation support

A letter, if written, will describe what the mentor reviewed. If they only saw a finished deck on the last day, they cannot say much — and they should say no. Students who want letters should make the review trail easy.

Becoming easier to mentor

Be on time. Label files. Accept that “I worked hard” is not a defense of an unsupported claim. Hard work that does not change the recommendation is still unfinished work.

Examples of comments that change files

“Cell D4 treats an unsourced blog number as a fact; move it to assumptions or drop the claim.” “Slide 3 contains two recommendations; the title should pick one.” “You rejected the cheaper option because it is cheaper — that is not a reason a budget holder will accept.” “Your next step is a slogan; name an owner and a first week action inside the brief’s constraints.”

Students can use this list as a pre-flight before sending a draft. If you can write one of these comments about your own file, you are beginning to self-mentor, which is the remote skill underneath coaching.

After the call, before the glow fades

Rename the file. Highlight the two changes. Put a date on the revision. If you wait until the weekend, you will remember the warmth and forget the joint that failed. Mentors may ask for a diff. That is not distrust; it is how a two-call model covers a multi-week pathway.

Asking for a diff without being a villain

“Please send v2 with the two agreed changes highlighted, by Thursday.” That sentence is mentoring. Students sometimes hear distrust. Teach them it is how remote teams work and how a two-call model covers many weeks. If v2 arrives with unrelated new slides and the same weak joint, name that in the performance review. Patterns, not vibes.

Mentors should keep a short private note: date, file, two asks, whether they landed. That note is what makes a later recommendation letter possible. Without it, memory will flatter or fail. GVI would rather a dry note than a glowing letter about work nobody saw.

If staffing is too thin to read diffs, reduce cohort size. Do not keep the poster and drop the marking. The school-cohort insight and the partner checklist both exist because this temptation is common.