Skip to content
MergeMate.aiENDE
The fictional campaign backpack and travel essentials arranged beside a cabin window
Fictional Northline campaign. Illustrative imagery.
Product/Working methods

Working methods

Write the routine down.
Use it on the next task.

Save a reviewed set of instructions for recurring work. Mergi can use it in a project: which sources to use, what to prepare and what you want to check.

Follow the work

The next batch needs the same product checks and caption style. Keep those instructions as a working method and update them when the brief changes.

Mergi works with saved project material.

Follow a task in this room.

See what goes in and what gets prepared.
Explore the fictional project.

Northline / Open GroundWorking methodsIllustrative workspace
Project skill · illustrative working method

A considered client package.

  1. Read the current brief
  2. Inspect approved references
  3. Map deliverables
  4. Flag missing inputs before production

Adapt this method to the current brief.

The backpack beside canvas samples and sewing materials
Source reference

The useful part is the order of checks. The new brief still determines the direction and deliverables.

A starting point you can explain.Fictional project. No work is processed or saved.
Northline / Open GroundWorking methodsIllustrative workspace
Project skill · illustrative working method

A considered client package.

  1. Current brief · outdoor journey
  2. Check the latest product reference
  3. Change the output requirements
  4. Keep creative choices open

Adapt this method to the current brief.

The backpack overlooking a coastal path
Source reference
Close detail of the canvas and metal clasp
Working context

The method provides a starting structure. The campaign’s direction and output plan still need a current decision.

The process repeats. The idea moves on.Fictional project. No work is processed or saved.
Northline / Open GroundWorking methodsIllustrative workspace
Project skill · illustrative inactive suggestion

A considered client package.

  1. Suggested improvement
  2. Review the evidence
  3. Accept or keep inactive
  4. No automatic model retraining

Review before activation.

Hands packing the backpack inside a cabin
Source reference
Two people resting beside the campaign backpack
Working context

The proposed instruction is a candidate for the next method. Someone still needs to check that it is useful and appropriate.

Keep improvements deliberate.Fictional project. No work is processed or saved.

A starting point you can explain.

For a recurring client package, keep the procedure: read the current brief, inspect approved references, list deliverables and flag missing inputs. This is an editable project skill, not a fixed creative answer.

The process repeats. The idea moves on.

This month’s brief may need a different message, product or audience. Review the instructions before applying them and replace assumptions that no longer hold. A previous approval does not automatically apply to new work.

Keep improvements deliberate.

A useful correction can become a suggested memory or working instruction in supported versions. Review and activate a suggestion before relying on it. Keep evidence and version changes visible rather than treating every successful task as a universal rule.

What you can work with.

The actions behind the example.

Save the instructions

Write down the steps and checks for recurring tasks. Activate the method in the project so Mergi can read it when working on the next request.

Check the current brief

Keep the routine, then adapt the message, product and audience for this job. Instructions from a previous campaign are not a new approval.

Review proposed changes

If the clasp was cropped too tightly, record that correction as an instruction for future work. Review proposed learning before treating it as an active rule.

What you check yourself.

Review the saved instructions against today’s brief. Accept a proposed correction only if it should apply to future work.

A little more detail.

Does Mergi automatically train itself on our projects?

No. Project instructions and reviewed knowledge are different from training model weights. The workflow described here keeps proposed learning subject to review and activation.

Can I buy or publish community workflows?

A community marketplace is not part of the implemented product described here. These pages discuss editable project skills and available curated methods.