> ## Documentation Index
> Fetch the complete documentation index at: https://nmem.guoxudong.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Save the result
> Turn the useful conclusion into a concise decision record for your next task.
AI Now · Lesson 5 / 5 \~5 min Intermediate
Question
Source
Evidence
Investigate
Save
Save the checked outcome of your Nowledge Mem AI Now task. Include the evidence you used and any gap that research or an internal follow-up still needs to resolve.
## What you'll accomplish
You'll save a checked result in Mem, then find it and reread its source from a fresh conversation.
## Learn with the simulation
On desktop, the simulation opens automatically when you enter this lesson. Follow the popup hints to send a request, read its reply, and check the result. If you close it, click **Start the simulation** under [Your turn](#your-turn) to reopen it. Each lesson prepares its fictional Atlas materials; no model setup is needed.
The simulation runs locally with preset replies. It does not call AI or search the web, and refreshing resets the practice. On mobile, read the steps first and use a desktop browser for the interactive exercise.
## Before practising in the real app
Have the evidence brief and the finding or open question from lesson 4 ready. Public research is optional; record only the work you actually completed. Keep the lesson 2 source available in Library. A saved conversation history and a durable memory are different: this lesson saves a takeaway that can stand on its own.
## Your turn
Follow the hints in the automatically opened simulation. After practising, use the steps below to repeat the workflow with your own material in Nowledge Mem.
In the AI Now task, ask:
```text theme={null}
Draft a short record of this Atlas planning task. Include its date, whether it is a recommendation or an approved decision, the evidence and source references, a key limitation, and one next action. Preserve any unresolved question. Do not save the record yet.
```
Keep the Library source name and sections in the record itself, and a source link or identifier if available. Add verified research links and their dates only if you used them. For the fictional interviews alone, an appropriate record could be:
```text theme={null}
Atlas planning recommendation — fictional practice scenario
Status: recommendation; no pilot approved.
Recommendation: try one cross-team planning review next quarter, using existing quarterly notes to check dependencies over the next twelve months. Future review frequency remains undecided.
Basis: Library source atlas-planning-interviews.md, Interview A describes a missed dependency; Interview B conditions participation on reusing notes.
Limit: Interview C confirms no trial or measured benefit. These are fictional course notes.
Next step: ask the review owner to define a preparation limit and a measure of dependencies found, then decide whether to approve the pilot.
```
Add the date you reviewed the record. Replace the example with your actual conclusion; an unresolved decision can still be worth saving.
After reviewing the draft, ask:
```text theme={null}
Save the version we just checked as a memory in the current Space. Keep the date, fictional-practice label, recommendation status, source references, limits, and next action. If this task already has a saved result, update that record instead of creating a duplicate. Identify the memory you saved.
```
Open the saved memory and check its contents; “saved” in the reply is not enough. If AI Now cannot save it, copy the checked record to **Timeline**, press Enter, and verify the entry there. Check for an existing copy before retrying.
Open a new **AI Now** conversation in the same Space and ask:
```text theme={null}
Find the Atlas planning recommendation in Nowledge Mem. Is a pilot approved? What evidence supports the recommendation, what is still unknown, and what happens next? Identify the saved record.
```
Compare the answer with the memory. It should preserve the unapproved status and the source references.
In that fresh conversation, ask:
```text theme={null}
Find and read atlas-planning-interviews.md in Library. Quote the passage in Interview B that states the condition for participation, and explain how it limits the recommendation. Read the source itself rather than repeating the saved result. If you cannot read it, say so.
```
Check the source-reading activity and compare the passage with the document in Library. If the source cannot be read, check its Space and Searchable status before claiming the evidence is reusable.
Follow the popup hints; reopen the simulation if you closed it.
You have completed all AI Now simulations. You can continue practising in Nowledge Mem.
## What you should see
A fresh conversation finds the record with its status, evidence, limits, and next action, and rereads the original source. It does not turn your recommendation into an approved decision.
If retrieval fails, first confirm that the memory exists in the current Space. Search by its title; if needed, select it with `@` and ask again. Record whether it was found unaided or read after an explicit mention. Pasting the result back into the conversation does not verify retrieval, and saving another copy does not fix it.
Correct wording mistakes in the existing memory. When a real decision changes later, save a dated update that names the earlier record and explains the change, so you can still trace the history.
## Why this matters
The result becomes useful beyond this AI Now session. Future tasks and AI tools connected to the same Mem can start from the result record and read the Library source when they need more detail.
## Go deeper
Read [Memories](https://mem.nowledge.co/docs/memories) for guidance on keeping durable decisions current.
## You've completed AI Now
You checked saved context, read a source, assessed the evidence, chose how to investigate a gap, and verified that a fresh task can find the result and reread its source.
If you completed the simulations, repeat the workflow in Nowledge Mem with your own material to verify real retrieval and source reading. Run and check any public research separately; the simulation did not perform it.
To reuse the checked brief in another form, ask AI Now for a short report for the review owner. Keep the same sources, open questions, and unapproved status, then compare the report with the saved record.