MySpend.info · Help & Guides

Advisors and Clients

Scope: How the MySpend advisor/client system works — both from the perspective of a user who wants to invite a trusted advisor and from the perspective of an advisor who wants to help their clients.

Quick-start summary

3-minute read
RoleWhat you doWhere
User — invite an advisorAI Settings → Advisors → Invite advisor → enter emailGear icon in bookkeeper
User — accept a requestAI Settings → Advisors → pending request → AcceptGear icon in bookkeeper
User — review a suggestionAI Settings → Memory → Pending from advisors → Approve / RejectGear icon in bookkeeper
User — revoke an advisorAI Settings → Advisors → active advisor → RevokeGear icon in bookkeeper
Advisor — send a suggestionAI Settings → Clients → New suggestion → select clients → SendGear icon in bookkeeper
Advisor — retract a suggestionAI Settings → Clients → Sent → Retract (while still pending)Gear icon in bookkeeper

What is an advisor?

An advisor is a trusted MySpend user — an accountant, bookkeeper, or colleague — who can suggest standing instructions for your AI Bookkeeper. When you approve a suggestion, it is added to your memory and the AI applies it going forward.

Privacy: Advisors cannot see your expense reports, transactions, or any of your financial data. They can only send suggestions to you. You decide what to approve.
As a user — receiving advice

Inviting an advisor

  1. Open the AI Bookkeeper and tap the gear icon (AI Settings).

  2. Go to the Advisors tab.

  3. Tap Invite advisor.

  4. Enter the advisor's email address (they must have a MySpend account).

  5. Tap Send invite.

The advisor receives a request notification in their own AI Settings. Once they accept, they appear in your "My advisors" list and can begin sending suggestions.

As a user — receiving advice

Accepting an advisor's request

If an advisor initiates the relationship, you'll see a pending request in the Advisors tab with an amber background, showing their name, email, and an optional note they included.

Tap Accept to activate the relationship, or Decline to dismiss it.

As a user — receiving advice

Reviewing suggestions

When an advisor sends you a suggestion, an amber badge appears on the AI Settings gear icon showing the number of pending suggestions.

Open AI Settings → Memory tab. At the top you'll see a Pending from advisors section. For each suggestion:

After approval

An approved advisor fact behaves like any other active memory fact: it is injected into every conversation, ages normally, and can be disabled (checkbox), pinned (📌), or deleted (trash icon). Advisor facts will not be auto-compacted — they are preserved with their source attribution intact.

Deleting an approved advisor fact

Open the Memory tab, find the fact (look for the from [Name] chip), and click the trash icon. This removes the fact from your memory permanently. The advisor's sent history will still show that it was sent and approved.

As a user — receiving advice

Revoking an advisor

  1. Open AI Settings → Advisors tab.

  2. Find the advisor in "My advisors".

  3. Click Revoke and confirm in the inline prompt.

Revoking an advisor ends the relationship immediately — they can no longer send you suggestions. Facts that were already approved are not deleted automatically; those remain in your memory until you delete them manually.

As an advisor — giving advice

Getting started as an advisor

The Clients tab in AI Settings is visible only once you have at least one active client relationship. There are two ways to start:

Once the relationship is active, your client appears in the Clients tab.

As an advisor — giving advice

My Suggestions Library

The My Suggestions Library (a collapsible section in the Clients tab) lets you save reusable suggestion templates — so you don't have to retype common instructions every time.

Adding to your library

  1. Expand My Suggestions Library.

  2. Click + Add.

  3. Enter the suggestion text (max 128 characters).

  4. Click Save.

From the library you can also Edit a saved suggestion, Delete it (confirmation required), or click Send to open the send panel pre-filled with that suggestion's text.

As an advisor — giving advice

Sending suggestions

Click + New suggestion (always visible below the library). A send panel opens:

  1. Enter the suggestion text (max 128 characters). If you tapped "Send" on a library item, the text is pre-filled.

  2. Check the boxes for the clients you want to send to. You can select multiple clients at once.

  3. Click Send to N client(s).

The panel shows per-client results — a green checkmark for each successful send, or a red error if a specific send failed (you can retry). On a fully successful send, the panel closes automatically.

As an advisor — giving advice

Tracking sent suggestions

The Sent section at the bottom of the Clients tab shows all suggestions you have sent, with a status label on each:

StatusMeaning
pendingThe client has not yet approved or rejected it
approvedThe client approved it; it is now in their memory
rejectedThe client rejected it
retractedYou retracted it before the client acted on it

Retracting a suggestion

If a suggestion is still pending, you can retract it. Click Retract on the sent history row and confirm. The suggestion is removed from the client's pending queue — they will not see it, and your history shows it as "retracted".

Retracting an already-approved suggestion removes the fact from the client's memory immediately, even though they already approved it. Use this carefully.

What this guide does NOT cover

Out of scope

  • Advisors reading client data — this is not possible. Advisors can only push suggestions; they have no access to a client's expenses, transactions, or financial summaries.
  • Pinning facts on behalf of clients — advisors can suggest facts, but only the client can pin them.
  • Prompt suggestions — a future feature will allow advisors to suggest AI prompt templates. This is not yet available.
  • Granular advisor permissions — all advisors currently have the same capability (suggest facts). Per-advisor permission levels are planned for a future release.
  • Memory compaction of advisor facts — advisor-sourced facts are intentionally excluded from compaction to preserve attribution and allow retraction.