Use the AI assistant with estimates and guardrails
Ask for drafts, summaries, plans, and recommendations while keeping expensive or public actions behind estimates and approvals.
Purpose
Get useful drafts and recommendations from the assistant while keeping expensive or public actions behind cost estimates and human approval.
Who this is for
Anyone using the LumenReach assistant for drafts, summaries, plans, or recommendations.
Prerequisites
- A signed-in workspace session with assistant access enabled.
- A project with brand and goal context for more relevant answers.
- A monthly budget so assistant usage stays inside a clear limit.
Steps
- Open the AI assistant inside a project so it uses the right context.
- Describe the draft, summary, or plan you need.
- Review the usage estimate shown before a larger action runs.
- Edit the draft and keep only what fits your brand and goals.
- Send public, costly, or account-changing work to approval before it goes live.
- Check usage against your budget so spend stays predictable.
Expected result
You get a usable draft or recommendation, see its cost signal, and nothing public or costly happens until a person approves it.
Verification
- A usage estimate appears before a larger assistant action runs.
- Public or costly actions wait in the approval queue rather than running immediately.
- The budget panel reflects assistant usage after each action.
If something goes wrong
- The assistant declines an action
- Read the plain-language reason; it usually points to a missing approval, permission, or budget.
- An estimate looks too high
- Narrow the request or reduce its scope, then review the new estimate before continuing.
- Assistant access is unavailable
- Ask a workspace admin to confirm assistant access and budget for your role.
Permissions
Using drafts is broadly available; approving public or costly actions and setting budgets needs an admin or approver role.
Related articles
Support
If assistant access stays unavailable after an admin enables it, ask your workspace admin first, then LumenReach support, which can help without exposing private cost formulas.