Scout limits, credits, and known issues

Scout works inside Excel using Daloopa data. This page covers the main limits, timing expectations, credits, and what to do when something goes wrong.

Where the edges are

Boundary	What it means
Daloopa data only	Scout does not use external connectors or browse the web. If Daloopa does not cover a company or line item, Scout cannot fill it from somewhere else. You can paste figures into the workbook and ask Scout to work with them.
Excel only	Scout does not work in Google Sheets or a separate web app.
Inside the workbook	Scout writes to your workbook. It cannot save files to your machine, so it does not export PDFs or PowerPoint files.
Excel has to be open	You start each run, and Scout works in the foreground while Excel is open. You can leave it running and do other things, but you cannot schedule a build or run one with Excel closed. For substantial requests, you approve a plan first, which is one click before the work starts.
Nothing is shared with your team	Chat history lives in your own browser profile. Global instructions are yours alone. There is no shared workspace or team prompt library yet.
You cannot save a named skill yet	You can give Scout your modeling conventions by pasting them into a prompt, or by putting them in Global instructions so they apply to every session. You cannot yet save those conventions as a reusable, shareable skill of your own.

Timing

Request	Typical timing
Question answered in chat	Under 10 seconds
A step that writes to the sheet	Around 2 minutes, with a lot of variation
A long continuation	10 minutes or more
Full three-statement model	70 minutes to a bit over 2 hours in our runs

Smaller requests finish sooner. You can run more than one session at once.

Credits

Scout usage is metered in credits. Each request Scout sends to its AI model uses credits, based on how much work the request involves.

Your monthly credit allowance is set by your plan.
Some firms share one pooled allowance across their Scout users, rather than giving each user a separate allowance.
Credits reset at the start of each calendar month.
If you hit the cap, new requests are blocked until the reset or until your account team raises the cap. Nothing is deleted; you just cannot start something new.
A request that is already running when you hit the cap will finish.
Failed requests do not cost credits. If a run errors out, nothing is charged.
Your account team can raise your allowance or grant one-off credits.

Roughly what things cost

Workflow	Typical credits, median
Build a new model	~400
Update a model with a new quarter	~175

A quick question answered in chat costs very little. A large build across many periods and companies costs more.

Telling us about a problem

If Scout gives an incorrect result: send your CSM the ticker and the prompt. That includes rows in the wrong place, a value in the wrong period, or a run that stopped. The prompt helps us reproduce the issue.

If Scout completes the request but the output is not useful: tell your account team what you expected and what was missing.