Troubleshooting

Frequently Asked Questions

No Cards Posted This Morning. What Happened?

A run that finds no new documents posts nothing at all: no card and no note saying the window was quiet. So an empty channel does not tell you whether the agent ran.

The archive is what answers it. Ask your operator to open Regulatory Digest under Compliance in the console and look for a run at that date and cadence.

• A run is listed: the window was genuinely quiet. Nothing to do.
• No run is listed: the run did not complete. That is an ingestion or service question for your operator, not a compliance one.

Two deliberate reasons a run is skipped and posts nothing: your operator has frozen the digest, or a previous run of the same cadence finished very recently and the agent declined to post the same cards twice.

Why Didn't A Filing I Know About Show Up?

A few plain reasons. It landed outside the window that run covered. The source it came from is not one your install ingests. The classifier labeled it routine, so it did not lead a card. Or, most often, it did arrive and was counted rather than named: a card leads with one item per area and counts the rest, so a filing can be inside the run's document count without being the headline.

If a source you expect to see is never represented, that is a coverage question for your operator: which sources your install ingests determines what the digest can ever classify.

Why Is An Item Tagged Notable When I Think It Is Material?

Materiality is judged conservatively on purpose. Items that change a rule, open or close a regulatory gate, or set a clear enforcement precedent get the material label. Items that are useful context but do not shift action sit under notable, and the classifier downgrades when it is unsure.

If you read a notable item and decide it warrants action, act on it. The label is a default reading, not a decision. If a type of item keeps landing under notable when your team treats it as material, tell your operator: the classifier reads from a prompt they can adjust.

Why Doesn't The Headline Link To The Filing?

Because the cards do not carry links to the public record. A card names the lead item, says in a sentence or two why it matters for distribution or compliance, and counts the documents behind it. Reading the filing itself means opening the regulator's own site.

The one exception is the "What we have on file" sidebar, which links each cited document to its entry in the operator console. Those links point at your own uploaded material, never at the regulator's.

What Does The "No Documented Position" Line Mean?

When a material item matches nothing in your uploaded compliance and playbook documents, the card adds:

*Action:* legal flagged this and the firm has no documented position on file.

It means the agent searched your own material and found nothing on the subject. It is not an error and it is not a warning about the filing. It is the most actionable thing the digest produces: something material moved in an area where your firm has never written down where it stands.

The way to clear it is to write the position down and have it uploaded. The next card on that subject will cite it instead.

Will Regulatory Digest Ever Contact An Advisor Or LP Directly?

No. The agent posts only to your internal feed channel and your in-app feed. It does not send outbound messages, does not draft follow-ups, and does not surface anything to any outside party. Any client-facing action arising from a digest item happens through a separate agent or a teammate, and goes through that agent's own approval flow.

Can I Ask For A Digest, Or Filter One To A Single Source?

No. Regulatory Digest has no chat interface at all: no mentions, no direct messages, no commands, and no source filters. Every run covers all of its sources on its own cadence. See Commands.

If a documented phrasing like @Alpheous latest digest is in a runbook your team wrote, it never worked and should come out.

Why Doesn't The Weekly Card Look Like A Summary Of The Dailies?

Because it is not one. Every cadence produces the same kind of card from the same classifier, over a different window. The weekly card leads with the most material thing of the week rather than of the day, and counts that week's documents. It does not group by theme, add a trend note, or compare one period against the last.

Common Issues

A Card Names Far Fewer Items Than The Document Count Suggests

Working as intended. Each card carries one lead headline, up to two more named material items, and a count of everything else. A card reading Plus 40 additional documents in this run is telling you the run took in 41 documents for that area and judged one of them worth leading with.

If you need the full list rather than the lead, the run's entry in the archive is the record to open.

The Same Area Posts Nothing For Days

An area with no new documents in a window posts no card, so a quiet area is simply absent rather than reported as empty. A few days of that is normal for the slower areas.

If an area you expect to be active has been silent for more than a week, ask your operator to check that the ingestion feeding it is still running. The run list in the console shows whether runs are happening at all.

A Card Reads As Routine Background With No Clear Point

If an area produced nothing above routine in the window, its card leads with the highest-volume routine item rather than staying silent. That is the card telling you the area was read and there was nothing in it. No action is expected.

A Card Looks Truncated In Slack

Refresh Slack first; long cards occasionally render short. If it still looks cut off, open the same run in the archive, which holds the full text. There is no way to ask the agent to repost it.