Working examples

Every example contains:

• Query — the content of the entered message.
• Suggested analysis depth — Surface, Balanced, or Deep.
• What the advisor does — the skills appearing in the message timeline.
• Response — the structure and type of citations.

The actual content of every response may vary. These are only examples; however, the general shape of the response and the invoked tools generally remain similar.

Example 1 — Votes

Query:

Which Civic Coalition stakeholders voted against the proposed amendment to the energy bill at the last Sejm session? Show the resolution result and link to the vote.

Suggested depth: Balanced.

What the advisor does:

1. Legislation — List votings — enumeration of recent votes; filters down to votes concerning the energy bill and the amendment.
2. Legislation — Get voting — retrieval of the per-mandate distribution and metadata.
3. Stakeholders — Get stakeholder — matching names to identifiers for voting Civic Coalition members.

What the response looks like:

The amendment was rejected in Voting 47 of Proceeding 12 on 28 March 2026, with a result of 198 for / 232 against — see [legislation://voting/12/47]. The following Civic Coalition members voted against: [feature://stakeholder/abc-123], [feature://stakeholder/def-456], [feature://stakeholder/ghi-789] (3 out of 134 club members).

Example 2 — Public stance versus voting record

Query:

The Climate Coalition has recently been vocal publicly regarding the coal phase-out. How did its members actually vote in the plenum on coal-related resolutions in the last month? Are there any discrepancies?

Suggested depth: Deep.

What the advisor does:

1. Stakeholders — Search legislators — filtering by Climate Coalition affiliation.
2. Public pulse — Search tweets — searching posts from the last month containing coal phase-out terms.
3. Legislation — List votings — filtering to the same date range.
4. Legislation — Get voting — details of each candidate vote.
5. Legislation — Votes by legislator — cross-verification by club membership.

What the response looks like:

Out of nine plenary votes on coal amendments in the last 30 days, the Coalition was united in six. Three discrepancies:

• Stakeholder Kowalska ([feature://stakeholder/...]) wrote publicly about the proposed transformation fund ([feature://public_pulse/tweet/...]), but voted abstain in the funding amendment ([legislation://voting/12/52]).
• The Coalition split 7–2 on the mine closure schedule amendment ([legislation://voting/12/61]).
• Stakeholder Nowak ([feature://stakeholder/...]) was absent for three of the nine votes.

Example 3 — Document analysis with an attached file

Query:

Summarize the regulatory recommendations in this briefing and check if any of them were passed at recent Sejm sessions (with attached Q1-2026-renewables-briefing.pdf).

Suggested depth: Balanced.

What the advisor does:

1. Documents — Read file — reading recommendations directly from the attached file (without a search step — the file identifier is provided).
2. Legislation — List votings — filtering by topics covering each recommendation.
3. Legislation — Get voting — details of matching votes.

What the response looks like:

The briefing contains five recommendations [Q1-2026-renewables-briefing.pdf, c1]. Among them:

• Recommendation 2 (simplification of the connection permit process) corresponds to amendment 14 in the energy bill, passed on 22 March — see [legislation://voting/12/41].
• Recommendation 4 (expansion of the contract for difference limit) was proposed but withdrawn before the vote [Q1-2026-renewables-briefing.pdf, c4].
• The remaining three recommendations have not appeared in proposed legislation so far.

Example 4 — Alert configuration with elicitation

Query:

Configure an alert tracking any reference to the new pharmaceutical reform in plenary speeches and uploaded documents.

Suggested depth: Surface.

What the advisor does:

1. Legislation — List votings — searching for the bill referenced in the question.
2. The match returns two candidate bills — the tool emits an elicitation requesting a selection.
3. (selection from the embedded form)
4. Alerts — Create alert — creating an alert with the specific bill, description, and narrow keywords.

What the response looks like:

Alert Pharmaceuticals Reform 2026 ([feature://alert/...]) has been created. It tracks plenary speeches and uploaded documents. Status: Active, visibility: Private — this can be modified in the alert detail view.

Example 5 — Extracting a document fragment

Query:

Find the fragment in the latest Justice Committee transcript where the chairman announced the schedule for the public procurement bill amendment.

Suggested depth: Balanced.

What the advisor does:

1. Documents — Hybrid search — searching committee transcripts for terms related to the procurement bill.
2. Documents — returning a link for the matching fragment.

What the response looks like:

The chairman announced the schedule in the transcript from 30 March 2026 ([documents://chunk/.../...]). Cited fragment:

„The procurement-law amendment will be considered in second reading at the next sitting on 5 April."

The button opens the file scrolled to the cited fragment.

Example 6 — Multi-source legislative briefing

Query:

I need a one-page brief on the state of the renewable energy bill: legislative status, recent committee actions, public party positioning, projected voting outcome, and stakeholder concerns.

Suggested depth: Deep.

What the advisor does:

1. Legislation — List votings, Get voting — the legislative trajectory of the bill.
2. Documents — Hybrid search — committee discussions over the last three weeks.
3. Public pulse — Search tweets — positioning of every major club over the last two weeks.
4. Legislation — Votes by legislator — projecting the vote outcome based on previous alignments.
5. Documents — Hybrid search — internal briefings regarding this bill.
6. Alerts — List alerts — verifying whether an alert covering this bill exists.

What the response looks like:

Status. The bill is in its second reading. The Energy Committee reported positively on 28 March ([legislation://voting/12/38]).

Recent committee actions. The chairman announced 5 April as the voting date ([documents://chunk/.../...]).

Public positioning. The Civic Coalition supports it with reservations ([feature://public_pulse/tweet/...]). PiS opposes it, citing industrial costs ([feature://public_pulse/tweet/...]).

Projected outcome. The bill will likely pass with a margin of 30–45 votes.

No alert covers this bill — an option to create one is available.

Recurring patterns in the examples

Three patterns recur in the examples above:

• The advisor starts with structural searches. Voting records, stakeholder identities, and committees provide an anchor — text searching is narrowed by what the structural tools found.
• Multi-source citations are the norm. Even simple queries produce several URI badges, enabling verification of every claim.
• The advisor gracefully closes its work at turn limits. A Deep query ending at the limit returns a full response for the covered portion and a single-line note regarding the deferred scope.