One platform. Many workflows.

Because the orchestration, context engineering, traceability and evals are reusable, every new workflow ships in days, not months. Here are five flows already running on portfoliochat.ai.

Portfolio reporting automation

Manual problem

Portfolio commentary, performance narrative and CIO views are still stitched together by hand for every reporting period, position by position, portfolio by portfolio.

How portfoliochat.ai handles it

portfoliochat.ai calculates portfolio measures, identifies what matters for this specific portfolio and period, enriches the story with research, CIO view and news, and produces a branded review-ready draft for the relationship manager to approve.

InputsHoldings, cash, transactions, performance, P&L, exposure, restrictions, model-portfolio mapping, research notes, CIO view, relevant news.

Agents involvedPortfolio-scoped analyst → news researcher → exposure analyser → synthesis.

OutputsPosition-level commentary, top contributors and detractors, news-enriched explanations, branded PDF draft.

Human reviewRelationship manager or portfolio manager reviews and edits before the report is sent to the client.

GovernanceFull Langfuse trace per draft, eval scores on factuality and tone, deterministic performance calculations separated from AI commentary.

Client-specific news and morning briefings

Manual problem

Market news is abundant, but a relationship manager needs to know which events matter for which specific client portfolio, and why.

How portfoliochat.ai handles it

portfoliochat.ai combines portfolio holdings with high-quality news data, identifies what changed for each portfolio, and produces a morning brief tailored to the actual exposure of each client.

InputsPortfolio holdings, look-through, exposures, news feed (RavenPack), CIO view.

Agents involvedNews researcher → portfolio-scoped analyst → exposure analyser → synthesis.

OutputsMorning brief per portfolio, 'what happened in this portfolio' summaries, client-ready talking points.

Human reviewRelationship manager reviews talking points before client meetings.

GovernanceEvery news item linked to the portfolio positions it affects; full trace from news event to talking point.

Restriction-compliant rebalancing

Manual problem

When a portfolio manager wants to add or switch an instrument, the decision touches model-portfolio bandwidths, client restrictions, exclusions, currency constraints, volatility contribution, topic bias and qualitative PM judgement, all checked manually across dozens of portfolios.

How portfoliochat.ai handles it

Deterministic restriction checks run first, with strict pass/fail logic. AI reasoning is reserved for the qualitative overlay. On a breach, portfoliochat.ai reasons over the failure report and suggests an adjusted trade size that still respects the constraints.

InputsHoldings, cash, restrictions, model-portfolio mappings, exclusions, exposure, the proposed trade idea.

Agents involvedPortfolio-scoped analyst → restriction checker (deterministic) → trade-size suggester → synthesis.

OutputsPer-portfolio fit verdict, restriction-aware trade size, structured trade blotter, rationale per portfolio.

Human reviewPortfolio manager reviews and approves before any trade is sent to the trade system.

GovernanceDeterministic checks separated from AI reasoning; every check and override is logged for compliance.

Investment proposal automation

Manual problem

Prospect proposals are manually assembled from a questionnaire, a portfolio match, a suggested allocation, a CIO view, research, company content and a branded PDF. Every prospect, every time.

How portfoliochat.ai handles it

portfoliochat.ai maps the prospect questionnaire to a suitable model portfolio, scales the setup to the prospect's wealth situation, runs a final restriction-aware check, and produces a branded proposal draft.

InputsProspect questionnaire, model portfolios, CIO view, research, brand-content library.

Agents involvedQuestionnaire mapper → portfolio-scoped analyst → restriction checker → synthesis.

OutputsSuggested model portfolio, scaled setup, restriction-aware final check, branded proposal PDF.

Human reviewRelationship manager and portfolio manager review and adjust before sending to the prospect.

GovernanceEval scores on proposal quality and factuality; every claim traced to the source it came from.

Research and CIO view storytelling

Manual problem

Research, CIO views and market commentary stay generic. Translating them into client-specific narratives is manual, slow and inconsistent across advisors.

How portfoliochat.ai handles it

portfoliochat.ai combines portfolio data, relevant research, CIO views and news to generate explanations that are specific to a client portfolio or investment proposal, in your firm's tone of voice.

InputsPortfolio data, research notes, CIO view, news, brand voice guidelines.

Agents involvedPortfolio-scoped analyst → news researcher → exposure analyser → synthesis.

OutputsPortfolio-specific research commentary, talking points for client meetings, text blocks for branded decks.

Human reviewAdvisor or relationship manager reviews and adjusts before client delivery.

GovernanceBrand-voice eval scored continuously; citation eval ensures every claim has a source.

Have a workflow we have not built yet?

Book a workflow walkthrough Contact us