How do we get started with Finster Tasks for comps, company profiles, and earnings updates using our team’s templates?

Most teams adopt Finster Tasks because they’re tired of rebuilding the same comps, company profiles, and earnings updates from scratch every quarter. The good news: you don’t need to rip up your existing templates or change how your front office works. You plug your templates into Finster once, and it keeps them fresh—data to deliverable—on demand.

This guide walks through how to get started with Finster Tasks using your own templates, and how to make them reliable enough for deal-speed, client-facing work.

Quick orientation: Finster Tasks are reusable, automated workflows that turn your instructions and templates into consistent, cited outputs—comps, profiles, earnings packs, and more—without manual data wrangling.

---

Step 1: Decide which workflows to automate first

Start with the workflows that are:

• Repetitive and time-consuming
• Template-driven
• High stakes for accuracy and recency

For most teams, that’s:

• Comps & valuation pages
  • Trading and transaction comps
  • Benchmark tables for pitchbooks and IC memos
• Company profiles
  • 1–3 page primers
  • “What’s changed since last earnings?” briefs
• Earnings updates
  • Quarterly refresh packs
  • Summary decks and one-pagers

Pick 1–2 versions of each that your team already trusts. The goal is not to design the perfect universal template; it’s to get one high-value workflow working end-to-end quickly.

Checklist:

• At least one “golden” comps format
• A standard company profile layout
• Your usual earnings update pack or memo

---

Step 2: Gather your “golden source” templates

Finster is built to work with the formats bankers, analysts, and PMs already use. You can start from:

• PowerPoint pitch pages (e.g., “Company Overview,” “Trading Comps,” “Earnings Summary”)
• Word memos / IC notes
• Excel models / tables that you regularly paste into decks
• House templates for internal newsletters or client updates

For each workflow:

1. Pick the reference output you’d be happy to send to a client today.
   This becomes the “golden” example that Finster will mirror.
2. Strip out any client-specific details.
   Replace with neutral or placeholder company names if needed.
3. Highlight what must never change vs. what should flex.
   • Fixed: structure, headings, disclosure language, caveats.
   • Flexible: tickers, peer set, time period, KPIs, charts.

Finster Tasks are built to respect your structure. You tell the system what should stay constant and what needs to be refreshed dynamically from filings, transcripts, and premium data.

---

Step 3: Turn your templates into Finster Tasks

Once you know which workflows you’re targeting and have your golden templates ready, the next step is to encode them as Finster Tasks.

At a high level, each Task has three pieces:

1. Inputs – what you tell Finster
   Examples:
   • “Ticker(s) or company name(s)”
   • “Peer list (if not auto-derived)”
   • “Time period (e.g., LTM, FY, Q/Q)”
   • “Special focus (e.g., margin, cash flow, guidance)”
2. Logic & instructions – how you want it to work
   You describe:
   • Which metrics to pull (revenue, EBITDA, EPS, leverage, etc.)
   • Where to source them (10-K, 10-Q, press release, IR deck, FactSet, Morningstar, PitchBook, etc.)
   • How to treat differences (IFRS vs US GAAP, adjusted vs reported)
   • How to handle missing data (and when to say “no answer” rather than guessing)
3. Outputs – what Finster should produce
   You define:
   • Tables (comps grids, KPI tables, segment breakdowns)
   • Charts (growth vs margin, leverage trends, valuation bands)
   • Narrative (investment summary, earnings recap, “what changed” section)
   • File formats (PPT pages, Word docs, Excel exports, email-ready summaries)

You can build these Tasks once and then reuse them on any coverage universe, deal, or portfolio name.

---

Step 4: Configure a Finster Task for comps

Comps are where traditional tools and generic AI often fall apart: definition drift, stale data, no citations. Finster takes the opposite approach: everything is traceable and auditable.

When you set up a comps Task, you typically:

1. Define the universe and peer logic

   • Base case: “Use our standard GICS/industry peers and size filters.”
   • Override: “Allow manual peer lists when I specify them.”
   • Screening: Combine quant filters (e.g., EV, margin, growth, leverage) with natural-language filters (e.g., “US-listed asset-light SaaS with >80% recurring revenue”).

2. Specify the metrics and calculations

   • Trading comps:
     • Price, market cap, EV
     • Revenue, EBITDA, EPS (LTM / NTM / FY1, FY2)
     • Multiples: EV/Revenue, EV/EBITDA, P/E, FCF yield, etc.
   • Transaction comps:
     • Deal value, enterprise value, premium, date
     • Target/Buyer details
   • Logic for:
     • Adjusted vs reported numbers
     • Currency conversion
     • Outliers (exclude / flag)

3. Lock the presentation format

   • Column ordering, row grouping, and subtotals
   • House style (decimals, rounding, labels)
   • Required footnotes and caveats

4. Enforce citations and traceability

   • Every value linked back to its source:
     • SEC filings, earnings releases, IR presentations
     • FactSet, Morningstar, PitchBook, Crunchbase where applicable
   • Finster’s citations run down to the table cell, not just document-level.

5. Set default outputs

   • One-click export to:
     • PPT comps pages
     • Excel grids
     • Word appendix for IC memos

After this one-time setup, running comps for a new ticker becomes a parameter change, not a night of Excel surgery.

---

Step 5: Configure a Finster Task for company profiles

Company profiles benefit from structure and storytelling, but assembling them manually is slow. With Finster, you keep your structure and let the system do the heavy lifting.

Your profile Task typically includes:

1. Core sections

   • Business overview and segment breakdown
   • Revenue model and key drivers
   • Historical financials and trends
   • Capital structure and liquidity
   • Management and governance
   • Recent events: M&A, guidance changes, major contract wins/losses
   • Risks and watchpoints

2. Source hierarchy

   • Primary: 10-K/20-F, 10-Q/6-K, annual reports, IR decks, earnings call transcripts
   • Secondary: FactSet/Morningstar fundamentals, PitchBook/Crunchbase (for private or earlier-stage), Third Bridge/Preqin where relevant
   • You can specify a strict order: “Use SEC filings first, IR second, then FactSet.”

3. Narrative standards

   • Tone: neutral, analyst-ready, not marketing copy
   • Length: e.g., “2–3 pages” or “single-page summary”
   • Formatting: bullet-heavy vs paragraph-heavy, use of tables vs text

4. Update behavior

   • “On rerun, highlight only what’s changed since last run.”
   • Track changes across:
     • Guidance updates
     • Capital raises / new debt
     • Management changes
     • Regulatory or legal events

5. Compliance and disclosure

   • House disclaimers embedded in the template
   • Default red-flag checks (e.g., major restatements, going-concern warnings, covenant risks)

Each generated profile is fully cited down to the sentence and directly exportable to your preferred format.

---

Step 6: Configure a Finster Task for earnings updates

Earnings season is where Teams/Slack fills with “Who’s got the latest deck?” and analysts get buried in recurring, manual work. Finster Tasks are built precisely for this cycle.

A typical earnings update Task covers:

1. Scope and cadence

   • Single-name coverage: “Generate an earnings pack after each release.”
   • Portfolio or coverage list: “Run this Task across all holdings/coverage names.”
   • Scheduled and/or triggered:
     • Time-based (e.g., 30 minutes after earnings release)
     • Event-based (e.g., when MT Newswires flags an earnings headline)

2. Document ingestion

   • Earnings release
   • 10-Q / 6-K filed the same day
   • Slides from the earnings deck
   • Transcript from the call (including prepared remarks and Q&A)

3. Key outputs

   • Headline summary: beat/miss vs consensus, key drivers, updated guidance
   • KPI dashboard: revenue, margins, EPS, FCF, user/subscriber metrics
   • “What’s new since last quarter” section
   • Management tone and key quotes (cited back to transcript lines)
   • Risks, surprises, and forward-looking statements

4. Comparison logic

   • Quarter vs prior quarter (Q/Q)
   • Quarter vs prior year (Y/Y)
   • Actuals vs guidance vs consensus

5. Delivery format

   • A short written summary for PMs or MDs
   • A PPT deck ready for client or IC
   • Tables or graphs that can be dropped into a larger pack

Again, every figure and quote is backed by sentence-level citations. If the required data is missing or ambiguous, Finster fails safely—returning “no answer” rather than guessing.

---

Step 7: Run, review, and tighten the loop

Once your initial Tasks are configured:

1. Run them on a small test set
   • 3–5 companies you know extremely well
   • At least one with messy or complex disclosure
2. Review with a skeptical eye
   • Are the numbers right, and are they cited correctly?
   • Does the narrative actually reflect the underlying filings and transcripts?
   • Does the output look like something your team would send to a client?
3. Adjust the Task instructions
   • Tighten definition of metrics (e.g., “always use adjusted EBITDA as defined in the latest filing”)
   • Refine peer logic or screening criteria
   • Add or remove sections to match your house style

Because Finster combines ingestion, structured search, and generation in one pipeline, any change to your Task logic applies across the entire workflow. You’re not chasing prompts across multiple tools.

---

Step 8: Standardize across the team

The real leverage comes when your whole team uses the same Tasks instead of ad hoc approaches.

To standardize:

• Publish “house” Tasks
  Make your comps, profile, and earnings Tasks available as the default patterns for your group, desk, or team.
• Lock critical parts, leave flex where needed
  • Locked: formatting, disclaimers, core logic
  • Flexible: ticker lists, focus areas, specific metrics
• Embed Tasks in day-to-day workflows
  • Kick off Tasks from your coverage universe or watchlists
  • Trigger portfolio-wide earnings updates automatically
  • Use Tasks as the starting point for client pitches and IC decks

This is how you move from “AI pilot” to “AI-native workflow”: the Task, not the individual, becomes the standard.

---

Accuracy, compliance, and security by design

For front-office teams, automation is only useful if it’s dependable under scrutiny. Finster is built for that constraint:

• Every insight cited, every source auditable
  Sentence- and table-cell-level citations from SEC filings, IR sites, FactSet, Morningstar, PitchBook, Crunchbase, Third Bridge, Preqin, MT Newswires, and your internal docs.
• Safe failure, not confident guessing
  When data is missing, inconsistent, or out-of-scope, Finster says “I don’t know” or returns “no answer” rather than inventing a value.
• Enterprise-grade security
  SOC 2 posture, Zero Trust model, encryption at rest and in transit, audit logging, RBAC with SAML SSO and SCIM, plus single-tenant and containerized VPC deployment options. Your data is never used to train the product.

This is the difference between a chatbot wrapper and an AI-native analyst that can survive risk, legal, and client scrutiny.

---

##Putting it all together

If you want Finster Tasks to work for comps, company profiles, and earnings updates using your team’s templates, the sequence is simple:

1. Pick your golden templates for comps, profiles, and earnings updates.
2. Define the inputs, logic, and outputs for each as a Finster Task.
3. Run and refine on a small coverage set until it matches your team’s standard.
4. Standardize and scale across your desk or firm, with house Tasks everyone uses.

You keep control of the workflow and templates; Finster automates the ingestion, search, refresh, and generation—with every number and quote fully traceable.

---

Next Step

Get Started