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

Most finance teams don’t need help writing prompts. They need a way to turn the templates they already trust—comps grids, company profiles, earnings packs—into repeatable, auditable workflows that run at deal speed without sacrificing control.

That’s exactly what Finster Tasks are for: automating comps, company profiles, and earnings updates end-to-end, using your team’s own formats and standards as the source of truth.

Below is a practical, operator’s guide to getting started.

---

What Finster Tasks actually do in your workflow

Finster Tasks let you:

• Standardize workflows like comps, primers, and earnings updates into reusable templates.
• Automate data sourcing and refresh directly from SEC filings, earnings call transcripts, IR sites, and premium providers (FactSet, Morningstar, PitchBook, Crunchbase, etc.).
• Generate client-ready outputs (tables, charts, write‑ups) with sentence‑level citations for every number and statement.
• Schedule or trigger runs so recurring packs are ready before internal and client meetings.

Think of Tasks as codified workflows, not a chatbot macro. You define the structure and expectations once; Finster handles ingestion, search, and generation every time with the same standard.

---

Step 1: Decide which workflows to automate first

Even within “comps, profiles, and earnings,” you’ll move faster if you pick a small number of concrete use cases to start.

For each, define:

• Trigger: What event or cadence kicks this off?
  • Quarterly: “Post-earnings update for coverage universe”
  • Ad hoc: “New name added to watchlist” or “Client asks for a sector primer”
• Audience: Who uses this pack?
  • Internal desk prep, IC, PM, MD/client-facing
• Output format: Where does it land?
  • PowerPoint pages, Word memo, Excel comps grid, email summary, or all of the above

Common high‑ROI starting points:

• Coverage‑universe earnings updates (quarterly)
• Standard company profiles for sector or strategy teams
• Repeatable trading or deal comps packs

---

Step 2: Gather your existing team templates

Finster Tasks work best when they reflect how your team already operates. That means you don’t start from a blank page; you start from the formats you’ve been polishing for years.

Collect:

• Comps templates
  • Your standard Excel or PowerPoint comps layout
  • Definitions of each metric: source (e.g., “LTM EBITDA from FactSet”), calculation rules, any house adjustments or normalizations
• Company profile templates
  • Example decks/memos with:
    • One‑pager overview
    • Business description
    • Segment breakdowns
    • Historical financials
    • Investment thesis / key risks sections
• Earnings update templates
  • The standard format you use for:
    • Headline beats/misses
    • Guidance changes
    • Management commentary
    • KPI tables and charts
    • “What changed vs last quarter” summary

You can use real, previously sent materials (with confidential data redacted if needed) as the blueprint for your first Tasks. Finster will learn the structure and expectations, not train on your data.

---

Step 3: Turn one workflow into a Finster Task

With a template chosen, you now codify it as a Task so Finster can run it repeatedly.

3.1 Define the Task parameters

For each Task, specify:

• Task name
  • “Coverage Earnings Update – North America Industrials”
  • “LBO Comps Pack – Mid‑Cap Software”
• Scope / universe
  • Fixed list (e.g., coverage names)
  • Dynamic (e.g., “All US publicly listed airlines with market cap > $2B” using Screener)
• Data sources & priority
  • Filings and IR sites first
  • Then FactSet/Morningstar/PitchBook/Crunchbase, etc.
  • Any specific exclusions (e.g., “exclude pre‑IPO data”)

This is where you anchor the Task in your investment process, not just in generic text generation.

3.2 Encode your structure and instructions

Next, translate your template into specific instructions Finster can follow every time:

• Sections and order
  • For earnings: “1) Snapshot table, 2) Beat/miss analysis, 3) Guidance, 4) Management commentary, 5) Key charts”
  • For profiles: “1) Business overview, 2) Segments, 3) Financial history, 4) Strategy / recent events, 5) Risks”
• Tone and depth
  • “Write in concise, neutral sell‑side style, suitable for internal PM prep”
  • “Highlight only changes that are likely to be investment‑relevant”
• Mandatory inclusions
  • “Always include a ‘What changed vs last quarter’ section”
  • “Flag any restatement, one‑off item or change in KPI definitions”

Finster then applies these instructions as a repeatable workflow, not as a one‑off prompt.

---

Step 4: Configure Tasks for comps, profiles, and earnings specifically

4.1 Comps Tasks: From custom grids to automated runs

Goal: Your standard comps layout, automatically filled and refreshed.

Set up:

• Universe logic
  • Sector, size, geography filters
  • Inclusion/exclusion lists for “true comps”
• Metric definitions
  • Valuation metrics: EV/EBITDA, P/E, P/B, EV/Sales
  • Operating metrics: margin, growth, leverage, FCF, unit KPIs
  • Explicit sourcing rules:
    • “Use FactSet for price and market cap”
    • “Use SEC filings for EBITDA if FactSet field is missing or inconsistent”
• House rules
  • Normalization policies (e.g., treatment of stock comp, restructuring charges)
  • IFRS vs US GAAP handling
  • Currency and FX conventions

Output:

• Comps grid in Excel (for further tweaking), with every figure cited back to filings or data providers.
• Optionally, summary pages in PowerPoint for client or IC decks.

Because Finster’s pipeline is built for retrieval + generation, each number in the grid is backed by sentence‑ or table‑cell‑level citations. No black box, no “trust us.”

---

4.2 Company profile Tasks: Always‑on primers in your house style

Goal: A standard profile for any company that looks like it was written by your team, not by a generic assistant.

Set up:

• Core sections
  • Business description
  • Segment breakdown and revenue mix
  • Recent performance and trends
  • Capital structure and ownership
  • Strategic priorities and notable events (M&A, leadership changes, product launches)
• Depth rules by importance
  • For coverage names: deeper historical context and positioning vs peers
  • For watchlist names: more concise, focused on “why it matters now”
• Source preferences
  • “Prioritize company filings, IR presentations, and transcripts for narrative sections”
  • “Use FactSet/Preqin/PitchBook for ownership, private market data, and transactions”

Output:

• A Word or PPT profile, structured to your template with:
  • Charts filled from the latest data
  • Narrative sections with inline citations back to filings, IR decks, and transcripts

You can trigger new profiles simply by adding a name to a list or running a Screener filter, then applying the Task.

---

4.3 Earnings update Tasks: From raw filings to post‑call packs

Goal: A complete earnings pack that lands in your inbox or shared drive shortly after the release/call.

Set up:

• Event detection
  • Link the Task to your earnings calendar or a coverage list so Finster can:
    • Pull 8‑Ks, 10‑Qs/10‑Ks, press releases, and call transcripts as they drop
• Analytical structure
  • Headline table: actuals vs consensus, QoQ/YoY, key KPIs
  • Beat/miss and guidance sections
  • “What changed vs last quarter” summary
  • Management commentary: themes and quotes
• Comparison logic
  • “Compare revenue, EBITDA, EPS and key business KPIs vs prior quarter and same quarter last year”
  • “Call out any downward or upward revisions in guidance with scale and rationale”

Output:

• A pack ready for internal review:
  • Tables and charts, exportable to Excel and PowerPoint
  • Narrative summary and quotes with citations back to transcript paragraphs and filing sections

When data is missing or ambiguous, Finster returns “no answer” rather than guessing, so you see gaps instead of silent errors.

---

Step 5: Set up scheduling, triggers, and distribution

Once the core Tasks are defined, you make them run on their own.

Configure:

• Schedules
  • Quarterly earnings Tasks to run on specific days (or on event triggers)
  • Weekly or monthly monitoring packs for your coverage or portfolio
• Triggers
  • When a name is added to a coverage list → auto‑run company profile Task
  • When a company files or releases earnings → auto‑run earnings update Task
• Destinations
  • Email to specific distribution lists
  • Save to shared folders (SharePoint, data rooms, internal drives)
  • Export to Outlook, PowerPoint, Word, and Excel for slide‑ready outputs

This moves you from “remembering to run the process” to “reviewing and editing the output.”

---

Step 6: Build trust with citations and auditability

For front‑office/credit teams, the real test is whether you can defend the output to a skeptical MD or risk.

Finster Tasks are built for that scrutiny:

• Every number and statement is cited
  • Sentence‑level and table‑cell‑level citations link directly back to filings, transcripts, IR decks, or premium data sources.
• No training on your data
  • Your templates and internal documents are used to shape outputs, not to train the underlying models.
• Enterprise posture
  • SOC 2, Zero Trust, encryption at rest and in transit
  • RBAC, SAML SSO, SCIM provisioning
  • Private deployment options (single‑tenant or containerized VPC) including “bring your own LLM”

The result: Tasks don’t just save time; they produce outputs that can survive compliance and IC questions.

---

Step 7: Iterate based on real runs

Your first Tasks won’t be perfect. That’s expected. The advantage of a productized workflow is that improvements compound.

Use early runs to:

• Tighten instructions
  • If analysis feels too long, too short, or off‑tone, adjust Task prompts and constraints.
• Refine metrics and house rules
  • Clarify how to treat non‑recurring items, segment reclassifications, or pro‑forma numbers in comps and earnings Tasks.
• Add new sections
  • For example: ESG disclosures, leverage covenants, covenant‑lite flags, or sector‑specific KPIs.

Each iteration raises the floor of future outputs without adding more operational overhead.

---

How this differs from “just using a chatbot”

Many teams try to replicate these workflows with generic LLMs and “clever prompts” and hit the same walls:

• No stable template, so outputs change from run to run.
• No clear entitlements model or data provenance.
• No citations robust enough for compliance or risk review.
• Manual copy‑paste from multiple systems into decks and grids.

Finster Tasks are built specifically to avoid those failure modes. Data ingestion, structured search, and generation live in a single pipeline, so you go from source data to client‑ready deliverable with traceability at every step.

---

Putting it together: a simple rollout path

To get started practically:

1. Pick one earnings pack and one comps pack your team already trusts.
2. Upload or reference those templates in Finster and define your first two Tasks.
3. Run them on a small coverage set and review:
   • Are sections structured correctly?
   • Do citations point where you’d expect?
   • Are any metrics or house rules missing?
4. Iterate instructions once or twice until the outputs are “good enough for internal use.”
5. Scale to coverage and add company profiles as a third Task once the pattern works.
6. Turn on schedules and triggers so these workflows run automatically around earnings and key events.

You move from hours of manual prep to minutes of review, while retaining the precision and auditability your process demands.

---

Next Step

Are you ready to turn your comps, company profiles, and earnings updates into AI‑native workflows instead of one‑off prompts?

Get Started