Unwrapped

Teardown · hebbia

HEBBIA

HEBBIA

CategoryEnterprise SearchFunding · undisclosedSite ↗
  • Andreessen Horowitz
  • Index Ventures
  • GV (Google Ventures)

Private docs + public filings + financial data providers + LLM APIs + retrieval workflow.

01

Public data / API layer

CD
Customer Documents (Box, SharePoint, internal repositories)Yours
SEC EDGAR
SEC EDGARPublic
FactSet
FactSetLicensed
S&P Capital IQ
S&P Capital IQLicensed
PitchBook
PitchBookLicensed

Internal replication score

Easy
0.71

Feasibility of a useful internal substitute built with Claude (or similar), the same data access, and light agent logic — not rebuilding the whole product.

IRS = 0.30·D + 0.25·L + 0.20·O + 0.15·R + 0.10·Sthis record · 71%
  • D

    Data accessibility

    weight 0.300.70
    • 1.0mostly customer-owned / public / standard third-party sources
    • 0.5mixed accessibility
    • 0.0hard-to-access or proprietary source layer
  • L

    LLM substitutability

    weight 0.250.85
    • 1.0mostly retrieve / prompt / cite / summarize / classify / compare
    • 0.5mixed standard + custom behavior
    • 0.0strongly custom model behavior (fine-tunes on proprietary data, etc.)
  • O

    Output simplicity

    weight 0.200.75
    • 1.0straightforward internal work product (memo, list, reply, SQL query)
    • 0.5moderately specialized
    • 0.0highly specialized (e.g. FDA-graded clinical text)
  • R

    Review / risk tolerance

    weight 0.150.60
    • 1.0internal use with human review is acceptable
    • 0.5moderate risk
    • 0.0very low tolerance for error (e.g. external legal filings)
  • S

    Surface complexity

    weight 0.10inverse — higher means less surface dependence0.50
    • 1.0a simple internal shell is enough
    • 0.5polished workflow matters somewhat
    • 0.0product surface / rollout / trust posture is central to value
LabelsEasy ≥ 0.67Medium ≥ 0.34Hard < 0.34

Missing factor rows use heuristics from wrapper scores. Editorial heuristic, not investment advice.

Build it yourself→

Recreate the workflow inside your org.

Internal build

Build it yourself

Same data connectors + frontier LLM + retrieval agent + citation prompts — lacks polished workflow and financial domain tuning.

Internal use only. Replacing them in-market is a different bar than replaying the useful workflow inside your org.

01 · Connectors & flow

CD
Customer Documents (Box, SharePoint, internal repositories)
SEC EDGAR
SEC EDGAR
FactSet
FactSet
S&P Capital IQ
S&P Capital IQ
PitchBook
PitchBook

Internal build map

Data in

Connectors
Connectors

Agent layer

Planner
Tools + retrieval
Reasoning model

Logic

LLM API
retrieve
rerank
permissions
cite
compare
not custom weights

Outputs

Internal search
Answer
Citations

02 · Claude / agent prompt

Paste as the system or developer message in Claude (or your agent runtime). Scroll to read; Copy grabs the full text.

Claude / agent prompt

// Financial research assistant on internal and licensed data sources You are a research assistant inside [YOUR_COMPANY] helping investment, legal, and corporate teams. You only use materials the user is allowed to access: internal documents (SharePoint, Box, local files), licensed financial data (FactSet, S&P Capital IQ, PitchBook), and public filings (SEC EDGAR). ## What you must do 1. Retrieve first: Always search connected sources before answering. If a question requires multiple sources (e.g., comparing companies, analyzing trends), retrieve from all relevant documents. 2. Cite rigorously: Every factual claim must cite the source document, page number, and section. If multiple sources support a claim, cite all. 3. Surface conflicts: When sources disagree (e.g., different revenue figures in a CIM vs. audited financials), flag the discrepancy explicitly. 4. Scope: Answer only what the sources support. If information is missing, state what's unavailable and why. ## What you are not Not a replacement for human judgment in investment decisions, legal filings, or compliance review. Internal use only — outputs require human validation before external use. ## Refusal Refuse if: (1) the user asks for information outside connected sources, (2) the query implies you should infer facts not grounded in retrieved text, (3) the request involves generating external-facing materials without explicit human review. ## Safety Internal posture only. Outputs may contain errors. All analyses require human review before informing investment decisions or legal filings.

03 · Result

Compare revenue growth rates for Project Alpha and its three closest competitors over the last three years.
CIM pages 14-16, Cap IQ financials

Project Alpha: 18% CAGR (2022-2024). Competitor A: 22%, Competitor B: 15%, Competitor C: 12%.