01
AlphaLens entity data
Enrich the deal with known company truth.
Firmographics, funding, people, products, addresses, and growth signals give every row a durable foundation before any agent runs.
Add trusted answers to every company row.
Start with robust AlphaLens entity data, then use agentic AI to answer the questions only live web research, pitch decks, and documents can resolve. Every answer includes sources and confidence interval.
cited
answers with sources
custom
agent columns
auditable
screening outputs
Pipeline staging
Enrichment pipelines are the staging layer before CRM.
Documents from parsing, companies from sourcing, and matches from saved monitors all land in enrichment pipelines before anything touches the CRM. This is where rows get resolved, enriched, scored, reviewed, and turned into trusted fields your team can use.
Natural Language Sourcing
Documents & Decks
+ Add columnManage columns Export CSV
| # | Organization | HQ Country | Good Fit? | Entity | Latest Accounts | Raise amount ($m) | Raise stage | Solution | Traction | AlphaLens URL | LinkedIn URL |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Sereact | Enriching... | Enriching... | Enriching... | Enriching... | Enriching... | Enriching... | Enriching... | Enriching... | Enriching... | Enriching... |
Native CRM sync
Mapped to specific fields within Affinity, Attio, or Hubspot.
How enrichment works
Two enrichment sources power the pipeline.
Start with AlphaLens entity data to enrich the deal with known company truth. When the question goes beyond that data, use agents to seek answers from documents, the web, or prior context. Together, entity data and agent answers become the backbone of the pipeline feature.
02
Agent-derived answers
Go beyond entity data when the question needs research.
Agents use the web, documents, or prior row context to answer questions that are not already covered by the entity data layer, with sources and confidence attached.
Entity data foundation
Start every enrichment run from resolved company truth.
AlphaLens keeps the stable facts close at hand: identity, location, funding, growth, product, and people data. Agents should fill gaps, not rediscover the basics from scratch.
Firmographic Data
Company attributes, links and more.
A
Address Data
HQ & branch locations (stree/city/country)
Funding Data
Funding history with investors amounts and stages.
Growth Time Series
Headcount, Web Traffic, Socials & Jobs.
Product Data
What do they sell, core features, and who buys it,
SaaSB2BEnterpriseCloud
People Data
Decision makers and leadership context for qualification.
JDVP
Key decision makers
Agentic answers
Use agents where entity data stops.
Configure typed agent columns, ground outputs in verifiable sources, and return answers your workflow can use immediately.
Configure AI Agent
AI Agent (Web Research) · Step 2 of 2
Column Name *
Defensibility Vector
Web Research Mode
Enable live web search and cite source URLs.
Output Type *
Text Number Boolean Date Select Multi-select
Options *
(a) proprietary data(b) regulatory capture / certification(c) physics IP / patents(d) network effects(e) workflow lock-in(f) capital intensity barrier(g) talent density
Instructions / Prompt *
You are a defensibility analyst for a frontier tech venture fund. Your job is to identify where a company's moat*actually* lives - not where the founders claim it lives. ## Your task
Context Fields *
Active DomainOrganization NameOrganization Descriptionorganization.products.first.product_description
Back
Cancel
Agent question recipe
Agent questions are built from inputs, instructions, and the right agent.
Start with the evidence the agent should use, choose the answer format the workflow needs back, then pick whether the answer should come from documents, web research, or row context alone.
01
Context
Give the agent the row material it should reason from.
Context can come from entity data, a document region, or another AI column that already answered a prior question.
Entity dataDocumentsPrevious agent answer
02
Output format
Choose the answer format the workflow needs back.
Each answer returns in the format you choose, so downstream filters, reruns, review queues, and CRM mappings know how to use it.
TextNumberBooleanSingle selectMulti select
03
Prompt
Write the rough instruction, then improve it.
The Improve Prompt button turns a loose instruction into a structured role, objective, and examples-style prompt the agent can reuse across rows.
Rough promptImprove promptRole / objective / examples
04
Visual Engine
For document-only questions.
Reads decks, PDFs, slides, and document regions, then references the specific parts of those documents that support its output.
DocumentsSlide regionsSource spans
05
Web Agents
For answers that need live web research.
Uses the web to answer the question and cites the specific webpages used as evidence for the returned field.
Web researchCited pagesFresh sources
06
LLM
For reasoning without tool calls.
Answers from the provided context, prompt, and output parameters only. Useful when the relevant evidence is already in the row.
No tool callsRow contextChosen format
Try the full stack
Start with 7 days of AlphaLens, for free.
Use 500 credits to test semantic search, deck parsing, enrichment, and CRM sync before your plan starts.
Workflow
A two-layer enrichment engine for private-market work.
Route saved-monitor matches, parsed documents, and selected companies into enrichment pipelines, configure the fields your team needs, then map clean outputs to the corresponding CRM fields.
Workflow runbook
3 controlled steps from input to usable output
Reviewed Auditable
01Monitors
02Columns
03CRM
01
Monitors
Wire aligned searches into a pipeline.
Set up a dozen saved monitors for the searches that match your thesis, then send every relevant match into the enrichment pipeline of your choice.
02
Columns
Build the pipeline and configure enrichment fields.
Create columns backed by AlphaLens entity data, document context, and AI agent answers, so every row is enriched against the criteria your team actually reviews.
03
CRM
Map clean outputs to the right CRM fields.
Wire each enriched column to the corresponding field in Affinity, Attio, or HubSpot, then sync only the records that are clean enough to trust.
Next step
Explore native CRM integrations
Enrichment is only useful when clean answers reach the system your deal team actually trusts.