Purpose: Internal overview of all functionality in Abstract’s first production release.
Summary:
Version 1 introduces the foundational architecture of Abstract, a vertically-integrated AI orchestration platform for professionals working to understand shifts in public policy. It combines customer-specific context (private documents) with a constantly updated “public footprint” (AI-curated external data) to generate dynamic, reasoned policy insights and strategic recommendations.
1. Data Context Layer
Uploads
Purpose: Establishes the customer’s private context — their internal documents, client materials, and reference files.
Core capabilities:
Upload documents (currently optimized for PDFs; other formats in development).
Add tags for organization and search.
Rename, filter, and view uploaded files.
Definition:
Private Footprint – customer-provided materials that describe the organization’s operations, interests, and sensitivities.
Public Footprint – AI-generated intelligence on the organization (or its clients), created by autonomous agents scanning and summarizing relevant public web data.
How it works: The combination of both footprints becomes the contextual foundation for all subsequent reasoning — every chat, report, and analysis in Abstract draws from this composite knowledge base.
2. AI Interaction Layer
Chat Environment
Users can interact with Abstract in two primary modes:
General Chat
Open-ended conversational mode.
Allows users to ask questions about:
Their uploaded documents
New or proposed public policy developments
Strategic or legal implications for their business
Functions similarly to ChatGPT, but grounded in the customer’s context and relevant public policy data.
Predefined “Abstract Reports”
Abstract currently ships with three structured workflows mirroring the standard cadence of policy intelligence and government affairs work:
Policy Discovery (“WHAT changed”)
Detects new or updated policies that could affect the customer.
Impact Analysis (“HOW it impacts us”)
Analyzes in depth how a given bill, rule, or executive action affects the organization or its clients.
Strategy Report (“WHAT to do next”)
Produces forward-looking recommendations and tactical plans for engagement.
Each chat session automatically displays the customer’s purchased data coverage — i.e., the jurisdictions, agencies, and governing bodies Abstract continuously monitors for them.
3. Report Workflows
3.1 Policy Discovery Report
Purpose: Identify relevant legislation, rules, or proposed rules.
Inputs: Jurisdictions, timeframe, and optional issue focus (e.g., “labor” in “California” over “last six months”).
Process:
Abstract cross-references the customer’s private footprint and public footprint.
It searches all monitored bodies within the selected jurisdictions.
Returns a ranked list of policies using a proprietary relevancy score (most to least relevant).
Output:
For each item: status, summary, and “Why Abstract Discovered This” explanation.
Users can request deeper analysis or extend reasoning to additional policies.
3.2 Impact Analysis Report
Purpose: Provide a comprehensive breakdown of how a specific bill or rule affects the customer.
Structure:
Basics: Summary and provisions.
Policy Context: Current law, status-quo changes, prior attempts.
Stakeholder Analysis: Impacted parties, winners/losers, supporters, opponents, and opportunities to neutralize.
Political Landscape: Sponsor motivation, vote math, feasibility, and executive stance.
Fiscal & Operational Impact: Budget effects, implementation timelines.
Legal & Regulatory Analysis: Constitutional issues, agency capacity, rulemaking implications.
Risks & Opportunities: Unintended consequences, amendment leverage points, trends.
Recommendations: Positioning, rationale, and coalition engagement plan.
Notes: These reports take several minutes to generate, as they perform deep multi-source reasoning and synthesis across data layers.
3.3 Strategy Report
Purpose: Translate insights into concrete next actions and messaging.
Structure:
Bill or Rule Summary
Position Statement: Monitor, support, or oppose.
Strategic Objectives: Legislative, regulatory, business-protection, and relationship goals.
Stakeholder & Political Mapping: Allies, opponents, skeptics, and potential coalition partners.
Messaging & Narrative: Key talking points and recommended framing.
Tactical Plan: Engagement sequence, harm mitigation, timeline, KPIs, and contingency planning.
Usage: Often follows an Impact Analysis report; provides the client-ready output for action planning.
4. General Chat Enhancements
Users can refine or extend any report in General Chat (e.g., draft letters, emails, or summaries).
Supports copy, regenerate, thumbs-up/down feedback, and iterative refinement.
5. Collaboration & Access Control
Groups and Organizations
Each organization can contain multiple groups (e.g., practice areas, clients, or departments).
Each group:
Has its own private documents and chat history, isolated from others.
Shares the organization-wide public footprint.
Is managed by an admin, who controls group membership and access rights.
Help Center
Accessible via bottom-right intercom widget.
Routes to Abstract’s support documentation and chat.
6. Orchestration and Data Architecture
AI Orchestration Layer
Abstract orchestrates multiple AI agents across:
Proprietary scrapers and data monitors for legislative and regulatory updates.
Constantly refreshed feeds from state, federal, and local entities.
Why this matters: Much of this source data isn’t reliably available for purchase or subscription — Abstract structures it in real time.
Future integrations (in development):
Contact information databases to identify and reach government staffers.
Extended data sets (e.g., lobbying disclosures, fiscal notes, etc.) to further contextualize recommendations.
7. Summary
Abstract v2.0 delivers a complete, end-to-end workflow to:
Ingest contextual data (uploads + public footprint).
Monitor policy developments across all subscribed jurisdictions.
Generate automated discovery, analysis, and strategy reports grounded in the organization’s unique context.
Enable secure, role-based collaboration within and across teams.
It lays the groundwork for AI-orchestrated policy intelligence — moving from static research toward dynamic, contextual decision-support.