Introduction to Askalot

Askalot is the Research Operating System with mathematically verified questionnaires. One platform carries the researcher across the whole lifecycle — ideation → design → campaign → execution → data quality analysis — coordinated by a central Research Paper that every AI agent reads from and writes to. Formal mathematical proof confirms the questionnaire’s logic is sound before fielding; the analysis stage measures and defends the resulting dataset rather than just exporting it.

Who Is Askalot For?

Organizations

• Academic Institutions: Universities and research centers conducting scholarly research
• Market Research Firms: Commercial organizations executing research projects
• Government Agencies: Policy research and public opinion measurement
• NGOs and Think Tanks: Evidence-based research and advocacy organizations

Roles

Role	Responsibility
Questionnaire Designer	Creates survey instruments with proper logic and flow in Armiger
Fieldwork Manager	Coordinates data collection campaigns and participant selection in Targetor
Interviewer	Conducts face-to-face or phone interviews on behalf of respondents in SirWay
Respondent	Completes surveys via magic links or with interviewer assistance
Data Analyst	Processes and exports collected survey data in Balansor

QML: Domain Specific Language for Questionnaires

QML (Questionnaire Markup Language) is a YAML-based domain-specific language designed specifically for AI-assisted questionnaire generation. QML uses declarative syntax with mathematical semantics, allowing AI to focus on question content while SMT solvers handle logical validation.

Why QML?

Current AI tools struggle with complex survey generation because Large Language Models face an overwhelming cognitive burden when attempting to simultaneously:

• Generate semantically meaningful questions
• Track conditional dependencies between questions
• Ensure logical consistency across all paths
• Maintain correct question ordering and flow
• Avoid contradictions and unreachable dead ends

QML solves this by separating concerns: AI generates questions, SMT validates logic:

• Declarative Syntax: Questions are defined with preconditions and postconditions, not sequential flows
• Mathematical Semantics: Every QML construct has formal mathematical meaning verified by Z3
• Separation of Concerns: AI generates meaningful questions while SMT solvers ensure logical consistency
• Human-Readable Format: YAML with diagram based navigation makes questionnaires accessible to non-programmers

This separation relieves AI from tracking complex conditional dependencies, logical consistency, and flow ordering—allowing it to focus solely on generating semantically meaningful questions while mathematical validation handles the complexity.

See QML Syntax for detailed documentation.

Survey Presentation Format

Askalot delivers each survey question as a universal JSON document for rendering, making integration into any application seamless. A Questionnaire is the verified template; a Survey is one respondent’s session through it — the Presentation Format always describes the next question of a live Survey, computed from the respondent’s previous answers. Each step carries all necessary information to render the question:

• Caption and Text: The question caption plus optional descriptive text
• Input Control Specification: Type of control (text, radio, checkbox, dropdown, slider, switch) with its configuration
• Render-Ready Options: Valid answer options pre-extracted from the control — integer answer codes paired with display labels
• Server-Side Validation: Answers are checked against QML postconditions on submission — rejected answers return the question’s hint message

Framework-Agnostic Design

This JSON representation makes integration dead simple because:

1. No Predefined UI Components: Applications use their own UI framework’s components
2. No Custom CSS Required: Styling matches your existing application design
3. Brand Consistency: Questions naturally adopt your application’s look and feel
4. Framework Flexibility: Works with React, Vue, Angular, vanilla JS, or any frontend technology

This approach enables any application to render surveys using its native components while maintaining full logical validation from the QML engine. See Survey Presentation Format for the complete format reference and flow loop.

How It Works Together

The four services form an integrated workflow, coordinated by the Research Paper:

1. Ideate & Design in Armiger: AI ingests source documents and natural-language objectives, drafts the Research Paper, then generates a QML questionnaire grounded in the paper; SMT validates logical consistency
2. Campaign in Targetor: Import target audience, configure demographic parameters and sampling strategy aligned with the paper, launch campaign
3. Execute in SirWay: Respondents walk the flow that was verified at publication, with dynamic navigation; persona agents can simulate responses for testing
4. Data Quality Analysis in Balansor: Rake to correct sampling bias, score response quality (straightlining, speeder, Cronbach’s α), measure representativeness; the Analyst agent reads the paper to interpret findings against the original research objectives

All services share a common authentication system, a common data repository, and a common Research Paper — enabling a coherent workflow rather than a relay between disjoint tools.

Roundtable’s research flow for a Demo project: every stage, from goals in the Research Paper to the paper chapter written from the measured data, with a link to the service where it happens.

Key Benefits

For Survey Designers

• AI assistance: Generate questionnaires from natural language descriptions
• Error prevention: Mathematical validation catches logical flaws automatically
• No manual dependency tracking: System handles question order and conditional logic
• Confidence: Know that deployed surveys are logically sound

For Respondents

• Smooth experience: Questions follow a flow whose logic was checked before fielding
• Relevant questions: Only see questions that apply to their situation
• No unreachable questions: A question nobody could reach is caught before the survey is fielded
• Dynamic adaptation: Survey adapts based on previous answers

For Analysts

• Quality data: Logical validation ensures response consistency
• Representative samples: Targetor enables demographic balancing
• Statistical correction: Balansor adjusts for sampling bias
• Standard formats: Export to familiar analysis tools

Formal Verification in Practice

Traditional survey platforms require designers to manually ensure:

• Questions appear in the correct order
• Conditional logic doesn’t create contradictions
• All questions are reachable
• Validation rules are mutually consistent

Askalot automates all of this through mathematical verification. Here’s a simple questionnaire example:

The SMT solver automatically verifies:

1. Reachability: Green nodes (ALWAYS reachable), yellow nodes (CONDITIONAL on preconditions)
2. Preconditions: Rounded nodes show when questions appear (e.g., consent only if age ≥ 18)
3. Postconditions: Curly nodes validate answers (e.g., children < household size)
4. Consistency: All constraints are simultaneously satisfiable

This happens instantly during design in Armiger, not after deployment when problems are discovered by confused respondents.

Next Steps

• Quick Start

  ---

  Follow the complete workflow from design to analysis

  → Start Now

• Creating Surveys

  ---

  Learn QML syntax and best practices

  → Guide

• Mathematical Theory

  ---

  Understand the formal foundations

  → Theory

• API Reference

  ---

  Integrate Askalot into your applications

  → API Docs