AI Features - Askalot.io | AI Agents for Survey Research

Designer Agent · Armiger

One agent. Two roles. One validated questionnaire.

The Designer Agent operates in two distinct phases: first as a researcher analyzing your documents and extracting measurable concepts, then as a questionnaire architect generating mathematically validated QML. You control the transition.

Role 1: Research

Document analysis & concept extraction

Ingest research materials (PDF, Word, spreadsheets, regulations)
Semantic indexing via pgvector for intelligent retrieval
Extract quantifiable dimensions and research themes
Produce a structured research document for review

You approve the research document

Role 2: QML Generation

Questionnaire design & formal verification

Generate QML from the approved research document
Create questions with appropriate controls and skip logic
Z3 SMT solver validates reachability and logical consistency
Iterative loop: generate → validate → fix → re-validate

armiger.dev.askalot.io

Armiger - AI-powered questionnaire design IDE with Designer Agent research and generation roles

Both roles operate through the same chat interface in Armiger

Designer Agent · Research Role

Upload documents. Get a structured research brief.

In its research role, the Designer Agent ingests your source materials — research papers, compliance frameworks, clinical guidelines, policy documents — and produces a structured research document that captures all measurable concepts and their relationships.

Document ingestion — PDF, Word, spreadsheets, and plain text
Semantic indexing — Documents chunked and embedded via pgvector for intelligent retrieval across large document sets
Concept extraction — Identifies quantifiable dimensions, measurement themes, and response scales
MCP document tools — Search, list, and summarize indexed documents conversationally
Project isolation — Indexed documents scoped to your active project, ensuring team-level data separation

Market Research Clinical Due Diligence Risk & Compliance IT Security Procurement

Research Role Pipeline

1

Document Upload

PDF, Word, spreadsheets, regulations

2

Semantic Indexing

Chunked, embedded, stored in pgvector

3

Concept Extraction

Measurable dimensions, themes, scales

Research Document

Structured brief ready for your approval

customer-satisfaction.qml

# Generated by Designer Agent
questionnaire:
  name: "Customer Satisfaction Survey"
  items:
    - id: overall_satisfaction
      type: question
      text: "How satisfied are you overall?"
      control: likert_5

    - id: dissatisfaction_reason
      type: question
      text: "What could we improve?"
      control: open_text
      precondition: "overall_satisfaction <= 2"

# SMT Validation: PASSED
# All items reachable
# All postconditions satisfiable
# No contradictions found

Z3 SMT Validation Passed

2 items verified · all reachable · 0 contradictions

Designer Agent · QML Generation Role

Generate questionnaires. Validate with mathematics.

In its generation role, the Designer Agent takes your approved research document and generates complete QML questionnaires with conditional logic, skip patterns, and appropriate response controls. Every questionnaire is then mathematically verified by the Z3 SMT solver.

QML generation — Complete questionnaire with blocks, skip logic, and validation rules
8+ control types — Likert scales, multiple choice, open text, numeric, date, matrix, and more
Z3 SMT validation — Proves every question is reachable and all logic paths are sound
Auto-correction — Agent fixes validation errors and re-validates until all checks pass
MCP integration — Agent calls validation, file listing, and document tools via MCP protocol

Launch Armiger

Campaign Manager Agent · Targetor

Set up campaigns with the Campaign Manager Agent.

The Campaign Manager Agent guides you through campaign setup via two interfaces: a step-by-step wizard for beginners and a conversational chat for power users. It recommends sampling strategies, proposes demographic distributions, and warns about potential biases.

1

Campaign design — name, questionnaire, mode

2

Target audience — demographics, sample size

3

Sampling strategy — stratification factors, distributions

4

Respondent pool — generate and preview quality

5

Review & launch — summary, validation, deployment

Conversational chat for power users: merge pools, compare rates, create custom segments
AI recommends stratification factors based on research topic
RMSE quality scoring measures pool representativeness

Launch Targetor

targetor.dev.askalot.io

Targetor - AI-assisted campaign management with sampling strategy configuration

roundtable.dev.askalot.io

RoundTable - AI persona simulation with demographic-based response generation

Respondent Agent · SirWay

Simulate campaigns with Respondent Agents.

Respondent Agents adopt demographic personas and complete surveys with life-like responses — for testing questionnaire flow, validating analysis pipelines, or generating synthetic datasets. Three simulation modes cover every use case.

Single

One survey, one persona

Campaign

Full 3-phase lifecycle

Mass Fill

Bulk synthetic data

Responses shaped by respondent demographics and persona profiles
Three-phase flow: Preparation → Execution → Analysis
Test analysis pipelines (Bronze → Silver → Gold) with synthetic data

Young Professional Married Family Retired Senior High-Income Executive Random

Launch SirWay

Semantic Clustering · Balansor

Discover themes in open-ended responses automatically.

Semantic clustering turns free-text survey responses into structured, analyzable themes. Embeddings capture meaning — not just keywords — so similar ideas group together even when worded differently. Choose keyword extraction or AI-powered labeling to name each cluster.

Multi-theme detection — Responses can belong to multiple themes with configurable similarity thresholds
Two labeling methods — Keywords via c-TF-IDF for speed, or AI-generated labels via LLM for natural descriptions
Gold-ready output — Creates indicator columns for each theme, ready for cross-tabulation and export
Medallion pipeline — Bronze → Silver (weighted) → Gold (publication-ready) with full lineage

Launch Balansor

balansor.dev.askalot.io/clustering

Balansor - Semantic clustering of open-ended survey responses with multi-theme detection and LLM labeling

Your AI provider. Your choice.

No vendor lock-in. Use premium cloud models with your own API key, or use our free built-in local models at no cost. Switch providers anytime without changing your workflow.

A

Anthropic

Claude Sonnet, Claude Haiku

Default provider

O

OpenAI

GPT-4, GPT-4o

Local AI

Ollama, Llama, Mistral

Free — hosted on our servers

G

Google

Gemini Pro, Gemini Ultra

80+ tools. One unified API.

Every platform capability is accessible via REST API and Model Context Protocol (MCP). AI agents from any provider can manage your research autonomously through programmable tools.

5 Project Management

5 Questionnaires

9 Campaigns

6 Respondents

8 Pools

6 Strategies

4 Workload

8 Surveys

2 Users

8 Datasets

3 Audit

3 Documentation

MCP Tool Call

// AI agent validates a questionnaire via MCP
validate_qml_file({
  qml_name: "customer-satisfaction.qml"
})

// Then creates a campaign and generates a pool
create_campaign({
  name: "Q1 National Survey",
  project_id: "proj_abc123",
  questionnaire_id: "qst_def456"
})

Ready to let AI transform your research?

Specialized agents, mathematical validation, and 80+ programmable tools — from research documents to publication-ready data.

Create Free Account Quick Start Guide

Specialized AI Agents for Every Stage of Research