Askalot — Investor Deck¶
Current State Document — February 2026
One-Liner¶
Askalot is developing an agentic survey platform to help research organizations manage the full research lifecycle — from questionnaire design to publication-ready data — with AI agents orchestrating 80+ tools and mathematical validation.
The Problem¶
Complex survey research is broken. Enterprise questionnaires (compliance assessments, clinical trials, market research) routinely contain 100–350 questions with intricate conditional logic. The industry still relies on manual flow design and manual testing.
| Pain Point | Impact |
|---|---|
| Development + QA cost per complex survey | \(10,000–\)50,000 |
| Time spent on manual flow testing | 2–4 weeks |
| Data pollution from logic errors | 5–15% |
| AI survey generation ceiling | ~20 simple, linear questions |
No major platform — Qualtrics, SurveyMonkey, Typeform — offers automated validation of survey logic. They all require manual branching, manual testing, and accept the resulting errors.
Source: Founder's direct experience as former Head of Data Collection at Szonda Ipsos (Hungary's largest market research firm).
The Solution¶
Core Innovation: Declarative, Flowless Design¶
Traditional platforms require manually wiring skip logic branches. Askalot inverts this: authors declare preconditions on each question, and the engine dynamically computes valid traversal paths at runtime.
This architectural decision enables two things no competitor offers:
-
Mathematical Validation — The Z3 SMT solver (Microsoft Research) formally proves that every question is reachable, every path is valid, and no contradictions exist. This is not testing; it is proof.
-
AI-Safe Generation — LLMs only need to generate question content and preconditions. They no longer need to reason about ordering, branching, or flow consistency. The solver handles that. This removes the ceiling on AI-generated survey complexity.
Questionnaire Markup Language (QML)¶
Askalot defines a domain-specific language for survey specification. QML files are declarative, version-controllable, and machine-readable. They serve as the single source of truth consumed by the AI design tools, the runtime engine, the validator, and the analysis pipeline.
What Is Built Today¶
Askalot is a working, multi-tenant platform — not a prototype. Six microservices are built, integrated, and fully operational in the test environment. Production deployment is in progress.
Platform Services (Deployed)¶
| Service | Function | Status |
|---|---|---|
| Armiger | AI-powered questionnaire design IDE with browser-based code editor, Z3 validation, and flow visualization | Complete |
| Targetor | Campaign management: sampling strategies, demographic targeting, respondent pools, interviewer workload distribution | Complete |
| SirWay | Survey execution engine with lazy evaluation, magic-link access, interviewer-assisted mode (phone & in-person), and AI campaign simulation | Complete |
| Balansor | Statistical analysis: Bronze/Silver/Gold medallion pipeline, post-stratification raking, multi-format export (CSV, SPSS, Excel, Parquet) | Complete |
| Portor | Unified API gateway exposing 80+ tools via REST and Model Context Protocol (MCP) | Complete |
| Roundtable | Organization landing page, user management, AI agent simulation orchestration | Complete |
Platform Infrastructure (Deployed)¶
| Component | Function |
|---|---|
| OIDC Identity Provider | Global SSO with OAuth (Google, Microsoft), privacy-preserving tenant discovery |
| Notification Service | Email delivery with magic-link survey invitations |
| Audit Service | Full event logging with InfluxDB time-series storage |
| Monitoring Stack | Prometheus, Grafana, Loki, Tempo, Pyroscope |
| Documentation Site | Public docs at docs.askalot.io (MkDocs Material) |
| Docker Registry | Private image registry with Buildx CI |
Infrastructure Architecture¶
- Proxmox hypervisor with LXC containers for isolation
- Multi-tenant: Each organization gets dedicated resources (separate database schema, custom domain)
- Two-tier routing: Traefik (HTTPS termination) → NGINX (per-tenant) → Services
- PostgreSQL with PgBouncer connection pooling
- Redis Streams for inter-service messaging
Key Technical Capabilities¶
| Capability | Detail |
|---|---|
| Z3 SMT validation | Reachability, contradiction detection, cycle detection, dead-end analysis |
| Lazy evaluation engine | Questions computed dynamically from preconditions — no static flow graph |
| AI questionnaire generation | Natural language or document ingestion → QML → validate → iterate |
| Sampling strategies | Greedy and random-constrained algorithms with RMSE quality scoring |
| Post-stratification weighting | Iterative proportional fitting (raking) to align sample demographics |
| AI campaign simulation | 5 built-in persona profiles (young professional, married family, retired, high-income, random) |
| Multi-provider AI | Claude, GPT-4, Bedrock, Gemini — no vendor lock-in |
| MCP integration | 80+ tools accessible to any MCP-compatible AI agent |
| Multi-format export | CSV, Excel (.xlsx), SPSS (.sav), Parquet with full metadata |
| Audit trail | Every action tracked with entity history and actor identification |
What Is Not Built Yet¶
| Gap | Description | Priority |
|---|---|---|
| Payment/billing | No subscription management or payment processing | High — required for revenue |
| Self-service onboarding | Organization provisioning is manual (CLI script) | High — required for growth |
| Regulatory template libraries | No pre-built SIG, CAIQ, ESRS, Wolfsberg questionnaire packs | Medium — accelerates enterprise sales |
| Mobile-optimized survey UI | SirWay is responsive but not mobile-first | Medium |
| Advanced analytics | AI-assisted insight generation in Balansor | Low — roadmap item |
| Performance testing | No load testing suite (Locust plan exists) | Medium |
| SOC 2 certification | Security controls in place, formal audit not started | Medium — enterprise requirement |
Market¶
Size¶
- Online survey software market (2023): ~$3.6 billion (Grand View Research)
- Growth rate: 13–15% CAGR
- Projected (2030): ~$9 billion
- Critical data still needs to be collected directly from people — not sensors or APIs
Target Verticals (Ordered by Entry Priority)¶
| # | Vertical | Why Now | Questionnaire Complexity |
|---|---|---|---|
| 1 | Third-Party Risk Management (TPRM) | 43% of companies use multiple assessment questionnaires; SIG/CAIQ standards expanding | 101–350 questions |
| 2 | ESG/CSRD Compliance | CSRD requires 1,200+ data points; ESRS has 82 disclosure requirements | Hundreds of indicators |
| 3 | AI Governance | EU AI Act phased obligations through 2027; NIST AI RMF adoption | Emerging — first-mover opportunity |
| 4 | Financial Services KYC/AML | Wolfsberg questionnaires standard in bank-to-bank due diligence | Complex risk-tiered branching |
| 5 | Clinical Trials (ePRO/eCOA) | $2B+ market (2025) at 13–16% CAGR; FDA requires validated instruments | Longitudinal adaptive protocols |
| 6 | Academic Research | Complex methodology, grant-driven timelines, institutional buying | Varies widely |
Competitive Landscape¶
| Feature | Askalot | Qualtrics | SurveyMonkey | Typeform | LimeSurvey |
|---|---|---|---|---|---|
| Flow design | Automatic pathfinding | Manual branching | Manual skip logic | Linear only | Manual conditions |
| Logic verification | Mathematical proof (Z3) | Manual testing | Basic validation | None | Manual testing |
| AI generation + validation | Integrated loop | Limited AI tools | Basic AI assist | None | None |
| Complex logic support | Unlimited (functional) | Advanced but limited | Basic | Very limited | Moderate |
| Deploy time (complex survey) | 3–5 days | 2–4 weeks | 1–3 weeks | Not suitable | 2–3 weeks |
| QA automation | 100% automated | Manual | Manual | Manual | Manual |
Moat: Z3 integration + declarative QML + agentic AI loop is 2+ years of specialized engineering. No competitor has attempted this combination.
Business Model¶
Pricing (Target — USA Market)¶
| Tier | Price | Target Customer |
|---|---|---|
| Standard | \(500–\)1,000/month | Research teams, small agencies |
| Enterprise | \(20,000–\)50,000/year | Corporations, compliance-driven organizations |
Revenue Model¶
Subscription SaaS with per-organization tenancy. Revenue drivers: - Platform subscription (recurring) - AI usage metering (future — bring-your-own-key initially) - Template marketplace (future — regulatory questionnaire packs)
Traction¶
Company¶
Askalot Survey & Research Kft. — incorporated in Hungary.
Environment Status¶
| Environment | Status |
|---|---|
| Test | Fully operational — all services deployed and functional |
| Production | In progress — infrastructure provisioning underway |
Pilot Partners (Signed, Not Yet Started)¶
| Partner | Type | Agreement |
|---|---|---|
| Zavecz Research | Hungarian market research firm | Pilot commitment signed |
| ELTE University | Faculty of Informatics | Research-use agreement in principle with Dean |
Pilots will begin once the production environment is operational.
Development Milestones Achieved¶
- Full multi-tenant platform built and deployed in test environment
- 80+ MCP tools accessible to AI agents
- Z3 mathematical validation integrated end-to-end
- AI questionnaire generation with iterative validation loop
- Campaign simulation with demographic persona profiles
- Medallion data pipeline with post-stratification weighting
- OIDC-based SSO with OAuth provider support
- Public documentation site live
Academic Foundation¶
- The founder's master's thesis on formal verification of questionnaire logic using SMT solvers was presented at ELTE's TDK (Tudományos Diákkör — Hungary's national scientific student research competition) and received high ranking.
- The mathematical foundation has been accepted for presentation and scientific paper at ICAI 2026 (13th International Conference on Applied Informatics, Eger, February 2026), with proceedings indexed in Scopus.
Team¶
Founder¶
Peter Saghelyi
- 25+ years in software development, AI, and research methodology
- Former Head of Data Collection at Szonda Ipsos (Hungary's largest market research firm at the time)
- Built the entire platform solo: 6 microservices, shared modules, infrastructure, AI integration, documentation
- Master's thesis on formal verification of questionnaire logic (ELTE University)
Team Gaps (To Fill with Investment)¶
| Role | Purpose | Priority |
|---|---|---|
| Backend/Full-stack Developer | Core platform development, AI integration | Immediate |
| Infrastructure/SRE Engineer | Platform operations, scaling, security, monitoring | Immediate |
| Product Manager | Roadmap validation, feature prioritization, customer feedback | 6 months |
| Sales/Business Development | Enterprise customer acquisition, US market entry | 6 months |
Investment¶
Amount Sought¶
\(600,000–\)1,500,000 USD (200–500 million HUF)
Use of Funds¶
| Category | Allocation |
|---|---|
| Team (18–24 month runway) | 60–65% |
| Infrastructure (US-hosted cloud) | 15–20% |
| Legal, compliance, company formation | 5–10% |
| Marketing and demand generation | 10–15% |
Equity Offering¶
| Investment | Equity | Implied Pre-Money Valuation |
|---|---|---|
| $600,000 | 7% | $7.9M |
| $900,000 | 10% | $8.1M |
| $1,200,000 | 13% | $8.0M |
| $1,500,000 | 15% | $8.5M |
Company¶
Askalot Survey & Research Kft. is incorporated in Hungary. IP is company-owned.
Milestones¶
12 Months — Market Entry¶
- Production environment live, US cloud infrastructure provisioned
- Pilot programs with Zavecz Research and ELTE University completed
- First paying US customers (pilot projects)
- 5–10 active enterprise trials
- Self-service onboarding operational
- 1–2 regulatory template packs (TPRM, ESG)
24 Months — Enterprise Growth¶
- 20–30 enterprise customers
- \(500,000–\)800,000 ARR (pre-bookings)
- Product-market fit validated in 2–3 verticals
- SOC 2 Type II certification initiated
36 Months — Scale¶
- 50–80 clients
- $1.5–2.5 million ARR
- Expansion revenue from existing customers
- Customer retention >85%
Exit¶
Timeline: 5–7 years¶
Target Acquirers¶
Large survey/research platforms seeking mathematical validation and AI-native architecture:
- Qualtrics (SAP)
- SurveyMonkey (Momentive)
- Typeform
- Enterprise compliance platforms (OneTrust, Whistic)
Target Exit Value: $100–150 million¶
Based on: - Year 5–7 ARR of $15–25 million (assumes Series A/B follow-on funding for sales scaling) - 5–7x ARR multiple (median-to-above-median for B2B SaaS with defensible technology) - Strategic premium for mathematical validation IP + AI integration
Investor Returns¶
| Investment | Equity | Return at $100M | Return at $150M | Multiple |
|---|---|---|---|---|
| $600K | 7% | $7M | $10.5M | 7–10x |
| $900K | 10% | $10M | $15M | 8–12x |
| $1.2M | 13% | $13M | $19.5M | 9–14x |
| $1.5M | 15% | $15M | $22.5M | 10–15x |
Risks¶
| Risk | Mitigation |
|---|---|
| Solo founder / key-person dependency | Codebase is well-documented (CLAUDE.md at every level); first hire is a developer |
| No revenue yet | Working product with pilot partners; investment focused on commercialization, not R&D |
| US market entry from Hungary | Remote-first company; US cloud infrastructure; founder has international experience |
| Enterprise sales cycle length | Start with mid-market; regulatory urgency (CSRD, AI Act) shortens cycles |
| Competitor response | 2+ year technical moat; competitors would need to rebuild their flow architecture |
Summary¶
Askalot is a working platform with an incorporated company — not a slide deck. The core technology (declarative QML + Z3 validation + agentic AI) is built, integrated, and fully operational in the test environment, with production deployment in progress and two pilot partners signed. The investment is for commercialization: building the team, entering the US market, and converting technical capability into revenue.
The founder built the entire platform — six microservices, shared infrastructure, AI integration, academic thesis — alone. With a small team and focused go-to-market, the path from working product to paying enterprise customers is short.