Grounded in published science. Verified through discovery.
AutoPraxys bridges peer-reviewed human–AI interaction literature with empirical behavioral observation. We separate measurable interaction signals from unverified causal assertions.
Independent Academic Research on Human–AI Interaction
Our product thesis is anchored in rigorous published findings on cognitive offloading, overreliance, and the jagged technological frontier. Note: External published studies document broader human–AI behavioral challenges; they do not claim to validate or endorse the Praxys platform specifically.
Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality
Dell'Acqua, F., McFowland, E., Mollick, E. R., Lifshitz-Assaf, H., Kellogg, K., Rajendran, S., Krayer, L., Candelon, F., & Lakhani, K. R. (2023).
“For tasks outside the frontier, AI decreased consultants' performance, with consultants using AI being 19 percentage points less likely to produce correct solutions compared to those without AI.”
Relevance to Praxys: Demonstrates the 'jagged frontier' effect: workers frequently overestimate AI capabilities on complex tasks, accepting flawed answers without critical verification.
Experimental evidence on the productivity effects of generative artificial intelligence
Noy, S., & Zhang, W. (2023).
“Generative AI technology increased productivity: average time to complete tasks dropped by ~40%, and output quality rose by ~0.45 standard deviations... AI compressed productivity inequality, but raises critical questions regarding skill development and independent problem-solving.”
Relevance to Praxys: Confirms that while AI dramatically accelerates production speed, it simultaneously introduces dependencies on foundational cognitive stages.
Overreliance on AI Systems: Path Toward Addressing Overreliance
Passi, S., & Vorvoreanu, M. (2022).
“Overreliance occurs when people uncritically accept AI recommendations, failing to identify errors even when they possess the requisite domain knowledge to do so.”
Relevance to Praxys: Directly describes the 'Capability Illusion' observed in AutoPraxys discovery interviews: users with sufficient training overlook errors due to cognitive offloading.
Initial Field Discovery Cohorts
Before engineering software, we conducted semi-structured qualitative interviews to understand how AI tools change working behavior.
STEM Students Interviewed
Qualitative discovery sessions across university cohorts assessing prompt iteration habits and independent problem articulation.
University Professors Consulted
Computer science and engineering faculty evaluating learning visibility, academic integrity, and pedagogical assessment challenges.
Industry Engineering Leads
Senior software engineering managers assessing junior developer code comprehension and code-review confidence.
How Praxys Compares to Existing Tool Categories
Different software categories serve different purposes. Praxys does not replace AI assistants; it introduces the missing behavioral observation layer.
| Category | Interaction Telemetry | Pre-Dispatch Assurance | Friction Analysis | Socratic Coaching | Privacy Boundary |
|---|---|---|---|---|---|
| Consumer AI Interfaces (Web) | Session length & text log only | Post-hoc filter / provider discretion | None (individual chat interface) | None (answers provided directly) | Cloud-hosted; provider terms govern training |
| Developer Copilot Extensions | Acceptance rate & active keypress | Public code match detection | Basic acceptance percentage | None (completion-oriented) | Telemetry governed by enterprise license tier |
| Traditional LMS Platforms | Submission timestamp & grade only | Plagiarism text similarity | None (static submission portal) | Summative test scores | Institutionally hosted student records |
| Praxys Intelligence Platform | Interaction cadence & edit physics | Pre-dispatch regex & entropy policy scans | Loop detection & friction diagnostics | Socratic reflection & autonomy calibration | Metadata-first; client-side isolation by default |
* Analysis based on documented public product documentation and API specifications as of Q4 2026. Enterprise tiers for frontier providers offer robust SOC2 and zero-data-retention guarantees for server-side processing; Praxys operates client-side at the interaction interface.
AutoPraxys Research Brief: Behavioral Observation Protocol
Our methodology working paper detailing client telemetry normalization, metric calculation indices, and early pilot observations is currently undergoing peer review with academic advisors.