How we use “IA”
From here on we write Intelligent Assistance (IA). We prefer that phrase because it keeps human agency primary: assistance that cites, declines, or asks—not autonomy that replaces responsibility.
Explico® Laboratory
Explico® Laboratory is a small, public statement of how we evaluate Intelligent Assistance (IA)—and how we work with others who want controlled trials instead of hype. We are building a product company, not a university lab; this page is honesty about what we measure and what we still have to prove.
We care about reducing reliance on resource-heavy guessing without throwing away what modern assistance tools do well. In practice that means repeatable runs through controlled environments, then comparing labeled experimental builds against the same frozen outcomes—so partners can see what changed and what did not. When something earns its keep, it can move into the product path; until then it stays in the lab. We also design for shared service and, over time, Laboratory research on hardened deployments for regulated workflows—we report cautiously, not finished promises.
The Explico® Intelligent Assistant is meant for depth over dazzle: faster paths to understanding, with critical thinking and authorship staying with people. Our first core tenet is security first—zero-trust segmentation and encryption on the wire—then applications designed into that posture, not bolted on afterward. We do not train on your prompts, documents, or work product.
Who leads the laboratory work
Morpheus (a builder name used for more than thirty years) leads engineering direction for Explico®. The laboratory pages describe intent and evaluation norms—not a large staff or an academic department. Off the keyboard, the same person is a husband, parent, and neighbor who would rather ship careful systems than perform biography on a web page.
Precision prompts, deterministic answers.
Explico® is Latin for we explain. The Explico® Intelligent Assistant (IA) is built for zero-trust, privacy-first work: precise questions, scoped context, and answers drawn from human-curated knowledge domains—not open-ended guessing. Secure interactions are not used to train Explico® on your data or to alter signed domain content without curator review; when information is missing, the assistant asks for what it needs so every explanation stays sourced and transparent. The same well-formed question is built to return the same result—repeatable in a meeting, defensible in a memo, trustworthy under review—not a paragraph that merely sounds right.
We want IA to be seen as a partner beside people—not a substitute for judgment. The laboratory is where we run controlled comparisons, share results with willing partners, and only then promote what survives review.
What the laboratory is (and is not)
- Is: a public write-up of evaluation principles and partnership norms; hands-on product exercise stays behind authentication and entitlements.
- Is: invite-only trials and side-by-side tests with teams who accept clear scopes, notices, and declines—not open-ended data collection.
- Is: a way to compare experimental designs on the same benchmarked outcomes, so evolution is measurable.
- Is not: a claim that we already operate at production scale for every workload—we report what we have measured.
- Is not: a substitute for your institution’s IRB, counsel, or security review; we aim to lower practical barriers to well-scoped research trials when both sides agree on controls.
- Over time we hope educator–curators and industry Subject Matter Experts (SMEs) will help domains evolve; that workflow is early and entitlement-gated.
Responsible iteration
When external references are needed, cited lookups are handled through controlled, cache-first services with explicit attribution. That material may supplement assisted answers; it does not replace curator-governed knowledge. Unvetted or disputed material stays visibly marked until reviewers accept, modify, or reject it.
We respect the intellectual property of everyone who uses the solution; how we handle data, retention, and prohibited uses is set out in our Privacy Policy and Acceptable Use Policy (see also Terms of Service, including the no-training commitments).
During assisted reference workflows, we do not capture your prompt text or customer work product for model training. We may use aggregated, minimized metadata from retrieval activity to spot gaps in knowledge domains—candidates for curator review toward stronger deterministic coverage—not a record of why you needed a lookup. Curators and delegated approvers work under phishing-resistant authentication; promoting configuration or content to production requires deliberate, audited confirmation.
For partnership and US trial timing, see Collaboration & trials. For what we measure, see What we measure. To begin, use Contact on the main site for an introductory video call—not an open self-serve lab signup.
← About overview · Sign in to the portal (verified accounts)