BMT-01.07 Executive Summary#
BlueMirror.tech | May 2026#
Margaret Chen has noticed something she has not told her daughter. Twice in the past month, she has walked into a room and lost the thread of what she came in for. Once, last Wednesday, the loss persisted for what felt like several minutes, although she suspects it was less. She felt unsteady afterward, the kind of unsteadiness that follows a moment of recognizing oneself unaccountably out of step with one’s own intention. She has not mentioned it to anyone. Her own mother had Alzheimer’s. Margaret knows what this might be. It might be nothing. It might also be the beginning of something. She is not ready to find out, and she is not ready to begin the conversation with her daughter that finding out would entail. So she does what most people in her position do: she watches herself, quietly, with a vigilance she keeps hidden from the people she loves.
The cognitive concierge is the agent that serves Margaret across this trajectory. It is the most ethically complex agent in the BlueMirror system because it serves people whose capacity to consent to being served is changing. The architecture must adapt to her cognitive state without requiring her to acknowledge the change. It must protect her dignity while providing support. It must involve her family at the right moments, in the right ways, without infantilizing her at any moment that does not warrant escalation. Every design decision in this agent has an ethical dimension that cannot be separated from the technical implementation.
The governing design principle is the dignity constraint. Every interaction must pass the dignity test: would a competent, caring human companion handle this the same way? This is not a soft principle; it is a hard constraint that drives technical decisions throughout the agent. Latency is a dignity metric: the person who is disoriented and asks “What day is it?” needs an answer in under one second. Five seconds of silence is five seconds of visible confusion that undermines her sense of competence. The latency requirement drives the edge-only deployment of the memory care infrastructure agents; network round-trips are too slow. Language is a dignity metric: the system speaks to the person Margaret is, not the person her diagnosis might suggest she is becoming. Memory is a dignity metric: the system remembers what Margaret has told it and does not require her to repeat herself. The constraint also shapes what the system refuses to do: it does not score Margaret’s cognitive state and report the score to her, even when she asks. The score is an internal signal that drives behavioral adaptation. Reporting it would communicate something the architecture does not believe is true, that her worth as a person is a function of a number a model produces.
The cognitive concierge composes from six infrastructure agents, more than any other concierge agent. Five of the six deploy edge-only because the latency requirements are non-negotiable. The Orientation Assistant answers time, place, and person questions in under 800 milliseconds. The Reminiscence Facilitator surfaces memory prompts that connect the present to her past in ways that comfort and orient her. The Routine Anchor maintains the structure of her day, one of the most powerful cognitive scaffolds available. The Wandering Prevention agent recognizes when Margaret moves toward exits at unusual times and offers gentle redirection (“Margaret, it’s three in the morning, would you like me to make you some tea?”) rather than alarm or restraint. The Sundowning Support agent addresses late-afternoon anxiety by adjusting environment and surfacing calming activities. The Communication Adapter runs underneath the others, continuously adjusting language complexity. All six on the edge. Five dedicated SLMs total 700M parameters: Orientation Assessor at 150M, Cognitive State Estimator at 200M shared with the health concierge, Confusion Detector at 100M, Reminiscence Prompter at 150M, Simplification Engine at 100M. Specialization enables sub-50ms latency on time-critical functions; confusion detection running inside a 700M general model would be slower than a dedicated 100M specialist.
The architecture’s most important behavioral property is adaptation without acknowledgment. Language gets simpler. Sentences shorten. Vocabulary narrows toward the words she still uses with confidence. Reminders increase in frequency, decrease in complexity. Choice presentations narrow from four options to three to two. The transition is continuous rather than stepped. The naïve approach would be to define modes (normal, moderate impairment, advanced impairment) and switch between them at thresholds, but the thresholds are visible to the person. The day the system switches modes is the day she notices something has changed about how it speaks to her. The notice is itself the harm. The architecture refuses the mode switch; adaptation is parameterized by the Cognitive State Estimator’s continuous output rather than discrete thresholds. The honest limitation: the technique works smoothly in the middle of the cognitive trajectory and works less smoothly at the extremes, where the escalation framework (Series 04) governs transitions.
When the Cognitive State Estimator detects a low-capacity day, every other concierge agent adjusts. The earning concierge reschedules the cooking class with the student in Brisbane, framing the rescheduling as a routine timezone reshuffle. The buying agent defers the substitution review. The financial concierge raises approval thresholds: today is not a day to authorize a Roth conversion. The home environment adjusts lighting and temperature to settings that have historically supported Margaret on her better days. One assessment, thirteen adjustments. This is the integration argument made concrete. The propagation is also one of the architecture’s most sensitive properties: every concierge agent sees the cognitive state estimate, no external party sees it, the family does not see the state, the clinician does not see the state without explicit consent and a specific clinical purpose. The state is internal.
Family notification follows clear boundaries. The daughter receives a weekly summary, not real-time cognitive scores. Daily fluctuations are normal; sharing them creates anxiety without actionable insight. The agent notifies family when a trend persists for three or more weeks, when a safety event occurs, or when the escalation hierarchy requires family involvement. The person’s privacy is preserved even within the family. Margaret has not told her daughter about the moments she has noticed. The system does not tell the daughter either, until Margaret tells the system to or the threshold for clinical concern is crossed and the family is part of the consented escalation. A different design would surface every signal to the family on the assumption that more information is better. The architecture rejects this design because the assumption is wrong. More information about Margaret’s interior life, distributed to family who cannot act on most of it, produces anxiety rather than help.
For the full treatment of the dignity constraint, the SLM stack for memory care, the cross-concierge effect, and the family notification boundaries, read the complete article on BlueMirror.tech.