Poko manifesto · 12 October 2026
The Poko technology roadmap
Software first. Intelligence in context. Hardware when earned.
Current execution phase — October 2026: Poko is now working with individuals, businesses, teams, organisations, and governments. Professional and team workflows are current work and development. The detailed features below retain their own release, scope, and evidence status; physical environments and sensing remain further horizons.
Design rule
We do not believe every human problem requires a device. We believe every new capability should have a specific outcome, a plausible mechanism, a route to proof and an acceptable cognitive, financial and privacy cost.
Poko therefore starts with the cheapest credible intervention that can improve a real-world outcome, and only adds sophistication when it pays for itself.
Layer 1 — Poko personal software
Why it matters: The gap between insight and action creates demand for a consistent personal support environment.
Potential core experiences: weekly state and planning, practical decision tools, editable personal operating playbook, lessons tied to current needs, restarts after disruption, consent-based data integrations, real-world practice and progress views.
Engineering priority: reliable structured state, understandable recommendations, corrections, accessibility, ability to leave the app, protected privacy and outcome measurement. AI may help translate complex information into a small useful next step, but should not be a black box or the only interface.
Validation priority: people voluntarily return because it helps them act, and they achieve meaningful outcomes without accumulating maintenance work.
Layer 2 — Practitioner platform and intervention catalogue
Why it matters: Many real cognitive-function problems benefit from structured supported practice, compensation, environmental changes or appropriate referral, not just education.
Potential product: a library of functional intervention objects that professionals can select, adapt, dose and review; practitioner workflow and permissions; programme builder; outcome and cost measures; provenance, evidence level and claims boundaries.
Guardrail: separate coaching, education, allied health, clinical treatment and medical review by role, purpose and authorisation. The sixteen functional categories organise target problems; diagnosis is contextual information, not a universal folder.
Validation priority: practitioner acceptability, appropriate patient/client outcomes, safe delivery, workload economics and a clear regulatory determination for each product function.
Layer 3 — Team and enterprise decision architecture
Why it matters: Individual coaching cannot repair every broken handoff, pointless meeting or overloaded decision rule.
Potential product: process-level attention and decision audits, organisational operating rhythms, workload and interruption maps, decision rights and escalation tools, cognitive-cost-informed workflow redesign, optional environment integrations.
Guardrail: no hidden 'brain performance' scores or personal neuro-monitoring for managers. Separate worker consent and sensitive data from employment power dynamics. Design measures around shared processes and bounded intervention studies.
Validation priority: time saved, error reduction, greater decision reliability, fewer breakdowns or other independently specified organisational outcomes—not vanity dashboards.
Layer 4 — Smart workplaces and furniture
Why it matters: Cognitive conditions are physical as well as digital. Noise, light, unpredictable interruption, temperature, layout and furniture design can influence work experience and task performance.
Potential product families: adjustable/ergonomic workstations, adaptive acoustic and lighting components, quiet collaboration/focus zones, furniture designed for task transitions, smart meeting interfaces, environment sensing with local controls, restorative micro-spaces.
Preferred entry: partner with OEMs, furniture manufacturers, facilities providers and design firms. Prototype a narrow use case; run a controlled field pilot; compare it with simpler alternatives; then decide between licensing, co-design, assembly and own-manufacture.
Validation priority: actual benefit relative to installation/operating cost, accessible use, maintenance burden, user acceptance and long-term reliability.
Layer 5 — Wearables and neurotechnology
Why it matters: In limited use cases, physiological, behavioural or environmental signals might provide useful context that self-report or ordinary workflows cannot capture as efficiently.
Discovery order:
- No sensor: see whether better task and environment design solves the problem.
- Existing device: when appropriate, explore optional integration with established wearable platforms and carefully defined inputs.
- Research pilot: assess validity, privacy, calibration, subgroup performance and actionable value of additional signals, potentially including EEG/qEEG in governed studies.
- Product pathway: decide whether the purpose is general wellness, workplace design, practitioner assistance or a regulated medical use, and follow the corresponding legal and scientific requirements.
- Proprietary hardware option: proceed only if differentiation, evidence, user demand and economics justify manufacturing.
Explicit non-claims: Poko does not currently claim to read thoughts, identify an individual's capability from a single biomarker, diagnose ADHD from wearables, treat a disorder with an unapproved device, or have validated whole-person cognitive scores.
Shared platform services, if warranted
Across layers, some reusable capabilities may be valuable: task-state memory, participant-controlled contextual profiles, intervention provenance, consent and access controls, structured outcomes, longitudinal review, transparent recommendations and interoperability. Build these as reusable components where a genuine need exists; avoid constructing a grand technical abstraction before the first products have repeatable demand.
Gate-by-gate roadmap
| Gate | Commercial question | Evidence required | Typical risk |
|---|---|---|---|
| G1: Personal value | Does the core digital support help enough to earn repeat use and payment? | Cohort outcomes, retention, paid behaviour | Weak differentiation / maintenance friction |
| G2: Professional value | Can appropriate professionals use it safely and efficiently? | Practitioner pilots, governance, scope checks | Clinical claims / workflow integration |
| G3: Organisational value | Can changing systems produce measurable customer ROI? | Bounded pilots, process and functional metrics | Change management / procurement |
| G4: Physical value | Does design of furniture/space measurably improve chosen outcomes? | Field trials, cost and human-factors tests | Manufacturing, facilities complexity |
| G5: Sensing value | Does added measurement justify its burden? | Independent validity, benefit vs simpler alternatives, privacy review | Poor signal quality / surveillance |
| G6: Hardware business | Should Poko license, partner or manufacture? | BOM, supply chain, service model, regulatory assessment, margins | Capital intensity / inventory |
Time horizons are planning assumptions, not delivery promises. Capital should be gated separately for the regulated and hardware tracks.
Standards and boundaries
Australian TGA guidance distinguishes general-wellness coaching software from software or wearables with medical intended purposes; the intended purpose, functions and claims matter. OAIC privacy guidance requires careful treatment of personal and sensitive information, with consent and other lawful-basis rules depending on context. Professional scope, human research ethics, accessibility and workplace safety may also be relevant. See source library.
The long-term product bet: The most valuable future intelligent environment may not be the one with the most sensors or the most automation. It may be the one that quietly allows people to think, work, decide and recover with less unnecessary burden.