01 · Purpose-led
Specific business question
Teho is used to answer defined business questions, not to conduct general monitoring.
HOW A STUDY WORKS
A bounded work-pattern study, not a monitoring layer.
A Teho Study helps organisations see where time and cost concentrate in the work Teho can safely evidence, and what should change first. This explainer covers the study boundary, approved signals, privacy thresholds, coverage limits, and control model used before any client rollout.
Five commitments shape each Study.
These are the public principles behind the Study privacy architecture.
02 · Minimum capture
Bounded evidence base
Approved work signals are governed by the agreed Study boundary. Excluded activity is kept out of the evidence base, and deployment-specific controls define what is technically blocked before capture.
03 · Optional evidence
Only with explicit agreement
Richer evidence, including screenshot, semantic, or LLM-enabled paths, is optional and used only where needed to answer the question.
04 · Team-level
Patterns, not people
Only repeated, privacy-safe work patterns become evidence. Reporting is team or workflow-level and subject to minimum cohort thresholds.
05 · Client-controlled
Settings agreed before rollout
Access, capture settings, retention, deletion, consent posture, and reporting audience are approved upfront.
What it is for.
A Teho Study surfaces repeated patterns in systems of work and supports better operational decisions. It is scoped to one defined team or workflow over an agreed period, rather than used as an open-ended monitoring layer.
The output helps leadership prioritise what to stop, redesign, automate, AI-enable, or keep human-led. It is not line-by-line oversight of individuals.
How patterns graduate into evidence.
A lightweight device agent is deployed to agreed participant devices. It captures approved work-pattern signals, sends them to the agreed processing path, and produces leadership reporting only from patterns that pass privacy, coverage, and confidence checks.
Stage 01
Approved work signals
The agent captures only the signal categories enabled for the agreed deployment.
Stage 02
Privacy and confidence checks
Teho classifies work-pattern signals, applies privacy controls, and suppresses low-confidence or too-small slices.
Stage 03
Reportable findings
Graduated patterns become a team-level Work Profile Map, opportunity portfolio, and role-redesign recommendation.
Coverage limits are part of the privacy design. The readout shows time and cost for activity that clears the reporting bar. Some real activity may be excluded, suppressed, emerging, low-confidence, or outside the observation window, so the picture is useful without pretending to be complete.
What the system can and cannot access.
Can be in scope
- Application context and active-window title metadata.
- Timestamps, duration, and interaction-pattern labels.
- Optional screenshot, semantic, or LLM-enabled evidence where richer signal is explicitly approved.
- Only the team, workflow, and Study period agreed in advance.
Can be excluded
- Specific applications, keywords, and time periods.
- Sensitive windows through privacy-guard rules.
- Screenshot capture entirely, if the Study does not require it.
- Idle periods and non-work activity outside the agreed evidence boundary.
Out of reach
- Open-ended device history or personal memory layer.
- Keystroke transcripts, credentials, audio, or video.
- Images that do not pass configured scope and privacy controls.
- Screenshot images when screenshots are disabled.
What it is not designed to do.
The product is not intended to operate as a covert or general-purpose monitoring tool. Its design aim is bounded workflow insight.
- No covert monitoring.
- No individual ranking or performance scoring outputs.
- No stored keystroke transcripts or credential capture.
- No audio recording, video recording, or open-ended screen memory.
- No unbounded reporting outside the approved team or workflow scope.
Capture and privacy controls.
Each deployment has a control model covering scope, capture settings, reporting posture, consent/indicator settings, Privacy Guard exclusions, optional evidence paths, and retention. The client can inspect these before device deployment begins.
The important distinction: Teho does not promise "nothing is captured". It promises controlled, purpose-led capture with client-approved boundaries and team-level reporting.
Reporting and access model.
Reporting is designed to remain aggregated at team or workflow level. Minimum cohort thresholds suppress slices that are too small, and low-confidence or unclassified activity is not forced into findings.
No portal surface is designed to produce a per-person score, leaderboard, or supervisor-facing activity feed.
Security, retention, and governance.
The Activity Analysis baseline runs on UK-hosted managed cloud infrastructure. Optional evidence or model-processing paths are named per deployment before contract signature.
Governance decisions are made upfront: in-scope team or workflow, participant group size, OS scope, screenshot mode, semantic policy, consent posture, blocked applications, blocked keywords, reporting audience, and retention settings.