Sospita Software develops digital products in which artificial intelligence may assist users with analysis, recommendations, summaries, coaching and other decision-support activities.

Our approach is based on a simple principle:

AI assists; people decide.

Artificial intelligence is not used by Sospita as a substitute for professional judgement, organisational responsibility or human approval where decisions may affect health, safety, operations or individuals.

1. Human Oversight

AI-generated information is advisory unless explicitly stated otherwise.

Where AI is used in safety, operational or professional workflows, authorised users remain responsible for reviewing the information and making the final decision.

AI does not independently:

  • approve a Permit to Work;

  • determine that work is safe to proceed;

  • make a final risk-acceptance decision;

  • establish an official root cause of an incident;

  • assign responsibility for an incident;

  • close or verify a corrective action;

  • provide medical diagnosis or medical clearance;

  • make autonomous employment, access or disciplinary decisions.

2. AI in Sospita Products

AI functionality is available only in selected Sospita products and features.

Sospita SafeGuard

Sospita SafeGuard includes AI-assisted permit review and recommendations.

The AI provides advisory support only.

It does not independently approve a permit, replace authorised workflow participants, or determine whether work is safe to proceed.

Final decisions remain with authorised people.

Sospita Athletic Performance

Sospita Athletic Performance includes AI-assisted features such as:

  • AI Daily Coach;

  • AI Monthly Athlete Evaluation.

These features may provide coaching guidance, summaries or athlete-evaluation support.

AI output is informational and advisory.

It does not provide medical diagnosis or medical clearance and does not replace a qualified coach, healthcare professional or trainer.

Incident Investigation & Root Cause Analysis

AI-assisted investigation capabilities have been developed for the Sospita Incident Investigation & Root Cause Analysis module but are not currently live for production use.

Planned AI capabilities may assist with tasks such as:

  • identifying information gaps;

  • suggesting investigation questions;

  • analysing barriers;

  • proposing possible causes;

  • suggesting actions;

  • assisting with draft report content.

AI suggestions require human review.

AI does not independently confirm official root causes, assign action owners, finalise incident reports or make formal investigation decisions.

Products Without Current AI

Sospita Risk Assessment and My Checklists do not currently use AI.

The presence of AI in other Sospita products should not be interpreted as meaning that every Sospita product contains AI functionality.

3. Professional Judgement Remains Essential

AI systems can produce inaccurate, incomplete, outdated or inappropriate output.

Users should evaluate AI-generated information against:

  • actual operational or site conditions;

  • available evidence;

  • professional knowledge;

  • organisational procedures;

  • applicable laws and standards;

  • information that may not have been available to the AI system.

AI recommendations should be treated as additional decision-support information rather than authoritative conclusions.

4. Data Minimisation

Sospita seeks to provide AI functionality using only information reasonably necessary for the relevant feature.

Where practical, unnecessary identifying or sensitive information is removed or reduced before information is sent for AI processing.

The exact information used depends on the relevant product and feature.

5. Health and Sensitive Information

Sospita does not design its AI features to make medical diagnoses.

In Sospita Athletic Performance, supported Apple HealthKit and Android Health Connect integrations may be used to derive limited coaching or recovery indicators.

Raw numerical health values, such as raw sleep or calorie data obtained from HealthKit or Health Connect, remain on the user's device.

Limited non-numeric derived indicators may be transmitted to and stored by Sospita for coaching and evaluation functionality.

Raw HealthKit or Health Connect numerical values are not sent to OpenAI as Athletic Performance AI prompt inputs.

6. Third-Party AI Providers

Some Sospita AI functionality may use third-party artificial intelligence service providers.

Where third-party providers are used:

  • Sospita seeks to limit information sent to what is necessary for the feature;

  • relevant providers are reviewed as part of Sospita's privacy and service-provider governance;

  • applicable providers may be disclosed in Sospita's Subprocessors and Service Providers information where appropriate;

  • processing depends on the relevant technical configuration and contractual arrangements.

Sospita does not make blanket claims that third-party AI providers never retain information unless that is supported by the applicable service configuration and contractual terms.

7. No Autonomous Safety-Critical Decisions

Sospita does not design its AI features to autonomously make final safety-critical decisions.

In HSE-related products, AI may assist with review, analysis or suggestions, but responsibility remains with appropriately authorised people.

The use of AI does not replace:

  • competent personnel;

  • supervision;

  • site procedures;

  • engineering controls;

  • legal requirements;

  • organisational approval processes.

8. AI-Generated Content

AI-generated content may contain errors, omissions or assumptions.

Users should review AI-generated:

  • recommendations;

  • explanations;

  • summaries;

  • proposed actions;

  • investigation suggestions;

  • coaching guidance;

  • report drafts;

before relying on them or incorporating them into formal records.

Where a product provides an acceptance or editing workflow, users should make any changes necessary to ensure that the resulting record accurately reflects the situation.

9. Security and Confidentiality

Sospita applies its normal security and privacy principles to AI-enabled functionality.

We seek to avoid unnecessary disclosure of confidential or personal information to AI providers and to apply appropriate technical and organisational safeguards.

Users should avoid entering information into AI-assisted features that is unnecessary for the intended task.

10. Transparency

Where AI materially contributes to a user-facing feature, Sospita aims to identify that functionality as AI-assisted or AI-enabled.

We avoid presenting AI-generated recommendations as human decisions.

Preferred public terminology includes:

  • AI-assisted;

  • AI-enabled.

We avoid language that suggests AI independently controls professional or operational decision-making.

11. Responsible Development

When designing AI functionality, Sospita considers:

  • whether AI is appropriate for the intended task;

  • the potential consequences of incorrect output;

  • whether human review is available;

  • what information should be provided to the model;

  • whether personal or confidential information can be reduced;

  • how recommendations are distinguished from formal records or decisions;

  • whether users can reject, edit or disregard AI suggestions.

Higher-consequence workflows require stronger human oversight.

12. Product Changes

AI technologies and product functionality evolve.

Sospita may add, modify or remove AI features as products develop.

Material changes affecting how personal information is processed will be reflected in the relevant privacy information where required.

13. Contact

SOSPITA YAZILIM LIMITED ŞİRKETİ
Talaytepe Mah. 4035. Sokak
Güneşyapı Sunrise Garden Sitesi H Blok No: 27
21070 Kayapınar / Diyarbakır
Türkiye

Email: info@sospita.io