The app does one thing: big cards on the home screen, three fields and a photo. Sign-in by employee number, no email needed. Turkish, English and Slovenian ready.
On-prem EHS and continuous improvement platform
Every worker on the floor, part of safety and improvement.
Occupational health and safety (EHS) reports plus 5S, Kaizen, Ethics Line and Eagle Eye records open from a phone in 30 seconds; the workflow starts on its own by report type and location, ties to an action, and turns into points and a reward. On your own server.
For factories, plants and sites; Turkish, English and Slovenian ready, more languages can be added.
Fieldloop in numbers
6
Modules: near miss, hazard, 5S, Kaizen, Ethics Line, Eagle Eye
205
REST API endpoints, documented with Swagger
3
Languages ready: Turkish, English, Slovenian
5
Scheduled jobs: reminders and reward reports
4
Management levels: specialist, manager, director, general manager
100
Points per reward; no threshold skipped, none produced twice
Sound familiar?
Five sentences, five mechanisms.
Each one is a sentence we heard from a production plant; opposite it, the product's answer, as a mechanism rather than a promise.
Hazards are seen on the floor but never reported; paper forms get lost.
Four big cards, three fields and a photo on the phone. Sign-in by employee number, no email needed; a report opens in 30 seconds and lands with the right department at once.
Reports get opened, nobody closes them.
Every report gets a due date, an owner and a risk level; owners of actions that are due or overdue get a reminder every Monday at 09:00.
Kaizen evaluation depends on the person and takes forever.
A suggestion decides for itself how far up it goes by its score: specialist, manager, director, general manager. The evaluator scores from the link in the email, without signing in to the panel.
Participation drops after the first month.
Every completed report writes points to the person who opened it; a reward record is created at 100 points, the reward report goes to the director on the 20th, and on approval it lands with HR as an Excel file.
In an audit, 'how was this decision made' goes unanswered.
Approved criteria are locked as a snapshot, every evaluation step is its own record, and the director's approval stores the IP address and browser.
Modules
Six report types from EHS to Kaizen, one data model, one permission model.
Six flows produced by the same people on the same floor, on one platform: the same lifecycle, the same action-plan discipline, the same statistics screen. Each module is switched on per organisation.
- ISG-2026000001
Near miss and hazard
Hazard zone, description, proposed measure, date found and a photo. The action plan sets risk severity and likelihood (0 to 12), the calculated risk level (1 to 4), a due date and more than one owner.
- Photo report, searchable hazard-zone list
- Severity × likelihood risk matrix
- One-page PDF report
- 5S-2026000001
5S
Order and cleanliness reports share the EHS lifecycle and action-plan discipline, with their own numbering, PDF template, statistics tab and due-date reminder.
- Zone-based report with photo
- Due date and owner
- Separate statistics and Excel
- KZ-2026000001
Kaizen
A suggestion is scored on configured criteria sets: type (EHS, quality, production, cost, environment, personal development), cost saving, man-hour gain, environmental waste, energy. The total score decides the next level.
- Automatic level by score band
- EHS and quality approval queue
- Reward per suggestion
- ET-2026000001
Ethics Line
Harassment, discrimination, conflict of interest, bribery or misconduct reports open confidentially or anonymously; they land in a separate flow only the ethics committee can see, and two-way communication continues through a case number.
- Anonymous or named report
- Queue visible to the ethics committee only
- Records fit for retaliation protection
- KG-2026000001
Eagle Eye
Behaviour-based safety observation: on planned walks, safe and unsafe behaviours, unsafe conditions and good examples are recorded; positive feedback earns points too, and the observation count is reported as a leading indicator.
- Observation checklists
- Safe / unsafe behaviour and condition
- Leading-indicator report
- Every report carries the organisation it was created in as a snapshot; if the employee moves, the old report stays with the right organisation.
Workflow engine
The report's type and location build the flow, not code.
When a report opens, the system looks at its type, the plant and department it came from, and its risk or score band; who sees it, who approves it and which channel carries the news are decided from that. Levels, score bands, approval thresholds and roles live in the database: when the process changes, the software does not.
Type and location pick the flow
A near miss lands with the maintenance manager, a Kaizen with the first evaluator, an ethics report with the ethics committee only; plant and department set the owner.
Score sets the level
A Kaizen score passes through a score band to role mapping: a low-impact suggestion closes at the specialist, a high-impact one climbs to the general manager. The mapping is edited from the panel.
Approval queue and lock
Above the band, the suggestion goes to EHS and/or quality approval; approved criteria are locked, the next level is not asked again, the score cannot change afterwards.
Decisions without signing in
Evaluators and directors land on the decision screen from a signed, time-limited link in the email. Email, SMS and push deliveries are logged: 'did the notification go out' is not answered by guessing.
Points and rewards
From report to reward, nobody has to remember anything.
What keeps participation alive is not the reward but the reward arriving on its own. The chain sits inside the system end to end, and every link leaves a trace.
Points are written
Every completed EHS, 5S, Ethics Line and Eagle Eye report writes a point record to the person who opened it; in Kaizen the points come out of the evaluation.
The threshold fills
As the total in a source type crosses each 100-point threshold, a reward record is created: 100, 200, 300. No threshold is skipped, none is produced twice.
The report goes to the director
On the 20th of each month a reward report per organisation is generated and emailed to the director; approval is given from a signed link, without signing in.
It lands with HR
The approved report goes to HR with an Excel attachment; on the 24th a reminder goes out for unapproved reports. Who marked the payout as given, and when, is on record.
A report's journey
From the floor to the general manager, from 30 seconds to a month.
Every step is real product behaviour; the times are an example, the rules are the system's defaults.
- 07:4201
The floor reports
At the start of the shift an operator photographs the missing pin on the die rack, picks the hazard zone from the list, sends. The report gets an ISG number.
- 08:0502
A counter in the panel
The EHS specialist opens the panel in the morning; the counter in the menu shows the new report. The app sends the operator an 'in review' push.
- 08:2003
Action plan
Severity and likelihood are entered, the risk level is calculated, the due date comes in as approval + 3 days, a maintenance owner is assigned.
- Mon 09:0004
Reminder
A weekly email to the owners of actions coming due: the week starts with the list of open work.
- 14:3005
Completed
The owner closes the action; who and when are on record. The report can be turned into a one-page PDF.
- 14:3106
Points
Points are written to the operator who opened the report; crossing 100, a reward record is created and shows under 'My points'.
- The 20th07
Reward report
A reward report per organisation is generated and emailed to the director; the director approves from the link, HR gets the Excel.
- Any time08
Statistics
The 30-day trend, department breakdown, critical open reports and best reports in the panel; Excel in one click.
Mobile app
With gloves on, under factory light, on a weak network.
The iOS and Android app, built in Flutter, does one thing: open a report in 30 seconds. It ships under your brand name through enterprise distribution.
Sign in by employee number
No email needed; an employee number or phone is enough. Employees migrated from the old system sign in without resetting their password.
Photos prepared on the device
In-app cropping and automatic size optimisation; uploads stay short on a weak network.
Searchable hazard zones
Long zone lists in a searchable dropdown: the floor's own vocabulary, 'Tandem press line', 'Robot spot welding', 'Die racks'.
My points and My reports
Points earned and their history, the status of every report opened; a push when one goes into review. Turkish, English and Slovenian.
Enterprise distribution on the App Store and Google Play, on your developer accounts.
Admin panel
In the panel everything is counted, scheduled and logged.
The PWA panel, built in Next.js, installs on phones and desktops; when the network drops, the last data comes from the cache. Definitions, flows, statistics and the audit trail in one place.
Scheduled jobs
- Monday 09:00EHS due-date reminder
- Monday 09:005S due-date reminder
- Monday 09:00Pending Kaizen evaluations
- 20th, 00:00Generate the reward report, send to the director
- 24th, 09:00Reminder for unapproved reports
All pinned to the Europe/Istanbul time zone; independent of the server clock.
Statistics and reporting
- 30-day trend and monthly department breakdown
- Cross-breakdown by plant, status and type
- Distribution by hazard zone and risk level
- Critical open reports: high risk and not yet closed
- Best reports, users who report the most
- Detail and summary Excel export
Definitions
- Organisations and logos
- Departments and job titles
- Hazard zones
- Processes and equipment
- Kaizen criteria and levels
- Languages and system settings
Audit trail
Who, when, on what grounds
The approving, completing and rejecting user; a rejection reason is mandatory.
Locked snapshot
Approved Kaizen criteria are locked as JSON; the score cannot change afterwards.
IP on approval
The director's approval stores the IP address, email and browser.
Delivery logs
Every email and SMS in its own collection; searchable from the panel.
AI agents
You connect your own assistant; there is no chat box inside the product.
A built-in MCP server, a Claude.ai connector and semantic search: Claude, Cursor and Codex list reports, ask about open actions and find similar records. The product generates nothing on its own; if your agent connects, it uses it.
Add the connector
In Claude.ai choose 'Add custom connector' and paste your organisation's API address; the browser goes to the panel, and the consent screen shows the client name and scopes.
Scopes
Reading is the default; writing is a separate scope and ships off. An agent sees only the organisation and modules its role sees.
Every call logged
Which agent, which key, which tool, when: it reaches the audit log. A persistent API key for Cursor and Codex.
Things you can ask
- ›Which EHS actions in Gebze are overdue this month?
- ›Has a similar report about the die racks been opened before?
- ›Are there reward reports waiting for the director's approval?
Semantic search
Reports and actions are indexed with meaning vectors (embeddings); type 'the rack wobbles' and the locking-pin reports come up too, no keyword to hit.
There is no AI component inside the product; risk level, evaluation level and rewards are calculated by rules. AI comes in when you connect your own agent, with your permission.
Your product
Adapted to your organisation, connected to your systems.
The source code is at Internative; flows, criteria, roles and screens are shaped at code level around your organisation's vocabulary and processes. You are not bound by the limits of a packaged SaaS.
Code-level customisation and SSO
A new module, a new criteria set, an approval chain or a report of your own: the team that builds the product writes it. For corporate sign-in, your existing SSO (Azure AD, Keycloak, SAML, OIDC) is integrated.
- Modules, criteria and screens of your own
- Corporate sign-in with your existing SSO
- Brand and terminology adaptation
Open API and enterprise integration
The 205-endpoint REST API is documented with Swagger; two-way data flows are built with HR, ERP and maintenance systems: SAP, Logo, Netsis, Dynamics and in-house applications. Push, SMS and email on your own providers.
- Employee and department data from your HR system
- Actions into maintenance and ERP systems
- Single sign-on with your in-house applications
Internative: since 2017, 100+ enterprise clients, 400+ projects.
Who it is for
Any organisation with a shop floor.
The common thread is the same: people working on the floor, EHS forms that get lost and actions nobody can follow up. From a single plant to multi-site groups.
Automotive suppliers
Press, welding and die lines; hazard zones defined in the floor's own words, a Kaizen culture kept alive with points.
Metal and machinery manufacturing
Machining, assembly and welding; near-miss reports tied to the risk matrix, 5S tracked by zone.
Chemicals, pharma and coatings
Eagle Eye observation walks and hazard reports; process and equipment definitions in several languages.
Food and beverage
Shift-based 5S and hygiene reports; due date and owner, weekly reminders.
Energy and facility operations
Mobile reports across scattered sites, sign-in by employee number, an app that works on a weak network.
Logistics and warehousing
Forklift, racking and loading-bay hazards; photo reports, a critical open reports screen.
Construction and site work
Reports by subcontractor and crew; daily observations, near misses and the Ethics Line in one app.
Holdings and multi-site groups
Organisation isolation, statistics and permissions per plant; one platform and one points system across the group.
On-prem and security
API, database and job runner on your server.
The .NET 8 API is packaged self-contained; no runtime install on the target server. MongoDB runs on the same infrastructure. The only mandatory external dependencies are push, SMS and email delivery, and those run on your accounts.
What gets installed
- .NET 8 REST API, self-contained publish, under PM2
- MongoDB; Hangfire scheduled jobs in the same database, no replica set needed
- Next.js admin panel (PWA), behind Nginx
- Flutter iOS and Android app, enterprise distribution
- PDF (from HTML templates) and Excel generation on the server
- MCP server and semantic search index
External services (delivery only)
Opened on your accounts or connected to what you already run; report data and photos never leave your server.
- Email: your existing SMTP server (Office 365 ready)
- SMS provider, with a delivery log
- Push: OneSignal, device matched per user
- Bot gate: Cloudflare Turnstile
- Corporate sign-in: your existing SSO
Identity and passwords
JWT sessions, role-based permissions; passwords stored with BCrypt. Migrated employees sign in without resetting their password.
Organisation isolation
An organisation admin sees only their own organisation's data; the constraint is enforced at every layer down to the statistics queries.
Immutable audit trail
Approved criteria as a locked snapshot, every evaluation its own record, the IP on the director's approval; a report carries the organisation it was created in.
Zero-loss data migration
Users, reports, images, points and reward history are moved from your old system; 13 migration runners and a layer compatible with the old password encryption are ready.
How you get it
Six steps from discovery to operations support.
The on-prem model: adapted to your organisation, installed on your server, your team trained, operated by Internative.
Discovery
Plants, departments, modules and the approval chain are mapped together.
Definitions
Hazard zones, processes, Kaizen criteria and levels are set up in your organisation's vocabulary.
Installation
The stack is installed on your server; SMTP, SMS, push and SSO are connected.
Data migration
Users, reports, points and reward history are moved from your current system.
Mobile rollout and training
The app goes out through enterprise distribution on the stores; EHS, HR and managers are trained on the panel.
Operations support
Updates, backups and support are operated by Internative.
Price and timeline by proposal, based on plant and user count; the discovery call is free.
FAQ
Frequently asked questions
No. The API and MongoDB run on your server. The only external services are push, SMS and email delivery, and those run on your accounts.
Levels, score bands, approval thresholds and roles are managed in the database. A different process is a different configuration; anything beyond that is adapted at code level.
No. Users, reports, images, points and reward history are moved from the old system; employees can sign in with their old passwords.
No. Near miss and hazard, 5S, Kaizen, Ethics Line and Eagle Eye are on the same platform; modules are switched on per organisation.
Only the users defined as the ethics committee. A report can be opened anonymously and two-way communication continues through a case number; managers of the other modules cannot see this queue.
Behaviour-based safety observation: on planned walks, safe and unsafe behaviours and conditions are recorded, and good examples earn points too. The observation count is the leading indicator that comes before an accident.
They do not. Evaluations and reward approvals are done in one click from the signed link in the email; the approval stores the IP address and browser.
The 205-endpoint REST API connects to HR, ERP and maintenance systems: SAP, Logo, Netsis, Dynamics and in-house applications. Your existing SSO is integrated for corporate sign-in.
Not inside the product; risk and points are calculated by rules. With the MCP server and the Claude.ai connector you connect your own agent, and semantic search finds similar reports.
The discovery call is free.
We map your plants, modules and approval chain together; the demo runs on an example close to your floor.
Internative: since 2017, 100+ enterprise clients, 400+ projects.
Request a demo
Tell us your plant and headcount and how you track today; we will come back with the install scope.