User Guide

Transformative —
Intelligence from Your Data

Your operational data already holds the answers. Transformative connects to the systems you run every day and turns the raw record into something you can question, track, forecast, and act on.

Fleet Operations Compliance Tracking AI-Assisted Analysis Predictive Forecasting Geospatial Intelligence
01 — Overview

What is Transformative?

Transformative is a compliance intelligence platform. It connects to the databases, form systems, and telematics platforms your organisation already uses — and gives you a single place to see, question, and act on all of that data together.

Think of it as the layer that sits above your existing systems. Your route data lives in SQL Server. Your field forms come in via DoForms. Your vehicle GPS streams through Geotab. Transformative ingests all of it, stores a complete bi-temporal history, and puts a powerful analytical interface on top.

You don't need to change anything about how you operate. The platform reads your data as it already exists, tracks every change over time, and lets you interrogate what happened, when it happened, and what's likely to happen next.

📊
Live Data Dashboard
See your operational data as it stands right now, updated on every poll cycle from your source systems.
🤖
Plain-English Queries
Ask questions in natural language. The AI translates your intent into precise data queries instantly.
🗺
Geospatial Intelligence
GPS tracks, stop sequences, and route maps — all drawn from your actual operational data.
🔮
Predictive Forecasting
Trend-based AI forecasting with divergence alerts when actuals deviate from projections.
🕰
Full History
Every change is preserved. View your data as it was on any date, not just as it is today.
What-If Scenarios
Model changes to your operations without touching live data. Explore outcomes before you commit.
02 — Navigation

Your Dashboard

When you log in, you land on your data source dashboard. Each connected data source appears as a named panel — your fleet management system, your forms data, your telematics feed. Select one to open it.

Dashboard home

Once inside a data source, the interface divides into three areas:

A
Tab Strip — Table Selector
Each provisioned table appears as a tab across the top: Routes, Vehicles, Employees, Assignments, Trip Log, and so on. Click a tab to load that table's data into the main grid. A pin icon (📌) appears on the active tab when you've explicitly selected it; a dashed tab means the AI inferred it from your last query.
B
Main Grid — Your Records
The central list shows the current records for the selected table. Each row is a single record — a route, a vehicle, a trip, an employee. Click any row to select it and see its full detail in the right panel.
C
Right Detail Panel
Selecting a record opens its full field set in the right column, along with related data — linked GPS tracks, child records, version history, and any connected data from other tables.
Data source view
💡
Tip
You can mark any tab as a favourite using the star icon that appears on hover. Your home tab — the one that loads automatically when you open a data source — can be set per data source from the tab menu.
03 — Data Navigation

Exploring Your Data

Records in the main grid represent the current state of each entity. Clicking a record reveals everything attached to it — not just the fields, but the full connected graph.

Drill-Down Navigation

For data with natural parent-child relationships — routes containing stops, stops with students assigned — the right panel shows a linked hierarchy. Click through from a route to see its stops. From a stop, see which students board there. From a student, see their care needs and medical requirements. Each level loads the next without leaving the page.

Drill-down navigation

The Toolbar

Above the main grid, a compact toolbar gives you three controls that work together:

ControlWhat it does
Alert RulesDefine threshold rules that fire when a field value crosses a limit — average speed below threshold, record count above expected. Active alerts show a count badge.
Data QualityRuns four automatic checks: field coverage, duplicate detection, timeliness, and data staleness. Flags issues without you having to look for them.
FilterApply one or more field filters to narrow the grid. Works alongside AI queries — you can combine both.
Toolbar controls
04 — AI Intelligence

Asking Questions in Plain English

The AI query bar sits between the toolbar and the data grid. It's the most powerful way to work with your data — type a question or instruction in plain English and the platform figures out exactly what to query, which table to look in, and how to surface the result.

There are three distinct modes. You switch between them using the mode selector on the left of the bar.

AI query bar
Filter
Filtering — Find the Records You Need
Describe what you're looking for and the grid updates to show only matching records. The AI understands your data's structure — it knows which tables are related and can cross-reference them in a single query.
"show routes with stops in zone DOH"
"show active vehicles assigned to the Yonkers district"
"show students with an EpiPen requirement"
— The grid narrows to exactly those records. No SQL required.
Insight
Insight — Summarise and Analyse
Ask a broader analytical question and get a written summary backed by aggregated data. The AI runs a two-pass analysis: first it calculates the numbers, then it interprets what they mean in plain language.
"what is the average trip speed by zone over the last 4 weeks?"
"which routes have the highest timesheet variance?"
"summarise vehicle utilisation across the fleet"
— Returns a narrative summary with the underlying figures.
Predict
Predict — Forecast a Trend
Ask the platform to project a metric forward based on historical patterns. The AI identifies the trend, generates a weekly forecast, and stores it so you can track whether actuals follow the projection. See the Forecasting section for more detail.
"forecast average trip speed for the next 8 weeks"
"predict timesheet hours over the next month"
— Generates and stores a forecast run you can monitor over time.
How table inference works
When you type a query, the AI reads your available tables and their relationships, then infers which table you're most likely asking about. A small chip appears in the input row showing the inferred table name. If it picks the wrong one, click a tab to pin your choice and re-run the query.
Insight mode result
05 — Geospatial

Maps and GPS Intelligence

For data sources with geospatial content — vehicle GPS, route stop coordinates, trip logs — Transformative renders an interactive map directly in the right panel when you select a record. No configuration needed; the platform reads latitude, longitude, and sequence information from the registry and draws the map automatically.

Route Maps

Select a route record and the map draws the actual GPS breadcrumb trail for that route — the path vehicles travelled, timestamped at the intervals your telematics system captured. The track is drawn in time order so you can see exactly how the journey unfolded.

Route GPS map

Stop Markers

Where a route has planned stop data registered, numbered amber markers are overlaid on the GPS track — each marker showing the stop sequence number and stop name on hover. This lets you compare the planned stop sequence against the actual path travelled at a glance.

Geotab Trip View

For Geotab-connected data sources, selecting a vehicle record shows a trip strip — each trip card displaying start time, end time, distance, and maximum speed. Click any trip card to isolate its GPS track on the map. This lets you drill into a specific journey without losing the context of the full day's activity.

Geotab trip view
📡
All positions data
The all-records scatter map — visible when no specific record is selected — plots every position point in the current table as a dot on the map. This gives an immediate overview of where activity is concentrated across your operation before you drill into any individual record.
06 — Temporal Intelligence

Viewing Your Data at Any Point in Time

One of the most distinctive things about Transformative is that it never overwrites historical data. Every time a record changes in your source system — a route is reassigned, a vehicle changes status, a timesheet is amended — the platform stores both the new version and the old one, with a precise timestamp for each.

This means you can see your operation not just as it is today, but exactly as it was on any date in the past.

The Time Scrubber

At the top of each data source view, a date/time control lets you set the "as at" date. Move it to last Tuesday and every table, every grid, every record refreshes to show the state of your data as it existed on that day. The header shows a persistent amber banner reminding you you're viewing historical state, with a "Return to Present" button.

Historical state view

Version History

Selecting a record and opening the Version History panel shows every version of that record across time — each change listed with its timestamp and the values that changed. This is your complete audit trail, available for every record, with no extra configuration.

Version history panel

Bi-Temporal Gap Detection

Transformative tracks two time dimensions for every record: when it was true in the real world (application time) and when the platform recorded it (system time). When these are significantly different — meaning data arrived late — the gap detection panel flags it. This is particularly useful for compliance: if a record needed to exist on a certain date but was only entered later, the gap is visible and measurable.

⚠️
Why this matters for compliance
In regulated environments, the question isn't just "did this happen?" — it's "did we know it happened in time?" Bi-temporal gap detection answers the second question, which traditional systems can't. If a vehicle compliance record was entered three days late, the platform shows that gap explicitly rather than presenting the record as if it arrived on time.
07 — Predictive Intelligence

Forecasting and Divergence Alerts

The Predict mode in the AI bar doesn't just show you what happened — it projects what's likely to happen next, based on the historical patterns in your data. And once a forecast is running, the platform monitors incoming actuals and alerts you if reality starts to diverge from the projection.

Running a Forecast

Switch the AI bar to Predict mode and type what you want to forecast. The platform identifies the relevant metric, analyses the historical trend, and generates a weekly projection for the period you specify. The forecast is stored against your tenant so you can return to it later.

1
Type your forecast request
e.g. "forecast average trip speed for the next 8 weeks" — the AI identifies the field, the aggregation method, and the relevant table automatically.
2
Review the projection
A sparkline chart appears showing the historical baseline (teal line), the AI's forward projection (dotted purple), and a written narrative explaining the trend it identified.
3
Set a divergence alert
Click the bell icon on the forecast history row. Set a divergence threshold — e.g. alert me when actuals diverge more than 15% from the forecast for 2 or more consecutive weeks. The platform checks this on every data poll.
4
Monitor the badge
When an alert fires, a ⚠️ badge appears on the forecast row showing the count of triggered periods. Click it to see the per-period breakdown — forecasted vs actual vs the percentage divergence.
Forecast divergence chart
Divergence alert panel
🔮
What makes a good forecast target?
Forecasting works best on metrics with a consistent weekly pattern — average trip speed, timesheet hours per zone, vehicle utilisation rate. Metrics that are highly variable day-to-day but stable week-to-week are ideal. The platform needs at least 8–12 weeks of historical data to identify a reliable trend.
08 — Scenario Engine

What-If Scenarios

Scenarios let you model a proposed change to your operation and see what the data would look like if that change had happened — without touching the live record. You're creating a branch of reality that exists alongside your actual data, not instead of it.

A scenario has a branch point — the date from which the hypothetical diverges from history. Everything before the branch point is your real operational data. Everything after is the modelled version.

Scenario panel

Creating a Scenario

From the Scenario panel (below the main data grid), create a new scenario and give it a name and hypothesis. Then make the proposed changes — reassigning a route, adding a vehicle, moving a resource. The scenario tracks those changes as a separate set of records branching from your chosen date.

Private and Published

New scenarios are private by default — visible only to you. When you're ready to share the analysis with others in your organisation, you can publish it. Publishing is permanent: a published scenario becomes visible to all authenticated users on your tenant. It cannot be unpublished, which ensures the integrity of any analysis that others have reviewed or relied on.

⚠️
Publishing is permanent
Once a scenario is published, it cannot be made private again. Only publish when you are confident others should see it. For exploratory or draft analysis, keep scenarios in the private (My Scenarios) tab until they're ready.
09 — Connectivity

What Feeds Transformative

Transformative is source-agnostic. It connects to the systems your organisation already runs. The provisioning of a new data source is handled by your platform administrator or reseller — as an end user, you see the data after it's been connected and ingested. This section is a brief overview of the types of sources the platform supports.

Source TypeDescriptionTypical Data
SQL Server Direct connection to any Microsoft SQL Server database. The primary ingestion path — if your operational system runs on SQL Server, it can be connected with no changes to the source. Routes, vehicles, employees, timesheets, assignments — any operational table.
DoForms Pulls submitted form records from the DoForms REST API. Ingests form submissions as bi-temporal records with full version history. Field inspections, compliance checklists, maintenance reports, driver declarations.
Geotab Connects to the Geotab telematics API. Ingests trip summaries and GPS log records for map and speed analysis. Vehicle trips, GPS breadcrumbs, speed profiles, device exceptions.
OpenAPI / REST Any REST API that exposes an OpenAPI 3.x specification can be connected as a data source with watermark-based incremental polling. Third-party compliance systems, scheduling platforms, external registries.
🔌
Adding a new data source
Connecting a new data source is done through the Authority provisioning wizard — a step-by-step interface that walks through connection credentials, table selection, field mapping, and temporal configuration. During the field mapping step, administrators can also mark individual fields as PII and configure encryption. This is an administrator/reseller task and typically only needs to happen once per source system.
11 — Privacy & Data Rights

How Your Data Is Protected

Transformative is built for regulated industries where personal data carries legal weight. The platform supports GDPR Article 17 (Right to Erasure) and CCPA Right to Delete requirements through a cryptographic approach that is fully compatible with its immutable, bi-temporal data store.

How PII Fields Are Encrypted

When your administrator connects a data source, they can mark individual fields as containing personal information during the field mapping step. Each marked field is assigned a PII category:

CategoryTypical fields
IdentityFull name, date of birth, national ID, employee code
MedicalHealth conditions, care needs, medical notes, medications
FinancialSalary, pay rates, cost codes, billing references
ContactAddress, phone number, email, emergency contacts

Once marked, the raw value of that field is never stored in plain text. Instead, the platform encrypts it with a unique AES-256 key generated specifically for that individual before writing the record into the database. The encrypted value can only be read back by the platform while that key exists. When you view a record in the platform, the decryption is automatic and invisible to you — the field appears exactly as it would otherwise.

🔒
One key per person, not per table
Each data subject (person, participant, employee) has their own unique encryption key. If that person appears in multiple tables — a student in both a students table and a medical needs table, for example — all of their records across all tables use the same key. This means a single erasure action removes their personal information everywhere simultaneously.

The Right to Be Forgotten

When a data subject requests erasure under GDPR Article 17 or CCPA, your administrator processes the request through the Erasure Management section of the Authority. The process works in two stages to protect against accidental or mistaken erasures:

Stage 1 — Schedule (immediate)
The erasure is scheduled with a grace period (default 72 hours, configurable per request). From this moment, any further ingest of data for that individual is blocked. However, the encryption key still exists — existing data remains readable during the grace period while the request is verified.
🔥
Stage 2 — Hard Purge (automatic)
Once the grace period expires, the platform automatically destroys the encryption key. The historical records remain in the database for structural audit purposes, but all personal fields become permanently unreadable noise. This satisfies GDPR and CCPA erasure requirements without breaking the immutable audit ledger.
Recovery within the grace period
If an erasure is scheduled in error — wrong subject reference, request withdrawn, or a process mistake — a super-administrator can recover it at any point before the grace period expires. Recovery is fully audited and requires a mandatory reason. Once the hard purge has run, recovery is no longer possible.

What Happens to Historical Records

The platform's bi-temporal records are never deleted. After a hard purge, the rows still exist in the database — you can see that a student was on a route, that a care recipient had a schedule, that an employee had shifts. The count, the dates, the relationships between records are all intact. Only the personal content of encrypted fields is gone.

In the platform UI, erased fields display as [Data Erased] rather than a blank, making it clear that the value existed but has been legally removed rather than that it was simply empty.

Blocked Re-Introduction

Once an erasure is scheduled (even during the grace period), the platform immediately blocks any attempt to re-ingest data for that individual from a source system. If the source system still holds the record and a poll cycle runs, the row is silently rejected and logged. This log is available to your administrator as proof that re-introduction was actively prevented — relevant documentation for regulatory compliance.

📋
Your administrator's responsibility
The platform blocks re-ingest from its own poll cycle, but it cannot reach into your source system to delete records there. Your organisation remains responsible for removing the individual's data from any upstream system that feeds the platform. The platform's blocked-ingest log is evidence of what it did — it is not a substitute for upstream deletion.

Erasure History & Audit Trail

Every erasure request — scheduled, recovered, or completed — is recorded permanently in an immutable audit log. This log cannot be edited or deleted. It records who requested the erasure, the legal basis, any verification reference provided by the regulator, and when each stage completed. Your administrator can view the full history at any time from the Erasure Management page in the Authority.

12 — Reference

Key Concepts

A short reference for the terms you'll encounter across the platform.

TermWhat it means
Bi-temporal Two time dimensions tracked for every record: application time (when the event actually happened) and system time (when the platform recorded it). Enables both accurate history and late-arrival gap detection.
As At The date/time you want to view your data from. Set to "now" for current state, or any past date to see historical state. Controlled by the time scrubber in the header.
Data Source A connected external system — your SQL database, your DoForms account, your Geotab account. Each data source has its own set of tables, displayed as tabs.
Poll Cycle The regular interval at which the platform queries your source system for new or updated records. New data typically appears in the platform within minutes of appearing in the source.
Forecast Run A stored AI-generated projection for a specific metric. Each time you run a Predict query, a new forecast run is created and appears in the history list below the AI bar.
Divergence Alert A rule attached to a forecast run that fires when realised actuals deviate from the forecast by more than a configured threshold percentage. Monitored automatically on every poll cycle.
Scenario A hypothetical branch of your operational data, modelling a proposed change from a specific branch point date. Private until published. Publishing is permanent.
Compliance Tier The subscription tier that determines which features are available: Compliance (data access and history), Scenario (adds what-if modelling), Predictive (adds AI forecasting and alerts).
ValidFrom The application-time timestamp on every record — when the fact it describes was true in the real world. Used for all as-at queries and forecast comparisons.
UTC All data is stored internally in Coordinated Universal Time. The platform converts to your local timezone for display, based on your account's display timezone setting.
PII Personally Identifiable Information. Fields marked as PII by your administrator are encrypted at ingest with a unique per-subject key. They appear normally in the platform while the key exists, and display as [Data Erased] after a hard purge.
Crypto-Shredding The erasure method used by the platform. Rather than deleting records, the encryption key for a subject is destroyed. Without the key, the encrypted personal fields are permanently unreadable noise — even though the rows remain intact in the immutable ledger.
Grace Period The window between scheduling an erasure and the hard purge. Default 72 hours. During this period ingest is blocked but data is still readable. A super-administrator can recover the erasure if it was scheduled in error.
Hard Purge The automatic, irreversible destruction of an encryption key once the grace period expires. Performed by the platform's background purge worker. After a hard purge, recovery is not possible.
Subject Reference The internal identifier used to associate encrypted fields with their encryption key. Typically a stable ID column from the source system — such as a student ID or employee number. All tables sharing data about the same person use the same subject reference.