The product family
Six capabilities. One model of your business.
Atlas, Radar, Ask, Signal, Helm and Act are not separate tools that happen to share a logo. They read and write to the same governed model of the organisation, which is what makes the progression from monitoring to action possible at all.
Arctellis Atlas
KnowBusiness model and connected intelligence
Your business, understood.
Atlas is the institutional model of the organisation. It connects operational systems, databases, APIs, spreadsheets, documents and external data, then holds what that information means: the definitions, hierarchies, rules and terminology the business actually uses.
This is the difference between copying data and understanding it. Underlying systems can be replaced while the Arctellis model of the business stays stable.
- System, database, API and file connections
- Metric and KPI definitions
- Organisational hierarchies and dimensions
- Business rules and terminology
- Source lineage and permissions
- Strategy, objectives and reporting standards
Gross margin
(revenue − cost_of_goods) / revenue
- Owner
- Group Finance
- Sources
- ERP · Ledger · Purchasing
- Dimensions
- unit · category · period
- Excludes
- Intercompany, returns
- Used by
- RADAR · ASK · SIGNAL · REPORTS
Defined once. Every surface in the platform uses this definition, and every figure can be traced back through it.
Arctellis Radar
MonitorOperational intelligence and performance monitoring
See the business as it is.
Radar is the instrument panel of the business. It answers how are we doing, and increasingly, what needs my attention. A dashboard is one presentation mode within Radar, not the whole of it.
Because Radar reads Atlas, the figures on screen are the same figures used in conversation, in forecasts and in the monthly pack.
- Configured KPI views by role
- Trends and period comparisons
- Unit, product and customer comparison
- Anomalies, exceptions and attention areas
- AI-surfaced observations
- Drill-down to the underlying movement
| Business unit | Revenue | Margin | vs LY |
|---|---|---|---|
| North | £1.42m | 43.8% | ↑ 5.2% |
| Central | £1.18m | 41.0% | ↑ 1.4% |
| South | £1.31m | 36.4% | ↓ 4.1% |
| Online | £0.91m | 44.9% | ↑ 8.7% |
South is the only unit where margin fell while revenue grew. Radar flags it rather than waiting to be asked.
Arctellis Ask
UnderstandConversational business intelligence
Ask the business. Get answers you can trust.
Ask lets people interrogate the connected organisation in natural language. Answers are grounded in governed business definitions and approved analytical operations.
Arctellis deliberately separates the conversation from the calculation. A language model interprets the question and explains the result. It does not invent the figures, and it does not write arbitrary queries against your databases.
- “What drove the change in margin last month?”
- “Which locations are underperforming compared with last year?”
- “Where has labour cost increased disproportionately to revenue?”
- “What changed this week?”
- QUESTIONA person asks in their own words.
- INTENTThe model maps it to metrics, dimensions and periods that exist in Atlas.
- CALCULATEApproved deterministic operations produce the numbers.
- EXPLAINThe model writes the explanation around those numbers.
- EVIDENCEDefinition, sources and method are shown alongside the answer.
Consistency is the point. Two people asking the same question in different words get the same number.
Arctellis Signal
PredictPredictive intelligence and forecasting
See what is coming.
Signal moves Arctellis from understanding the past to anticipating the future. It combines historical organisational data with the variables that genuinely move the business: seasonality, events, promotions, pricing, operating conditions, customer behaviour and relevant external signals.
Which variables matter depends on the business and the problem. What stays constant is the model underneath: Signal forecasts the same metrics Radar reports, so a forecast and an actual are directly comparable.
Forecast targets
Demand forecast, next 13 weeks
80% BAND- Peak weekW+9 · 24.8k units
- ConfidenceHIGH TO W+6, MEDIUM AFTER
- Largest driverPROMOTIONAL CALENDAR
The forecast carries its own uncertainty. A number without a range is a guess with better formatting.
Arctellis Helm
DecideDecision intelligence and scenario modelling
Test the decision before testing it on the business.
Signal asks what is likely to happen. Helm asks what happens if we change something, and eventually, what we should do. It holds the constraints the business actually operates under and shows the trade-offs between the options rather than a single confident answer.
Decision categories
- Changing price
- Changing capacity
- Reallocating resources
- Workforce levels
- Product mix
- Marketing investment
- Inventory commitments
- Expansion decisions
Scenario branches from one decision point
- Constraint held
- VOLUME ≥ −2%
- Selected
- SCENARIO B
- Passed to
- ACT · AWAITING APPROVAL
Arctellis Act
ExecuteWorkflow, automation and execution
Turn intelligence into action.
Act is the execution layer. It connects analysis and decisions back into operational work: scheduled processes, automated reporting, workflow orchestration, approvals, exception queues, tasks, sign-offs and actions in downstream systems.
Signal anticipates. Helm decides. Act executes. Human approval, governance and confidence thresholds are part of how it works, not a caveat attached afterwards.
On autonomy
Arctellis does not offer uncontrolled autonomous AI. Where action is automated, it is automated inside thresholds the business sets, with an audit trail and a human accountable for the boundary.
- Scheduled processes and automated reportingRECURRING
- Workflow orchestration and approvalsHUMAN IN LOOP
- Exception queues and notificationsRULE DRIVEN
- Tasks and sign-offsASSIGNED
- Actions in downstream systemsWRITE BACK
- Agents within governed limitsTHRESHOLD BOUND
Supporting capability
Arctellis Library
Arctellis also produces persistent outputs. Library holds the automated management reports, board packs, saved analyses, forecast reports, scenario outputs and exports the business relies on, generated from the same governed model that answers questions and drives forecasts.
Automated recurring reporting is one of the platform’s most immediately useful capabilities. Not report scheduling. The same trusted intelligence, written down and circulated on time.
- Automated management reports
- Board packs
- Saved analyses
- Forecast reports
- Scenario outputs
- Presentations and exports
Start with the decision that matters most.
We identify which recurring decisions carry real value, quantify what is at stake, and start there.