From Reactive Maintenance to Decision Governance.
AutoHalo is a decision layer that converts maintenance uncertainty into bounded probabilities — before commitment, not after.
The governance deficit in maintenance.
Reactive by Default
Maintenance decisions are made after faults occur. Scope deviations during service are unplanned, undisclosed, and absorb capacity that could have been allocated in advance.
Uncertainty Without Structure
Experienced technicians hold probabilistic knowledge in their heads. None of it is structured, communicated to decision-makers, or used to pre-position resources.
Operational Cost
In defence: unplanned downtime reduces fleet readiness and forces mission reallocation. In automotive: unapproved deviations destroy customer trust and create idle bay time.
This is not a technical problem. It is a decision problem — and it costs more than most organisations measure.
A decision layer — not a prediction system.
AutoHalo converts maintenance uncertainty into bounded probability ranges. Instead of saying ‘the vehicle needs service’, the system states: ‘for a vehicle of this type and mileage, the likelihood of a scope deviation in this range is X%.’ That is a different class of information.
This enables decisions that currently cannot be made: pre-ordering spare parts, allocating resources before a vehicle enters the queue, and communicating to decision-makers on data — not intuition.
Same logic. Two decision contexts.
The workshop knows that a vehicle of this type, at this mileage, frequently requires brake work. AutoHalo structures that knowledge into a probability range communicated to the customer at booking — not as a surprise when the car is on the lift.
Before a vehicle enters the maintenance queue, the system knows — based on operational profile and historical deviation patterns — which components are likely to require action. Spare parts are pre-ordered. Availability increases. Mission-critical capacity is preserved.
The engine uses Bayesian updating — a self-correcting statistical framework that produces reliable probability estimates even from limited initial data. Confidence intervals narrow as observations accumulate. No requirement for homogeneous datasets or OEM integration.
Sequenced market entry.
Defence & High-Readiness Fleets
Centralised vehicle data ownership. Defined maintenance cycles. High consequence of failure. The structural conditions for validation exist — unlike the fragmented automotive aftermarket.
Automotive Workshops
6 Stockholm-region workshops validated the problem. Fragmented data, high adoption barrier, but a massive market. Entry follows the defence pilot with a validated engine.
Built on empirical ground, not slideware.
Empirically Grounded
Field research across 6 Stockholm workshops and defence maintenance contexts. Not a theoretical model — validated against operational reality.
Defence Access
Direct relationships with Swedish Armed Forces decision-makers and procurement stakeholders. The beachhead is not theoretical — it is accessible.
EDHEC Research
The governance layer architecture is grounded in EMBA-level research conducted at EDHEC Business School, supervised by Dr. Serge Besanger.
Henrik Karlsson is the founder of AutoHalo and EMBA candidate at EDHEC Business School. With a background in the Swedish Armed Forces and a career focus on decision architecture under uncertainty — in both operational and commercial environments — AutoHalo was founded on the insight that the maintenance problem is a decision problem, not a technical one.
Open the conversation.
We are seeking a pilot partner to validate the governance layer in an operational defence context. Discussions are open.
AutoHalo is seeking investors and strategic partners ahead of the first defence pilot and automotive validation phase.