AI Assisted Solver MRO 日本語 Demo

Aircraft engine maintenance planning | Aether Platform × Secretary.io

Every plan drifts. Decide in advance how you will respond when it does.

Longer turnaround times at MRO (maintenance, repair and overhaul) shops, late LLPs (life-limited parts), unscheduled removals. Instead of rebuilding the plan every time the assumptions move, keep a plan for each scenario and a clear rule for when to switch plans, and decide which plan to follow at the monthly review meeting (the engine review). AI agents do the calculations and prepare the meeting pack.

Try the demo Talk to us See the use cases we tested

A 24-month shop-visit plan, viewed by scenario

Probability reality is closest to this scenario 62%
Decision Stay on plan

To decide this month: 1 induction slot booking, 1 LLP kit order.

Removal window Shop visit (TAT) Drift from assumptions Illustration based on synthetic data
  • Model built and tested on the fleet and contract terms of two major carriers (synthetic equivalents)
  • Decision points identified across 11 engine management processes
  • Runs in closed networks (Aether Platform)

Does this sound familiar?

Executive and finance

  • Year-end closes in, and nobody can say, as a range, how far maintenance spend may exceed the budget
  • The plan is rebuilt every year, so there is no way to check whether last year's decisions were right
  • Buying or leasing spare engines, renewing shop contracts and investing in a domestic engine shop cannot be compared on the same basis

Maintenance planning, engineering and procurement

  • Longer shop TAT and late parts break the plan month after month
  • How much additional work found at teardown to approve depends on who is asked
  • Induction slot bookings, LLP kit orders, reviews of service bulletins (manufacturer modification instructions): decision deadlines are all over the place and some get missed
  • Preparing the review meeting pack and recording what was decided takes too much time

The problem is not how plans are built. It is how decisions are made and checked.

Existing maintenance information systems (MIS) and forecasting tools are good at building plans. But what to do when the assumptions drift still depends on individual experience and meetings, and there are almost no tools to check whether those decisions were right. MRO AI Assisted Solver builds both into the day-to-day workflow.

Check
Can we trust how we decide? Do the long-term and short-term plans connect?
Compare past plans with actuals. Check the link from the annual budget down to the monthly plan. Feed actuals back into the next plan's assumptions
Most teams have no tool for this
Decide
Which assumptions do we plan on? What do we do if they drift?
A plan for each scenario. Update the probability of each scenario from actuals, and show the cost or benefit of a plan switch and its decision deadline
Relies on experience and meetings
Build
How many, how much, and when
Optimize shop-visit timing, workscope, shop selection and spare engine count with the uncertainty of cost overruns, TAT delays and unscheduled removals factored in. Show the year-end outturn as a range
Works with your existing tools

PDCA for the review meeting, OODA for the day-to-day

Engine planning runs on two loops. PDCA looks far ahead, through the annual plan version and the review meeting. OODA follows the facts in shop quotes and alerts and updates the plan. MRO AI Assisted Solver connects both loops with the same scenarios and simulation, so the review meeting can focus on decisions that go beyond the rules.

PDCAAnnual plan version, monthly review meeting

Look far ahead, early and broadly

  1. PPlanOnce a year, lock in the 24-month shop-visit plan as a version, together with a plan for each scenario and the rules for switching plans
  2. DDoAt the review meeting, decide slot bookings, kit orders and budget revisions, and turn them into tasks with deadlines
  3. CCheckEvery month, check how actuals match the plan, which scenario reality is closest to, and the cost or benefit of a plan switch
  4. AActCompare past versions with actuals, and correct the assumptions that drifted (TAT, lead times, failure rates) and the decision rules in the next version
OODAOn the spot (same day to a few days)

Follow the facts, update the plan on evidence

  1. OObserveEvery week, take in shop quotes and slot replies, parts lead-time replies, and engine health monitoring alerts
  2. OOrientDetermine which scenario the change fits. If it fits none, don't decide by rule; escalate it to the review meeting
  3. DDecideApply the rules agreed in advance at the review meeting: approve additional work found at teardown up to the limit the same day; on a health-monitoring alert, arrange a removal the same day if remaining life is short or peak season is near; if an engine's return from the shop is late, cover it with a short-term lease of a spare engine
  4. AActArrange it the same day and record the reason and the amount. Escalate to the review meeting when a plan switch is needed or the case goes beyond the rules

OODA results feed back into PDCA. The review meeting checks how many items were decided by rule, what they gained or cost, and how long it took to react to a change (time to detect, assess, decide and act). The next version then adjusts the rule limits and the assumptions.

What changes in the meeting room

Executive and finance

"Looks like we'll go over budget again this year."

"The probability of going over budget is about 60%. This is the contingency needed to stay within budget with 80% probability." You talk in ranges and evidence.

Maintenance planning

"TAT is getting longer. Let's rebuild the plan."

"We're moving toward the MRO congestion scenario, but costs under the Baseline plan barely rise under congestion. Switching would cost more, so we stay." You have what you need to decide whether to act at all.

Procurement and contracts

"Should we add spare engines?"

"An engine pool agreement gives the same effect as two more spares at about one-sixth of the cost." You compare options on the same basis.

Running the monthly review

"Preparing the pack takes several days every month."

AI agents prepare the pack: the conclusion, what changed since last month, what to decide this month, and the year-end outturn. Decisions and decision deadlines are recorded. Your team spends its time checking and discussing, not building slides.

Figures are validation results on synthetic data estimated from public information.

We can also run estimates tailored to your fleet and contract terms.

Try the demo Discuss your fleet

How it differs from the way you work today

Spreadsheets and experienceMIS forecastsMRO AI Assisted Solver
Building the planEach planner builds their ownStrongOptimized with uncertainty factored in, and linked to your existing forecasts
Budget outlookA single numberA single numberProbability and range (probability of going over budget, P80)
When assumptions driftRebuild the planRebuild the planShows the cost or benefit of a plan switch and its deadline, from a plan for each scenario
Checking decisionsNot possibleHardly anyCompares past versions with actuals and feeds results into the next assumptions
Meetings and recordsManualNot coveredPrepares the pack automatically and records decisions, approvals and deadlines

Use cases we tested

Real business questions, answered with synthetic data estimated from public information.

  • Annual plan and monthly tracking

    Engine shop-visit planning and spare engine decisions

    On 737-800 fleets modeled on two major carriers, we tested a 2-year shop-visit plan, the probability of going over budget, spare engines versus an engine pool agreement, and plan-switch decisions when the assumptions drift.

    • About 60%Probability that next year's maintenance spend goes over budget (Carrier A, synthetic)
    • About 1/6Cost of an engine pool agreement that gives the same effect as two spare engines
    • Month 2Leading indicators flag early signs of MRO congestion. Even so, in this test switching away from the Baseline plan costs more in every month

    Aether PlatformSecretary.io

  • Capital investment and negotiation

    Staged investment in a domestic engine shop

    We evaluated an investment to bring next-generation engine maintenance in-country with a 20-year Monte Carlo simulation. We estimated the value of deciding in stages, the break-even number of third-party shop visits, and each party's share if three companies invest together.

    Aether Platform

  • Workflow automation

    Running the monthly review with AI agents

    Two agents, a planner and a reviewer, handle ingesting quotes, plan updates, review, and preparation of the review meeting pack.

    See it in the demo

    Aether PlatformSecretary.io

Built on Aether Platform and Secretary.io

We combine two products from the RE-X Integrated Development Support Solution and apply them to engine maintenance planning.

Aether Platform

Analytics execution platform

  • AI agents run optimization and simulation in a secure environment
  • Planners watch the agents' work and give instructions from a browser
  • A planner agent works alongside a reviewer agent that checks against contracts, requirement documents and manuals
  • Runs in closed networks

Secretary.io

Meeting and decision records

  • Records the review meeting and organizes decisions, approvals and action items automatically
  • Ingests quotes, teardown findings reports, manuals (PDF) and shared folders as a knowledge base
  • Notifies owners of items with approaching decision deadlines
  • Connects to your existing MIS and engine health monitoring data via MCP

  1. Ingest

    Ingest quotes and revisions, engine returns from the shop and invoice actuals, and parts lead-time replies

  2. Compute

    Agents update the plan, scenario probabilities and option comparisons, and prepare the pack

  3. Decide in the meeting

    Review the conclusion, changes since last month and what to decide this month, then decide

  4. Record and feed back

    Record decisions and deadlines and notify owners. Actuals flow into the next plan's assumptions

How we get started

  1. Consultation

    We learn about your current planning process and challenges

  2. PoC

    We build a model for your fleet and contract terms from public information and data you provide, and compare it with your current plan

  3. Production rollout

    We design the connection to your MIS and engine health monitoring data, and operation in a closed environment

  4. Ongoing support

    We run the monthly cycle with you and revise the assumptions based on actuals

Frequently asked questions

Can we start before our internal data is in good shape?
Yes. The PoC is built on assumptions estimated from public information, and any figure without a source is clearly marked as an assumption. We then replace assumptions with your internal data, starting with those that move the conclusions most.
Do we need to replace our existing maintenance information system?
No. We connect to your existing systems and engine health monitoring data via MCP and use their forecasts and actuals as inputs. This solution covers what to decide when the assumptions drift, and how to check those decisions.
Can we use it without data leaving our company?
Aether Platform supports operation in closed environments. We propose a setup where the AI agent runtime, knowledge base and meeting records stay inside your company.
How long does a PoC take?
It depends on the fleet and processes in scope. In an initial consultation, we will give you an indication of scope and timeline.
Are the figures on this page from real airlines?
No. They are validation results on synthetic data estimated from public information. They do not represent the actual plans or figures of any specific airline.

Start with a demo built around your questions

Tell us about your fleet and contract structure. Using our synthetic-data model, we will show how the solution answers your questions.

Book a demo

Opens the contact form on the RE-X corporate site (form in Japanese). Please mention "MRO AI Assisted Solver" in your message.