Every answer arrives with its evidence attached.
This is the block valve you opened on the line above. Ask about it in plain language and Muono returns the reading, the record, the calculation and the document each one came from.
Put your own numbers in.
Somebody has to take this to the person who signs, and a page of capability will not do it. So here is the arithmetic on your figures rather than ours. Every input below is yours to change, and the working is shown so your finance lead can repeat it.
This is the size of your problem, not a claim about our effect on it. Muono does not recover a barrel or take a minute out of a changeover. It shortens the distance between a number moving and the reason for it. What share of that you convert is yours to estimate, and it is the number worth arguing about first.
Five surfaces on one model.
The twin is the one people ask to see first. The other four are where the shift actually happens.
One model of the site that changes as the site changes: assets, sensors, documents, calculations, open work.
- Tags in model
- 41,280
- Source systems
- 6
- Model age
- 12 s
Where Muono runs
Written for the people who answer for barrels.
Upstream, midstream and refining argue about the same three things: rate against plan, remaining life on a line, and what goes into the next turnaround. All three are answerable from data the site already holds. It just sits in the historian, in SAP and in a folder of ILI reports, so the answer takes most of a day to assemble and is stale by the time it lands.
Deferred production has a cause. Usually it is already written down.
Choke position, separator level and water cut live in three systems and one engineer's head. Put all three on the same well and a drop in rate arrives with its cause attached. The work order that answers it can be drafted before the morning call rather than after it.
- Choke position
- 34%
- Water cut
- 18.2%
- Deferred volume recovered
- 0 bbl
Differential pressure tells you something changed. The ILI run tells you what.
Line pressure, pig runs, ILI history and the permit covering the repair all hang off the same KP. Your integrity engineer stops assembling PDFs before every deferral meeting. What eventually goes to the regulator is read off the same record the control room is looking at.
- Throughput
- 8,240 bbl/d
- Differential pressure
- 0.42 bar
- Inspection tool
- KP 12
Turnaround scope, argued with data instead of memory.
Vibration, bearing temperature and load are already trending on your critical rotating equipment. Read together, they give the reliability engineer a window: weeks before a planned bearing swap turns into an unplanned outage. That is enough time to get the job into turnaround scope with its readings attached.
- Overall vibration
- 2.1 mm/s
- Bearing temperature
- 68 °C
- Intervention window
- 21 days
Output, curtailment and availability all have to agree.
SCADA output, the dispatch instruction, the availability declaration and the outage that drove it are held once and read the same way by the control room, the asset manager and settlement. When a turbine comes back from an outage early, the figure that proves it comes off the same model that scheduled the work.
- Turbine output
- 1.8 MW
- Wind at hub height
- 7.2 m/s
- Curtailment this hour
- 12%
- Availability
- 91%
What changes on your shift.
Muono is bought by a site and used by six people who each need a different thing from it. This is what each of them stops doing by hand.
Find the column your organisation would accept and read down it. Where a certification is still in progress, the row says so.
Built around the systems you already run.
Muono connects to the registers, historians and document stores your organisation runs today. Where information is still captured by hand, field teams enter it on mobile and it is available site-wide immediately. Nothing is ripped out to make room.
You have been sold this before
- We already have a historian, a data lake and a unified namespace. What is different here?
- Those store. They do not answer. What usually stalls them is that nobody agrees what a tag means, so the model never gets trusted enough for anyone to decide from it, and it becomes a place data goes rather than a place work happens. Muono's model carries a plain-language description on every entity and versions every change, so meaning travels with the data and an answer can name the system behind each figure. If your existing layer already does that, you do not need us and we will say so.
- The last single source of truth project died in data cleansing. Why would this not?
- Because that one almost certainly tried to model the whole site before anybody could use any of it, and the value arrived after the patience ran out. This starts on one unit and answers one real question inside two weeks. If the record is not trustworthy your own engineers find out in week three, not in month nine, and the money you have spent by then is a pilot rather than a programme.
- How is this not another AI pilot that never leaves the lab?
- A lab pilot fails because nobody will act on an output they cannot check. Every answer here shows the readings it used and the arithmetic it ran, and every agent runs at an authority level your team wrote down, with anything irreversible stopping for a named person. That is the difference between something an operations manager will put on shift and something that stays a demo.
- Our OT team will not allow this near the process network.
- Then scope it that way. Many deployments never go below Purdue Level 3 and read the process historian instead, and the collector initiates outbound connections only. The security specification states the position per connector rather than making one claim for every site, and it is written to be sent to your OT reviewer before any commercial conversation.
- What happens when the person who set it up leaves?
- This is the failure mode nobody puts in a proposal and everybody has lived through. The model documents itself, so there is no separate dictionary going stale on a drive, and local practice is recorded against the equipment it applies to rather than in one operator's head. A new engineer reads the model instead of finding out who to ask.
- You have no customers and no certifications. Why would we go first?
- Both true, and both stated plainly elsewhere on this site rather than buried. What you can test in a pilot is the architecture, the deployment model and the audit trail, which are the things a certificate attests to anyway. If your policy requires a certificate before anything starts, tell us on the first call and we will tell you honestly that we cannot meet it yet.
Measured on your plant, not ours.
Every row below is empty, and it stays empty until a pilot closes a baseline on a real site. We could fill it with an industry average today and you would have no way to check it. The method is the part we can publish now, so that is what this is.
Two weeks to the first sourced answer.
Every step below is a working system on your own data, not a slide. Simple cases run in weeks. Custody transfer, emissions against permit and anything a partner will dispute take months.
- Week 2One unit, answeringHistorian and asset register connected. Ask about conditions in plain language and get the reading and calculation behind the answer.Read-only · no change to systems
- Week 6Documents and open workP&IDs, procedures and work orders attached to the assets they belong to. First predictions on rotating equipment, with the evidence.SAP · document store · historian
- Week 12Supervised actionAgents draft the work and stop at your approval gate. The unit is walkable in 3D, every object opening to its own record.Limits signed by name · fully logged
- Month 4The hardest casesMulti-unit model, custody transfer figures, emissions against permit, and closed-loop routine work running inside written limits.Site-wide · audit-ready
A paid pilot on one unit, then a licence per site.
The pilot is scoped to a single unit so the cost is small and the result is unambiguous. What it costs depends on how many systems it has to reach, so we quote it after the scoping call rather than before.
Start on one unit. Prove it there.
A guided pilot connects one unit to Muono, on your data, with your limits written down. If it does not hold up in front of your engineers, it does not go further.