The agent proposes. Your operator approves. The record keeps itself.

KLM Consulting builds supervised AI agents for energy and industrial operations — and the systems architecture they have to live inside. We take the work ourselves; nothing is brokered out.

Houston, Texas  /  founded 2015
21 years of architecture for banks,
insurers and energy operators
NominationNOM-88214 / email Berth schedulerev 41 Tank 214 gauge06:00 reading Agent berth-exception ground reason check loop reads only what it was pointed at finding berth 3 committed to MV Anwen — 19 h overlap Approval gate R. Salazar, terminal scheduler approved 06:24 CT, unchanged nothing downstream moves until a named person signs Scheduling system rev 42 / marine ops notified Decision record 4417 / inputs retained elapsed 06:12 → 06:26 CT demurrage exposure avoided
drag the drawing sideways to see all of itWorked example — marine terminal, berth exceptionLevel 2 autonomy
A.01 — Diagnosis

Most AI work stalls in one of three places.

We get called in after the pilot. The model is usually fine. What failed was everything around it — and it fails in a small number of recognisable ways.

Knowing which one you are in changes what the next eight weeks should cost.

It cannot reach your records.

The demonstration ran on a clean extract. Production runs on thirty years of scanned tickets, four naming conventions, a mainframe export and a spreadsheet one person maintains. Nothing was ever built to reach across them, so the agent answers confidently from a fraction of the truth.

Nobody will put their name on it.

Accurate enough to be interesting, not accurate enough for a scheduler, a measurement analyst or an HSE lead to sign. With no defined approval step and no named approver, there is no route from interesting to operational, and the pilot renews forever.

It has no way to act.

Reading is the solved part. Writing an approved change back into the scheduling system, the ERP or the ticket queue is an integration project with its own permissions, failure modes and rollback — and it was never in the budget. So a person retypes what the agent found.

A.02 — Mechanism

Six parts. Five of them are not the model.

An agent is a small amount of reasoning wrapped in a large amount of plumbing. We build the plumbing first, because that is where the risk and the cost actually sit.

Each part below is something you can inspect, test and switch off independently.

Intake email, file drop,queue, webhook, cron Grounding your records, indexed;every claim is cited Reasoning the model, swappable,scored on your cases Tools named actions only,each one permissioned Approval gate a named person, withthe evidence attached Ledger inputs, sources,actions, approvals every stage writes to the ledger — the run can be reconstructed later without asking the model anything the part you can defend
drag the drawing sideways to see all of itReference pipeline — applies at every autonomy levelA.02 / rev 1
A.03 — Autonomy

Pick the level you can defend on a bad day.

Autonomy is not a maturity score to climb as fast as possible. It is a decision about what a mistake costs, made once, in writing, before anything is switched on.

level 1

Read and answer

The agent reads across your records and answers questions, with the source page attached to every claim. It changes nothing. This is where a well file search, a drawing lookup or a contract question belongs — and it is where almost every engagement should begin, because it earns trust on read-only ground.

cost of a mistake — a wrong answer, caught by the person who asked
level 2

Draft and route

The agent prepares the action — the reschedule, the reconciliation, the permit check — and hands it to a named approver with its reasoning and its evidence. Edge cases route to a human. They are never resolved quietly. Most of the value we have delivered sits at this level.

cost of a mistake — a rejected draft, and one more test case
level 3

Act inside a fence

The agent executes within limits you wrote: which actions, which value ranges, which counterparties, which hours. Anything outside the fence stops and escalates. Everything it did is reversible and logged. We recommend this only for narrow, high-frequency work where the fence is genuinely enumerable.

cost of a mistake — bounded, reversible, and visible the same day
read draft + route act inside awritten fence 123 blast radius throughput
We have never advised a client to start at level three. Not once in twenty-one years has the constraint been the model's capability. It has always been whether the organisation could live with being wrong at that speed.
A.04 — Application

Where an agent pays for itself.

These are jobs with a clear trigger, a bounded set of records to consult, and a person who already signs off today. That combination is what makes an agent worth building; its absence is what makes one a science project.

terminals & movements

Berth, yard and rail exceptions

Watch nominations against berth windows, tank availability and prior line-ups. Surface the clash while it is still cheap, propose the reshuffle, and put it in front of the scheduler with the conflict shown rather than asserted.

measurement

Ticket and movement reconciliation

Reconcile meter tickets against recorded movements, flag unaccounted volume the day it appears instead of at month end, and attach the gauge reports and tickets that produced the variance.

engineering records

Well files and drawings

Ask across decades of scans in plain language. Every answer returns the page it came from. Read-only, which is why it is usually the first thing we build.

commercial

Tender and contract intake

Pull commercial terms out of a bid pack, mark what changed between revisions, and write the summary into the system of record rather than a mailbox.

HSE & compliance

Permit and certification checks

Read the request, verify contractor certifications and the isolation list, route to the authorised signatory. Issuance stays with a person, always.

back office — the same pattern, lower stakes

Invoice and statement matching

Match, flag the exceptions, and leave the clean ones alone.

Shift handover assembly

Draft the handover from the day's records; the outgoing lead edits and signs.

Vendor performance alerting

Notice the pattern across purchase orders before the renewal conversation, not after.

A.05 — Evidence

What you hand the auditor.

Everyone in this market promises guardrails. Almost nobody will show you what the guardrail actually produces. Here is the record written by the run in the drawing at the top of this page — the real shape of it, with the names changed.

record 4417 agent berth-exception level 2 opened 2026-03-04 06:12 CT closed 06:26 CT
trigger
Nomination NOM-88214 received — MV Coral Trader, laycan 06–08 Mar.
read
Berth schedule rev 41. Tank 214 gauge report, 06:00. Prior three movements, same counterparty. Nothing else was in scope.
found
Berth 3 committed to MV Anwen until 07 Mar 14:00. Overlap of 19 hours.
proposed
Move Coral Trader to Berth 2; shift the 07 Mar line-up back four hours.
declined
Any change touching the Anwen commitment — outside the fence.
routed to
R. Salazar, terminal scheduler. Evidence attached.
approved
06:24 CT, unchanged.
written
Scheduling system rev 42. Notification to marine ops. Both reversible.
Every line above is reconstructable from retained inputs without re-running the model.

Three things this settles before they become arguments:

01

What it saw

The record names the documents in scope. If an answer was wrong, you can tell immediately whether it was the reasoning or the retrieval — and those have different fixes.

02

What it refused

The declined line matters more than the approved one. It is the proof that the fence exists and held.

03

Who signed

A person, by name, at a time, with what they were shown. Accountability does not move to the software, and no regulator has ever accepted that it did.

A.06 — Underneath

The unglamorous half of the work.

An agent is a thin layer over whatever integration you already have. Where that layer is thin, the first job is not an agent at all — and we will say so before you spend anything.

This is the work KLM did for a decade before any of it was called agentic, and it is still most of what we are hired for.

Platform and data architecture

One place the applications agree on: the model, the lineage, the access rules, and the answer to who owns which field. Most reconciliation pain is an architecture problem wearing a reporting costume.

Integration between systems that will not change

Events, contracts and queues between the scheduling system, the ERP, the historian and the twenty-year-old application nobody is allowed to touch. Designed so a failure in one does not take the others with it.

On-premises to cloud, without a big bang

A sequenced path with a working system at every step, costed before it starts. We have never recommended a rewrite where a strangler pattern would do.

Architecture review and second opinion

A written answer to what to build, what to buy, what to retire, and where the operating cost lands in year three. Delivered as a document you can hand to a board, not a slide deck.

Working languages and platforms — .NET 8 / C# Azure Azure Kubernetes Service Hot Chocolate GraphQL Apache Kafka Elasticsearch Azure DevOps Microsoft 365 Google Workspace
and the model and retrieval services layered on top of them.
A.07 — Engagement

Three ways to start, all of them scoped.

Every one of these ends with something written that survives the engagement: a decision, a running system, or an architecture your own people can maintain after we are gone.

012–3 weeks, fixed fee

Assessment

We look at the actual work, not the org chart, and come back with a written answer on where an agent pays for itself here and where it will not.

  • Candidate workflows, ranked by payback
  • The autonomy level each one can defend
  • Build and run cost, annualised
  • What to do instead, where that is the honest answer
026–10 weeks

First agent in production

One workflow, at level one or two, live — with its approval step, its evaluation suite and its ledger. Not a demonstration environment.

  • Grounding built against your real records
  • A named approver and a defined route for edge cases
  • Test cases written from your own history
  • Handover to your team, with the runbook
03monthly

Architect on retainer

Standing architecture ownership for organisations that need the judgement without the headcount. The same person every month.

  • Design authority on new work
  • Review of vendor and platform decisions
  • Roadmap with operating cost attached
  • Available to your team, not only your executives
A.08 — Principal

You are hiring the person who does the work.

KLM is deliberately small. The architect you meet is the architect who writes the design, sits in the review, and is still reachable when something breaks eighteen months later.

photograph
to be added
Landry Kammogne
Founder & chief executive, Houston, Texas

Twenty-one years in software and enterprise architecture, built on work for American banks, insurers and energy companies. Landry founded KLM Consulting in 2015 and has led it since, taking the client work himself rather than brokering it out.

He currently serves as platform and enterprise architect across a portfolio of oil and gas engineering applications — owning the architecture end to end: the data platform beneath it, the integration between the applications, and the path from on-premises software to cloud services.

Before that he was principal technical architect on the dock and export platforms of a Houston midstream operator, built on .NET 8, Hot Chocolate GraphQL, Kafka, Elasticsearch and Azure Kubernetes Service, and used every day by the teams scheduling vessels and moving cargo. The worked example at the top of this page is drawn from that world.

Earlier came architecture and principal consulting roles at IBM / Neudesic, Improving Enterprises, Hitachi Consulting and CSC, delivering systems for banks, insurers and Fortune 500 operators across Texas and the Midwest.

Clients bring him the questions with no obvious answer: what to build, what to buy, what to retire, and where AI pays for itself rather than adding another subscription. He answers them in writing, with the operating cost attached.

MBA in management and finance, Avila University, Missouri. Doctoral research in business administration, in progress.

Azure Solutions Architect Expert (AZ-305) Certified Kubernetes Administrator Certified Kubernetes Application Developer Professional Scrum Master I

Send us the workflow that keeps going wrong.

One paragraph is enough to tell whether this is an agent, an integration problem, or something you should not automate at all. You will get a straight answer either way, and it costs nothing to ask.

email
info@klmconsultingusa.com
telephone
+1 469-656-8057
based in
Houston, Texas — working across the United States