Q AOSIQ

IBM i Knowledge Risk Assessment

Your IBM i knowledge is retiring. Capture it before it walks out the door.

I find which of your systems depend on one or two people, capture what they know, and make it safely searchable. Fixed fee. Four to six weeks. Read-only from start to finish.

Request a 30-minute call Score your knowledge risk

69%

of IBM i professionals surveyed list skills among their top concerns. It now ranks #1, ahead of cybersecurity (64%).

Source: Fortra, 2026 IBM i Marketplace Survey (315 respondents)

Built for IBM i (AS/400, iSeries) shops in every industry Infor XAInfor LXJD Edwards WorldVAI S2KManhattan PKMSMEDHOSTRentalManJack HenryCustom RPG

The problem

The person who knows how month-end really closes is planning to retire.

Most IBM i shops run on decades of custom RPG, CL and job schedules that one or two people truly understand. The documentation is thin. The real knowledge lives in their heads.

When they leave, every change gets slower, riskier and more expensive, and a replacement is hard to find.

AI code tools can read your source. They can't read your people.

  • Order-entry rules no one wrote down
  • EDI maps with your trading partners
  • Overnight jobs that fail quietly
  • The month-end close sequence

Free scorecard · 5 minutes

Not ready to talk? Score your IBM i knowledge risk.

Ten statements. Answer honestly: Yes, Partly or No. Your score updates as you go, and nothing is sent anywhere unless you choose to.

1More than one person can fix the nightly batch when it fails at 2 a.m.
2Our most important RPG and CL programs have current written documentation.
3We know which programs and files each core business process depends on.
4No one who holds critical IBM i knowledge plans to retire or leave in the next 3 years.
5We could fill an open RPG or IBM i admin role within 90 days.
6Our job schedule and its dependencies are written down outside the scheduler itself.
7Every interface (EDI, FTP, APIs) is listed with an owner and what breaks if it stops.
8All changes go through change management, and we can see who changed what.
9Our backup, high-availability and recovery steps were tested in the last 12 months and are written down.
10Changes to our code, vendor or in-house, are tracked, so an upgrade or fix won't silently break them.

Answer all ten to see your score. 0 of 10 answered.

What you get

Five deliverables, one fixed fee.

01

System inventory

A read-only map of your jobs, programs, files and integrations, built from what is actually running.

02

Knowledge-risk map

Which systems only one or two people understand, ranked by what it would cost the business if they left.

03

Searchable knowledge base

A working pilot: your docs, runbooks and expert interviews, searchable in plain English, with every answer citing its source.

04

Governed AI access plan

How AI can safely help your team on IBM i: read-only access, human approval for anything that matters, and a full audit trail.

05

Roadmap and quote

Phased next steps in priority order, with a fixed quote for each. You decide what, if anything, comes next.

How it works

Four steps, mostly on your schedule.

  1. Step 1 · Free

    30-minute call

    We talk through your systems, who holds the knowledge and what's at stake. If it's not a fit, I'll say so.

  2. Step 2 · Weeks 1–2

    Read-only discovery

    You run my read-only collector yourself, or give me read-only access. Your choice. Nothing in your applications is changed. At the end of week 2 you can stop, keep the inventory, and pay only the first part of the fee.

  3. Step 3 · Weeks 2–4

    Knowledge capture

    Short interviews with your experts, who review what's written about their systems before it goes in. The goal is fewer 2 a.m. calls for them, not replacing them.

  4. Step 4 · Weeks 4–6

    Report and roadmap

    I walk your leadership through the findings, the risk map and a fixed quote for next steps.

Sample report

Here's what the knowledge-risk map looks like.

A full sample report, following a fictional distributor through the inventory, the risk map and the roadmap, is in production. Ask for a copy and I'll send it as soon as it's ready.

Ask for the sample report

Knowledge-risk map

Illustrative sample · fictional company

System People who know it Risk
Order entry and pricing 1 High
EDI with trading partners 1 High
Warehouse batch jobs 2 Medium
Month-end close 2 Medium
Customer master maintenance 4 Low

Safe by design

Built so it can't change anything.

Read-only, always

Enforced by your own IBM i object authority, not by a promise: a profile with read-only access, and a query pack that an automated test proves contains only reads. Nothing in your application libraries changes. The only output is written where you choose, and you can delete it.

You control your data

You decide what, if anything, leaves your network. The knowledge base can run on a model hosted inside your environment.

Approved and logged

Human approval gates and a tamper-evident audit trail, built on AOSIQ's governance runtime.

No migration agenda

Whether you stay on IBM i, modernize or move off it, knowing what you have comes first.

Who it's for

Any business that can't afford to lose what its IBM i experts know.

Manufacturers and distributors, hospitals and health plans, insurers, banks, logistics firms, retailers and local governments.

  • IBM i runs your orders, inventory, claims, billing or finance in production.
  • One or two people hold knowledge nobody else has.
  • Someone is retiring soon, or an RPG opening has gone unfilled for months.
  • You want to use AI, but only in a way you can control.
Stephane Jolicoeur, founder of AOSIQ

About

26 years on IBM i, seen from three sides.

I'm Stephane Jolicoeur, founder of AOSIQ. I started on the vendor side, supporting a major billing platform on IBM i. I then migrated client data between IBM i, Oracle and SQL Server. For the last two decades I've run a large IBM i estate in production, with 20 years of releases and no rollbacks.

I build the tools I use: read-only IBM i access, search that cites its sources, and AI agents that ask before they act.

AOSIQ is deliberately small. You work with me directly: I run the discovery, conduct the interviews and write the report myself, so nothing is handed to a junior team.

Based in metro Atlanta, Georgia, working with IBM i shops across the Southeast and remotely nationwide.

Stephane Jolicoeur on LinkedIn (opens in a new tab) →

Articles

What I've been writing about.

Practical notes on putting AI near production systems safely: retrieval you can trust, access you can control, and knowledge that outlasts the people who hold it.

Follow on LinkedIn →

  • LinkedIn article ·

    The 2 a.m. Problem: RAG and Secured MCP for Faster Issue Resolution

    Most of an outage is spent looking, not fixing. How well-designed retrieval and properly secured MCP put trustworthy evidence in front of engineers in seconds, with security boundaries that hold even when the AI gets something wrong.

    Read on LinkedIn (opens in a new tab) →

After the assessment

If you want help going further, there are two ways.

Build and hand-off

I build the full knowledge base and governed AI access for your team, priced by milestone, then hand it over with training.

Fractional architect

An experienced IBM i architect on a monthly retainer, for when your team needs a second set of senior eyes.

FAQ

Common questions

Do you need access to our production system?

Never write access. You can run my read-only collector yourself and send me the results, or give me read-only access. Your choice.

What does the collector read, and what access does it need?

Metadata only: release and PTF levels, libraries, objects and programs (names, types, owners, sizes, dates), database structure (tables, columns, indexes, constraints) and source member names and line counts. It never reads source code, business records, job logs or spool files. Row counts come from catalog statistics, not from reading your tables.

It needs one dedicated profile with *USE on the libraries in scope and no special authorities: no *ALLOBJ, no *SECADM. It connects over TLS and nothing is installed on your system. On production it runs at a gentle pace by default, and nothing leaves your network until you've reviewed it. If the profile can't see a library, the report says "incomplete," never "empty." The profile is disabled at the end of the engagement, and I confirm it in writing.

You can read every line before anything runs: I give you the collector's full source, including every query it will send.

Will you sign our NDA or BAA?

Yes. I sign your NDA before discovery starts, and a business associate agreement (BAA) if you're a covered entity under HIPAA. At the end of the engagement I return or delete everything I hold, as you choose, and confirm it in writing.

We already use an AI code tool like IBM Bob. Why would we need this?

Good. Code tools explain what your programs do. They can't tell you which people and processes the business depends on, or capture what was never written in code: the month-end sequence, the EDI exceptions, the job someone restarts by hand. The assessment finds those risks, captures that knowledge, and works alongside whatever tool you choose. I don't resell any of them.

Will our data go to an AI company?

Only if you approve it. The knowledge-base pilot can run on a model hosted inside your own environment.

How much does it cost?

A fixed fee, quoted after the free call once I understand the size of your environment. No hourly billing and no surprises.

What if we want to stop partway?

You can. The fee is paid in two parts. After the system inventory at the end of week 2, you can stop, keep the inventory, and owe only the first part.

Are you going to push us to migrate off IBM i?

No. The assessment is useful whether you stay, modernize or migrate. I don't sell migrations.

How much of our team's time does it take?

A few hours from your IT lead, plus short interviews with the people who hold the knowledge, scheduled around their work.

Free 30-minute call

Start with a free 30-minute call.

Tell me a little about your IBM i environment. I read every message myself and reply within two business days.

Prefer email? admin@aosiq.com

I never share your details. Privacy policy