In short: The honest test of any judgment system, and of anyone who installs one, is whether it keeps working after they leave. A system whose operators cannot be removed without the record degrading has the same flaw it was built to fix.
Key points
- The organisation should own the records, sources and accounts outright, not hold a licence to view them.
- Everything should export with history, supersession links and outcomes in a format readable without the original tools.
- Every engagement should end with a handover pack of the builder's own reasoning, including rejected approaches.
- A suggested maintenance rhythm: weekly decision review, quarterly falsifier sweep, annual read of what keeps going wrong.
- Test a record by exporting it and having someone else run a week of routines from the documentation alone.
Most enterprise software, and most advisory work, is built to be needed. The vendor holds the data, the model and the definitions. The consultant holds the context. Leaving is expensive by design, and that expense is part of the business model.
A judgment record argues for the opposite. Its whole value is that reasoning survives the people who produced it. That has to include the people who built the system. If the record only makes sense while a particular vendor, team or advisor is present to interpret it, it is not institutional memory. It is a dependency with good documentation.
What "kept without us" requires
Ownership. The organisation owns the records, the sources and the accounts. Not a licence to view them; ownership.
Portability. Everything can be exported, with history, supersession links and outcomes, in a documented format that can be read without the original tools.
Operability. The routines that keep the record alive, capture, conflict adjudication, the falsifier sweep, are documented and can be run by the organisation's own people.
A handover at the end. Every engagement closes with a handover pack of the builder's own reasoning: what was recommended, what was rejected and why, what to watch, and the graveyard of approaches tried and dropped during the engagement.
The honest caveat: a static record depreciates
Ownership is necessary and not sufficient. A judgment record that nobody maintains loses value. Precedents age. Conditions change. Views that were current become superseded in the world but not in the record, and an unmaintained record starts to mislead rather than inform.
So say it plainly, in the same breath as handing it over: you own this, and it depreciates unless someone keeps it current. Then give the organisation what it needs to do that itself, a maintenance rhythm small enough to survive: a weekly decision review, a quarterly falsifier sweep, an annual read of what keeps going wrong.
The standard we hold ourselves to
At lowob the client owns the records, the sources and the accounts, and can export everything with its history at any time. Every engagement ends with a handover pack of our own reasoning. Cancellation never holds a client's memory hostage. And when we hand the record over, we say clearly that it is a depreciating asset without maintenance, and show how to maintain it.
Handing over the memory gives up one reason to be rehired. That is deliberate.
The bet
Organisations rehire the people and products that left them better off, not the ones they could not get rid of. In a market where every vendor can reach the same data through the same protocol with the same models, trust earned that way may be the most durable advantage left.
The enterprise brain worth having is the one that can show what it decided, why, what would have proved it wrong, and when it changed its mind, and that still can, after everyone who built it has gone.
In practice
- Test whether your judgment record survives the removal of its builders: export it, and have someone else run a week of routines from the documentation alone.
- Put ownership, export and exit terms in every contract that touches decision records.
- End every engagement with a handover pack of the builder's own reasoning.
- Tell clients, in writing, what maintenance the record needs to keep its value.
What would prove this wrong. Evidence that clients consistently prefer, and get better outcomes from, judgment systems operated entirely by vendors, with no ownership or portability, over several years.
Common questions
How do I avoid vendor lock-in with a decision record or knowledge system?
Put ownership, export and exit terms in every contract that touches decision records, and make sure everything exports with its history in a documented format.
What should a consultant hand over at the end of an engagement?
A handover pack of their own reasoning: what was recommended, what was rejected and why, what to watch, and the approaches tried and dropped.
Does a judgment record keep its value once it is handed over?
No. An unmaintained record depreciates and starts to mislead. It needs a maintenance rhythm such as a weekly decision review, a quarterly falsifier sweep and an annual review.
How can I test whether my knowledge system depends on the people who built it?
Export it and have someone else run a week of its routines from the documentation alone.