Glossary

An option that was on the table before a decision was made, including the ones ruled out. A Memolok Decision Record keeps the alternatives and the reason each one lost, not just the option that won, so a rejected idea does not return eighteen months later looking new.

A software team's one file, one decision convention: a short Markdown file with a title, a status, the context, the decision, and its consequences, committed alongside the code. ADRs inspired Memolok's format but stay software only, carry no status you can query, and record no measured expectation to check later. See what you use today.

One of the five ways decision debt costs an organization. A choice gets copied forward into the next project because it is already there, not because anyone checked it still fits, since its original justification is unavailable. See decision debt.

The practice of giving an AI system the right information at the right time: what to orchestrate, what to retrieve, what to remember, and what rules govern all of it. A Memolok Decision Record sits in the governance layer. See Context engineering.

One of the three pillars of Decision Intelligence: rules engines, ML models, and workflow systems that make or recommend decisions on behalf of the organization, such as credit scoring, fraud detection, or demand forecasting. See the Decision Intelligence hub.

The compounding cost of choices that were made, often carefully, and never recorded. It is also the root cause behind the more familiar version of the same phrase, the hard calls that never get made at all, since those usually stall for lack of anything earlier to stand on. See decision debt.

One of the five ways decision debt costs an organization. The world a choice was made for changes, and because no premise was recorded, nothing flags that the choice has quietly stopped being right. See decision debt.

The category covering how an organization uses data and structure to decide, and to remember what it decided. It has three pillars: Decision Automation, Decision Support, and the Decision Record. See the Decision Intelligence hub.

A written account of one decision: what was decided, what else was considered, what was known at the time, and what the decider expected to happen as a result. See the Product page.

One of the three pillars of Decision Intelligence: dashboards, scenario models, and analytics platforms that hand a decision-maker data about the world outside, such as the market, operations, or customers. Forward-looking, unlike the Decision Record, and not a record of the organization's own past reasoning. See the Decision Intelligence hub.

How Memolok classifies an observed outcome by whether it was foreseen when the decision was made.

• Expected: assesses something the decision already committed to, and comes back satisfied, violated, or inconclusive.
• Emergent: a result nobody could have reasonably predicted.
• Deducible: a result nobody logged in advance, but that could have been reasonably foreseen.

What a decider expected to follow from a decision, written down at the moment of commitment rather than remembered afterward, so it can be checked honestly against what actually happened. See the Product page.

Where a system's rules come from, not the trace of a rule being enforced. A Memolok Decision Record is a governance concept in that sense: not a rule executed at machine speed, but the reasoning a rule can be checked against. See Context engineering.

A short, freely revisable statement of what a Memolok Decision Ledger is for: its purpose, its audience, and the rough direction it is heading. It orients the next decision rather than governing it; a sealed Memolok Decision Record never leans on it to justify itself, so updating the intent never rewrites the past.

The gap between what a decision expected to happen and what actually happened. One miss is a data point; a pattern of misses across many decisions is a signal worth raising at a strategic level, not only in a retrospective.

Something raised for possible decision work, in the raiser's own words: a problem, an idea, an opportunity, a mandate, or a concern ("I can't log in", "legal now requires seven-year retention"). A matter can be parked on the ledger with nothing decided yet, and picked up later. It may lead to one or more Memolok Decision Records, or be closed without one. See how Memolok works.

The collection of Memolok Decision Records for one line of reasoning: a person, a team, a project, or a whole organization. An organization can keep as many MDLs as it needs. Not the same as a decision log: a decision log is free text rows with no shared shape, while every record in an MDL has the same six parts and a status you can query. See the Product page and what you use today.

Memolok's structured, immutable record of a single decision: the need, the alternatives and what ruled them out, the premises it rested on, the commitment, the expected outcome, and what it deliberately left open. See the Product page.

The open protocol Memolok's plugin uses to read and write a decision ledger directly inside Claude Code, Claude chat, and Cowork, through a tool interface rather than by scraping pages. See decision records for AI agents.

What had to be true for a decision to succeed, stated sharply enough to be tested afterward, not a vague goal like "improve performance." See the Product page.

Something a decision deliberately left unsettled, recorded as its own item rather than lost in prose, so a later reader can tell a choice from an oversight. See the Product page.

One of the five ways decision debt costs an organization. Nobody will touch a decision because nobody knows what it was load-bearing for, so the safest available action is to build around it. See decision debt.

One of the five ways decision debt costs an organization. An issue that was never decided but that everyone treats as settled, with no verdict to look up because none was ever reached. See decision debt.

The flag Memolok raises on a decision that relied on a fact later corrected in the world almanac. It does not mean the decision was wrong, the choice was reasonable given what was known at the time, only that its foundation has since shifted and is worth a second look.

A fact the choice rested on: a budget, a contract term, a regulation, a supplier with no real competitor. Kept separately from the decision itself, so when a premise turns out to be wrong, every choice standing on it can be found. See the Product page.

A disposable working note inside a Memolok Decision Ledger: a pasted quote, a scraped page, rough numbers, an idea that may go nowhere. Unlike a decision record or a world fact, a scratchpad makes no promise and can be rewritten or deleted at any time.

What happens to a decision instead of a silent edit. Later thinking does not rewrite an approved record, it supersedes it, and both stay on the ledger, each honest about the moment it was made. A supersession chain turns "is this still the decision?" into a question with an answer rather than a search. See the Product page.

A real decision, sealed as Accepted or Rejected, with the bare minimum behind it: one option considered, little or no argument recorded. Still a genuine commitment with a verdict on record, not to be confused with a phantom decision, which was never decided at all. See the Product page.

The commitment itself: "we chose X." The verdict is only part of the decision, which also includes the options ruled out and the reasons X was the only one left.

Everything that happens in the world after a decision is committed. A Memolok Decision Record is sealed at the moment of the decision, so what follows, including whether it worked, is recorded separately, as one or more observed outcomes linked back to the record that caused them.

The collection of world facts a Memolok Decision Ledger has admitted: the budgets, regulations, and other outside-the-decision truths its records rely on. Keeping these separate from the decisions themselves lets Memolok flag every record that rests on a fact that later turns out to be wrong.

A fact that is true regardless of any particular decision, such as a budget, a contract term, or a regulation. Decisions reference world facts from the ledger's world almanac rather than repeating them, so when a fact changes, every decision that relied on it can be found.

One of the five ways decision debt costs an organization. A decision that was made, with alternatives explicitly ruled out for explicit reasons, comes back years later and gets re-litigated from scratch by people with no access to the original reasoning. See decision debt.