Lore

Your business knows things AI doesn't.
Lore is where it remembers.

Every team carries knowledge that never makes it into a document — the vendor your predecessors tried and abandoned, the compliance edge case that burned you two years ago, the judgment your best estimator carries in her head. Lore captures it, organizes it, and makes it available to both the people and the AI tools that need it.

The pattern that matters: "When we tried to do X last year and it failed, here's why." Lore is the layer that holds that sentence — and makes sure neither your team nor your AI makes the same mistake twice.
See the full platform

Four kinds of memory.
One place to query them.

Institutional knowledge is scattered — in email threads, in meeting notes, in the heads of senior staff who may not be here next year. Lore pulls these into a structured, queryable substrate that grows more valuable the longer your team uses it.

Lore memory architecture: four capture types flowing into a shared queryable substrate DECISIONS Why we chose this TRACES What actually happened PRIOR ART Attempts that didn't land PATTERNS Signals across projects Knowledge Substrate QUERYABLE BY YOUR TEAM + YOUR AI WORKFLOWS GROWS MORE VALUABLE WITH EVERY PROJECT, EVERY QUARTER
Decisions

Why your team made the choices it made

When a decision is recorded in Lore, it carries the full context: the alternatives considered, the reasoning behind the choice, and the conditions that would change it. Future decisions — by people or by AI — start from what you already know, not from scratch.

Traces

What actually happened, not just what was planned

AI agent sessions, project timelines, workflow runs — the real record of how work unfolded. Lore indexes these against the decisions and documents they relate to, so you can always answer: "what did we do and why did we do it that way?"

Prior art

The attempts that didn't make it into the official record

Abandoned approaches, vendor evaluations, experiments that ran before you joined. This is the knowledge most likely to be repeated — the vendor your predecessor tried, the approach that failed in Q3. Lore surfaces it before your team wastes time rediscovering it.

Patterns

The recurring signals across projects and quarters

When Lore can see across multiple projects, it finds connections no single project owner would notice: the same cost overrun pattern appearing in different contexts, the compliance issue that shows up in a new form, the workflow shape that keeps working. These become durable guidance, not tribal knowledge.


Memory is the compound advantage.

Generic AI tools have no memory of your business. They start fresh every session. Lore changes that — for your people and for the AI working alongside them.

AI that knows your history

An AI answering a question about a vendor, a project approach, or a compliance edge case should know what your business already learned about that topic. With Lore as its context layer, it does. The answer it gives reflects your experience — not a generic training set.

"The model didn't know we'd already tried that vendor. Now it does."

People who can query what came before

A new operations manager shouldn't spend six months re-learning what the previous one already figured out. Lore makes institutional knowledge searchable — decisions, rationale, outcomes — so onboarding accelerates and senior knowledge doesn't walk out the door.

"Six months of context, available on day one."

Knowledge that compounds instead of decaying

Most institutional knowledge degrades: people leave, documents go stale, tribal knowledge disappears when the wrong person exits. Lore inverts this. Each engagement, each project, each decision adds to a substrate that becomes more useful over time — not less.

"Gets smarter every quarter. Not the other way around."


Context for the whole Foundry platform.

Lore is the knowledge substrate the rest of the platform draws from. It doesn't work in isolation — it feeds context into every layer that needs to know what your business already knows.

Forge

Context inside the cockpit

When your team is reviewing an AI workflow run in Forge, Lore surfaces related decisions and prior attempts so they're reviewing with full context — not just the action in front of them.

Temper

Evidence that accumulates over time

Temper measures what's working. Lore remembers why. Experiment results, quality trends, and evaluation outcomes land in Lore and become searchable — so the next time a similar question comes up, your team has evidence to draw from.

Your workflows

Informed answers instead of generic ones

Any AI workflow running on your Foundry deployment can query Lore mid-execution. The estimator workflow knows your historical bid data. The compliance workflow knows your prior exception history. The output reflects your business, not a generic model.


Talk to us about Lore.

Lore is part of every Foundry engagement. When we embed inside your business, we capture institutional knowledge from day one — decisions made, approaches tried, patterns observed. By month six, your AI and your people both draw from the same growing foundation.

See how an engagement works →