Praxis

Platform

Four engines. One platform.

Atlas and Fabric form a shared, trustworthy model of the business. Probe investigates on top of it. Forge acts on what Probe finds and measures the result. They are four jobs inside one system, not four products.

FabricConnections, definitions, evidenceAtlasThe living modelProbeInvestigation and priorityForgeApproved action, measuredActions and outcomes are written back, so the model learns what worked
  • FabricAtlas feeds trusted, linked records
  • AtlasProbe gives investigations a shared model
  • ProbeForge hands over ranked, evidenced findings
  • ForgeAtlas writes back actions and outcomes

The living model of the business

Atlas

What exists, what happened and how it is linked: customers, jobs, deals, invoices, people, events and the relationships between them.

Things, links and events

A customer requested a job, a technician was assigned, it was billed on an invoice, a payment settled it. The model reads like a precise description of how your company works.

Sequences across tools

Questions like “customers whose job ran late and then did not pay” need order and relationships. Atlas holds both.

Grows with every project

Kinds of things and starting definitions are reusable. Each project configures names, sources and rules instead of rebuilding the model.

Example · One customer, six systems, six IDs. In Atlas she is one person with one history.

Connections, definitions, evidence and identity

Fabric

Fabric keeps the model honest. It connects sources, keeps every record's origin, and never treats a guess as a fact.

Identity, confirmed by people

Praxis suggests same-person links across sources, scored by match strength. A person confirms or rejects each one. Records are never merged on a guess.

Conflicts kept, unknowns recorded

When sources disagree, both values are kept with their source, and a person resolves the conflict. Unknown is never counted as no.

One versioned definition per term

“Active customer” or “late job” has one approved definition. Definitions are versioned, and changing one flags every result that used the old version for review.

Example · Billing says 14 Oak St, the CRM says 12 Oak St. Both are kept with their source, and a person resolves the conflict.

Investigation, discovery and economic priority

Probe

Probe asks why. It walks the model across systems, tests competing explanations and estimates what each finding is worth.

Several explanations, each with a check

A drop in bookings could be missed follow-ups, lower demand, or calls made and never logged. Probe tests each against the data and marks it supported, weakened, inconclusive or not testable.

Statistics and sample records shown

Every verdict shows its statistics and the records behind it. Findings are presented as associations until a measured change confirms them, so people can check before they act.

Value at stake, with its assumptions

Probe estimates the value at stake for each finding as a range, with the assumptions stated. An estimate is never presented as money earned.

Example · A pattern is a lead, not a verdict. Probe shows its reasoning and how sure it is.

Approved execution and measurement

Forge

Forge turns a finding into work that gets done: work orders for people, steps for software, all within your rules and all measured.

Approval bound to the exact plan

A record-by-record preview, then a second person's approval tied to the exact change set. Change anything and the approval no longer applies.

Two lanes of work

Work that needs a person becomes a work order with an owner, a due date, instructions and evidence of completion. Software steps run behind business rules that can block them. The write path into your systems is configured in the engagement.

Done means checked

Changes are verified by readback after they run. Results are measured against a fixed holdout group, assigned by hash and stable per record, plus before and after.

Example · Standing operations run on a schedule, in dry run or live, with a global kill switch, per-channel switches and a manual pause.

How the engines work together

  • FabricAtlas feeds trusted, linked records
  • AtlasProbe gives investigations a shared model
  • ProbeForge hands over ranked, evidenced findings
  • ForgeAtlas writes back actions and outcomes

Software at the core. AI when useful.

Praxis connects and understands your business through structured data, evidence, and software logic. AI can assist with parts of an investigation when it adds value, but the platform doesn't depend on it to operate.

AI output is validated before it is used. A number that is not in your data is rejected.

Start with one problem.

The first workflow builds the model. Every one after it reuses what is already there.