Services

The graph behind the agents

Knowledge Graphs

Palmetto builds the knowledge graph that gives your agents and your teams the same context. It connects fragmented data into one shared model and tracks the entities, rules, and provenance a decision depends on.

Built for Organizations where systems disagree on meaning and teams reconcile the same facts by hand.

Scattered facts.
One map.

Your agent can follow the path to the answer, and you can check every step.

Claims systemPolicy adminShared drivePolicy HO-3Rider R-17Flood exclusion 4.2Answer, with sourcesClaims systemPolicy adminShared drivePolicy HO-3Rider R-17Flood exclusion 4.2Answer, with sources
The question
Which exclusions apply to this claim?
The answer
Flood exclusion 4.2 applies. Rider R-17 does not override it.
Where it came from
Policy HO-3 §4.2, Rider R-17, claim note 3

01 · When to call

Where this shows up.

  1. Context for agents and search

    Agents get real relationships, definitions, and sources, on top of document search.

  2. Connected customer, product, and risk views

    Match the same entity across systems, and keep track of where each fact came from.

  3. Policy, control, and evidence traceability

    Connect obligations, controls, owners, and evidence, so the gaps are easy to see.

  4. Data and integration semantics

    One set of business definitions for your APIs, analytics, migrations, and integrations.

02 · What you get

What you get.
What you keep.

  1. The decisions and questions the graph has to answer
  2. Domain ontology, with terminology and documentation
  3. Graph architecture, source mappings, identity, and provenance
  4. A working ingestion, query, or GraphRAG pattern on representative data
  5. Governance, quality checks, and an adoption roadmap

The ontology, mappings, quality rules, and governance playbook are documented and handed to your team, platform choices included.

Discovery begins with the decisions and questions that matter, then expands the model and platform around those uses. How engagements run →

coverage.ttl
:Policy       a owl:Class .
:Coverage     a owl:Class .
:Exclusion    a owl:Class ;
              rdfs:subClassOf :PolicyTerm .

:hasCoverage  a owl:ObjectProperty ;
              rdfs:domain :Policy ;
              rdfs:range  :Coverage .
:excludes     a owl:ObjectProperty ;
              rdfs:domain :Coverage ;
              rdfs:range  :Exclusion .
:source       a owl:AnnotationProperty .  # provenance on every assertion
Classes
Policy, Coverage, Exclusion
Relations
hasCoverage, excludes
Provenance
A source on every assertion

We build and run our own product

We built membl.io.

Your GitHub Actions jobs, on the runner that fits.

membl runs each job on its own managed runners or on compute you already have. It picks the runner, says why, and shows what the job cost.

We designed it, built it and run it: the site, the dashboard, billing and docs. Our own client portal runs its CI on it.

Visit membl.io

03 · Good questions

Before we begin.

What is the difference between a knowledge graph and an ontology?

An ontology defines the concepts, relationships, and rules of a domain. A knowledge graph uses that meaning to connect real entities and facts. They usually work best together.

Do we need a new graph platform before we begin?

Not necessarily. Early work tests the model against representative data before any platform decision. Existing databases and tools get evaluated before we recommend anything new.

How do you keep it under control in production?

  • Provenance, confidence, effective dates, and lineage where you need them
  • Access controls that match your source systems
  • Versioned ontology and mappings, with reviewable changes
  • Quality checks and a clear path for facts that contradict each other

How will we know it's working?

  • Coverage and accuracy against the priority business questions
  • Entity match, duplicate, reconciliation, and exception rates
  • Time required to find, explain, and validate an answer
  • Reuse across workflows, integrations, analytics, or AI applications

Who do you need from our side?

  • A business owner who can prioritize the decisions and users the model must support
  • Domain experts empowered to resolve terminology and relationship questions
  • Data owners who understand source quality, identifiers, access, and lineage
  • Application, platform, governance, and security participants matched to the intended use

Next step

Bring the context.
We'll define the next step.

Tell us the workflow or decision, who owns it, and what it costs today. A rough version is fine.

Discuss what your agents need to know →

Or email support@palmettointeractive.com.