Phoenix AI Core · Platform Deep Dive

One Governed AI Layer
Across the Systems You Already Run

Phoenix AI Core sits on top of SAP, ServiceNow, Salesforce, and Oracle — not in place of them. Four pillars, one control plane: model serving, cross-system automation, unified retrieval, and policy enforcement on every AI action. Below is how each works and the value it returns.

A Layer Above Your Systems of Record

Applications, agents, and copilots sit on top. The Phoenix AI Core sits in the middle as one control plane. SAP, ServiceNow, Salesforce, and Oracle stay exactly where they are — reached through governed connectors, never migrated. Phoenix is not a new system of record and does not replace any of them.

Business Applications AI Agents Embedded Copilots Analytics
Phoenix AI Core Inference · Agentic Workflows · Knowledge Graph · Governance & Trust
SAP · BTP / OData / RFC ServiceNow · REST / Integration Hub Salesforce · REST / Bulk / Platform Events Oracle · OIC / REST
SAP S/4HANA · ECC ServiceNow Instance Salesforce Org Oracle Fusion / EBS · Files

The Cross-System Layer
No Native Vendor AI Can Be

Almost no enterprise runs SAP, ServiceNow, or Salesforce in isolation. The processes that cost money to run badly — order-to-cash, case-to-resolution, procure-to-pay — span all of them. Every vendor's own AI stack is strong inside its walls and blind outside them.

The problem today

  • A ServiceNow-native agent can resolve a ticket but can't see the SAP invoice behind it.
  • A Salesforce Agentforce agent can draft a customer reply but has no governed way to check an on-prem SAP record before promising a ship date.
  • Every new AI initiative means another point integration, another model-serving bill, another set of guardrails from scratch.
  • Compliance faces three separate governance postures — one per vendor stack — with no unified audit trail of what an agent did across a full transaction.

What Phoenix does instead

  • Connects to each system through native connectors and leaves it as the system of record — no multi-year migration before the first workflow runs.
  • Runs one agent that plans and acts across SAP, ServiceNow, and Salesforce in a single governed workflow.
  • Resolves a customer's full context — CRM case, support ticket, ERP invoice — in one query instead of three.
  • Gives compliance one control plane, with every AI action traceable to the specific record it touched.

What makes it different

Cross-system by design, not by accident

SAP Joule, ServiceNow AI Control Tower, and Salesforce Agentforce each govern only their own platform. Phoenix's only reason to exist is being the layer that spans all of them — a position single-vendor stacks can't take without abandoning their own platform's gravity.

Wrap your stack, don’t replace it

Native connectors into SAP, ServiceNow, Salesforce, and Oracle. Each stays the system of record; Phoenix reads and acts through governed scopes. Value on day one, layered on systems that already work.

Governance that blocks, not just logs

Every agent action — SAP, ServiceNow, or Salesforce — is architecturally impossible without clearing policy first. Enforcement at the point of execution, applied uniformly, not a dashboard of what already went wrong.

Pricing you see before sales does

Workload- and connector-based pricing published up front — not a quote gated behind a multi-month enterprise sales cycle. Build a budget before the first meeting, not after.

No forward-deployed-engineer dependency

See a real cross-system workflow run live in a short demo. Production in weeks, operated by your own platform team — not a permanent embedded engineering squad on-site to keep it running.

Sovereign by default

Full in-region and air-gapped deployment for markets where hyperscaler dependency or cross-border data flow is a legal blocker — the reality for the multinational SAP / Oracle base this is built for.

Where this leaves the alternatives

  • Native vendor AI — SAP Joule, ServiceNow AI Control Tower, Salesforce Agentforce / Data 360 — is well funded and improving fast, but each governs only its own walled garden by construction.
  • Palantir AIP and C3.ai chase "unify everything" from the opposite direction — a new core that absorbs your data rather than a layer that reads from where it already lives.
  • Forward-deployed models lend you the vendor's engineers instead of teaching your team the system — a cost and dependency risk for mid-size buyers.
  • Phoenix's defensible ground is specific: organizations running two or three of SAP / ServiceNow / Salesforce / Oracle together that need one governed AI layer across all of them.

What that means for you

0
Forward-deployed engineers required to go live
Public
Workload- and connector-based pricing, no quote wall
Wrap
Systems of record stay exactly where they are
Blocks
Bad actions stopped at execution, not just logged
Component 01

AI Inference Engine

The serving layer that runs your models. It deploys, fine-tunes, and hosts large language models and classical ML behind a single API — so the same governed model access powers every downstream agent and copilot, regardless of which connected system triggered it.

How it works

Optimised model serving

A vLLM / TGI serving layer with batching, KV-cache reuse, and speculative decoding on GPU pools targets sub-100 ms first-token latency with predictable throughput under load.

Multi-provider router & failover

A policy layer routes each request to the cost-optimal model that meets the quality bar — Claude API alongside open models — with automatic failover across providers and regions.

Fine-tuning & adapters

LoRA / QLoRA and full fine-tuning pipelines adapt open and commercial models to your domain language, tone, and tasks without training from scratch.

Any deployment target

Run in Phoenix cloud, your VPC, or fully on-prem / air-gapped. The same API and tooling works everywhere, so data residency is a config choice, not a rebuild.

Autoscaling & cost caps

Traffic-aware scaling, per-team and per-connector token budgets, and hard spend limits keep inference cost bounded and attributable to the workload that spent it.

Observability built in

Every call is traced — latency, tokens, cost, and quality-eval metrics — through OpenTelemetry to your dashboards and alerting, tagged by connector.

Business value

  • One model bill, not one per initiative. Every agent and copilot across every connected system shares the same governed, cost-routed serving layer.
  • AI you can put in the critical path. Latency targets and automatic failover mean customer-facing and revenue processes can depend on it.
  • No model lock-in. Swap or add models as the market moves without touching application or agent code.
  • Data stays yours. VPC, on-prem, and air-gapped options satisfy legal, security, and residency review up front.
  • Faster time to first workflow. Teams build on a managed serving layer instead of standing up MLOps plumbing.

What you get

<100ms
First-token latency target
1 API
One governed entry point for every system
Multi-model
Cost-based routing plus automatic failover
Any deploy
Cloud, VPC, or on-prem / air-gapped
Component 02

Agentic Workflows

The orchestration layer that turns a model into a worker. Agents plan and execute multi-step processes across connected systems through a governed tool catalog — an SAP RFC call, a ServiceNow table update, and a Salesforce API action are all just scoped, audited tools — with human approval gates for high-value actions.

How it works

Plan & decompose

A stateful planner breaks a goal into ordered steps, chooses a tool for each, and adapts as results come back — no hard-coded flowchart per process.

Governed tool registry

Every read and write is a discrete registered tool with a schema, scope, and rate limit — never open-ended API access. An agent can only do what a specific tool explicitly allows.

Scoped, auditable credentials

Connectors run under least-privilege credentials — never a standing admin account into SAP, ServiceNow, or Salesforce — with every call attributable to an agent and a user.

Human-in-the-loop gates

Configurable approval thresholds pause the agent for sign-off on high-value or low-confidence actions, with the full cross-system context handed to the reviewer.

Durable & resumable

Long-running workflows persist state, survive restarts, and can be paused, replayed, and audited step by step.

Graceful degradation

If a connector is unavailable, the workflow serves last-known-good synced data with a staleness indicator rather than blocking the whole process.

Business value

  • Cross-system processes stop needing a human stitch. A billing dispute across a Salesforce case, a ServiceNow ticket, and an SAP invoice runs as one governed workflow.
  • Automate work RPA can't. Agents handle unstructured inputs, exceptions, and judgement calls that break rule-based automation.
  • Process owners build without IT tickets. Workflows are configured from the tool catalog instead of commissioned as custom integration code.
  • Consistency and auditability. Every decision is logged with its reasoning, its inputs, and the systems it touched.
  • People move up the value chain. Staff review approvals and exceptions instead of doing repetitive cross-system lookups.

What you get

Scoped
Discrete audited tools, no standing admin access
HITL
Approval gates on high-value actions
Durable
Pause, resume, replay, and audit any run
1 workflow
Spans SAP, ServiceNow, and Salesforce
Component 03

Knowledge Graph

The memory layer. It ingests and links records from SAP, ServiceNow, Salesforce, Oracle, and document stores into one permission-aware graph — so a query resolves a customer's full context, CRM case, support ticket, and ERP invoice, in one place instead of three.

How it works

Native connectors & ingestion

SAP (BTP, OData, RFC, IDoc), ServiceNow (REST / Table API, Integration Hub, MID Server for on-prem), Salesforce (REST & Bulk API 2.0, Platform Events), and Oracle (OIC) feed the graph on a configurable sync cadence.

Cross-system entity resolution

A Salesforce Account, a ServiceNow Company record, and an SAP Business Partner are resolved to one customer entity — so agents and queries reason about the real-world thing, not three copies of it.

Hybrid retrieval

A graph store (Neo4j-class) plus a vector index (pgvector / OpenSearch) combine graph traversal, semantic similarity, and keyword search — the basis for accurate, grounded answers.

Permission-aware context

Source-system access rights travel with the data. A user or agent only ever retrieves what it is entitled to see in the system the record came from.

Freshness & lineage

Incremental sync keeps the graph current; every fact carries its source system, timestamp, and confidence so answers cite back to the exact record.

One query API

Applications and agents ask questions through a single semantic API instead of integrating SAP, ServiceNow, and Salesforce separately for every use case.

Business value

  • Answers you can trust. Grounded, cited responses over live records cut hallucinations and the rework they cause.
  • Integrate each source once. Every future agent, copilot, and query reuses the connector instead of commissioning a new one.
  • No data migration. Records stay in SAP, ServiceNow, Salesforce, and Oracle; the graph reads and links, it doesn't absorb.
  • Compliance-safe by construction. Permission inheritance and full lineage keep AI within data-governance policy.
  • Faster new projects. New copilots launch against ready cross-system context instead of a fresh integration effort.

What you get

3 → 1
Systems touched to answer one question
Inherited
Source-system permissions travel with the data
Lineage
Every fact cited back to its source record
0
Data migration before the first query
Component 04

Governance & Trust

The control layer that wraps everything else. Identity, guardrails, immutable audit logging, and policy-as-code are enforced on every AI interaction across every connected system — giving compliance one control plane instead of three.

How it works

Identity & RBAC

SSO / SCIM integration (Okta, Entra ID) with OAuth2 + OIDC ties every request to a user or service identity, with role-scoped access to models, tools, and connectors.

Policy-as-code that blocks

Rules evaluate every action and resolve to block, require approval, or log — enforced at the point of execution, uniformly across SAP, ServiceNow, and Salesforce.

Guardrails & filters

Input and output screening for PII, secrets, prompt injection, and off-policy topics — configurable per application and per connector.

Immutable audit log

Every prompt, retrieval, tool call, and response is written append-only — never updated or deleted — attributable to SAP, ServiceNow, and Salesforce individually and to the overall transaction.

Evaluation & monitoring

Automated quality, bias, and drift evals run against golden datasets, with per-connector policy-decision tracking and alerts when a model regresses.

Compliance path

Field-level encryption, permission inheritance from source systems, and a SOC 2 Type II certification path built into the roadmap.

Business value

  • One governance posture, not three. Compliance reviews the same dashboard and audit log for every connected system instead of one per vendor stack.
  • Enforcement, not hindsight. A bad action is prevented at execution — the platform's core trust claim, tested exhaustively before GA.
  • Audit-ready at any moment. Pull the full trace of any transaction, with every AI action tied to the record it touched.
  • Central visibility and cost control. One pane shows every AI workload, its spend, and its policy posture.
  • Confidence to scale. New teams onboard under the same guardrails without a fresh governance project each time.

What you get

Blocks
Policy enforced at execution, not after the fact
1
Control plane instead of three vendor consoles
Append-only
Immutable audit log, per system and per transaction
SOC 2
Type II certification on the roadmap

One Request,
Every Layer, Seconds

"Resolve this customer's billing dispute" — a Salesforce case, a ServiceNow ticket, and an SAP invoice — handled as one governed workflow.

Governance authenticates and scopes the request

The user's identity, role, and data-access rights are resolved across every connected system before anything runs. Guardrails and policy for this application are attached.

The Knowledge Graph assembles context

The Salesforce case, the ServiceNow support ticket, and the SAP invoice and payment history are retrieved through hybrid search — filtered to what this user may see, each citation linked back to its source system.

An Agentic Workflow plans and acts

The agent checks the SAP invoice via the SAP connector, drafts a resolution, and validates it against refund policy — pausing for human approval if the amount exceeds a threshold.

The Inference Engine powers every reasoning step

Each planning, extraction, and drafting call is routed to the right model at the right cost, with latency guarantees and automatic failover.

Governance records the outcome

The full trace — context used, systems touched, decisions made, approvals — is written to the immutable audit log, attributable to SAP, ServiceNow, and Salesforce individually as well as to the overall transaction.

Connector design principles

  • Every connector runs under scoped, auditable credentials — never a standing admin account into a system of record.
  • Reads and writes are registered as discrete tools with schemas and rate limits, not open-ended API access.
  • Connector failure degrades gracefully — last-known synced data with a staleness indicator — rather than blocking the workflow.
  • Each customer is an isolated tenant; enterprises can bring their own API credentials and deploy Phoenix fully within their VPC or air-gapped.

Four Roles,
One Platform

Each works from a single Phoenix surface instead of jumping between vendor consoles.

IT / Platform Administrator

Connects SAP, ServiceNow, and Salesforce through the admin console, scopes credentials per connector, sets model-routing cost caps, and monitors connector health from one dashboard instead of three.

Business Process Owner

Builds an agentic workflow that spans systems — check SAP inventory, update a Salesforce opportunity, open a ServiceNow task for manual review — from the tool catalog, without custom integration code.

Compliance / Risk Officer

Reviews policy violations, blocked actions, and pending approvals across every connected system in one place, and pulls the immutable audit log for any transaction under review.

Front-Line Employee

Uses a Phoenix copilot embedded in the ServiceNow workspace or Salesforce Lightning console — it pulls context from all connected systems and takes governed actions without the employee knowing which backend it came from.

See a Cross-System Workflow
Run on Your Stack

Bring a real process that spans SAP, ServiceNow, or Salesforce. In a short demo we'll show inference, agents, knowledge, and governance handling it live — no slides.