Synthetic Intelligence Layer
and AI-native Platformization

Your firm’s knowledge gathered into governed memories and put to work by department agents, or one secure platform built from your prototype or requirements. Both are built and run inside your environment.

REVIEW FEEDS BACK INTO MEMORYRECORDSSeveral systemsDOCUMENTSSeveral placesJUDGMENTHeads and inboxesFIXED-FEE DISCOVERYRecommends the routeROUTE 1SyntheticIntelligence LayerDepartment by departmentROUTE 2AI NativePlatformizationOne service line or clientREVIEWERApproves every outputONE FOUNDATIONGoverned dataFull audit trailYour AI contractsYour environment
Who we work with
BCG
EY
Fidelity Investments
HCL
Reliance
Maruti Suzuki
Adani
TATA
Mitsubishi Motors
ICICI Bank
EXL
Oracle
Wipro
Samsung
The World Bank
Schneider Downs
Moore Colson
United Nations
HDFC Bank
UNDP
JPMorgan Chase & Co.
State Bank of India
IDBI Bank
Benetton
Care Health Insurance
LIC
IFFCO-Tokio
Axis Bank

Partner Judgment is your Least-protected Asset

How a client likes to work, why a judgment was made, which precedent applies: the systems hold the numbers and the files, while the judgment sits in heads, inboxes and local drives. Most of what makes a partner valuable is held outside the firm’s systems, and three pressures put it at risk.

Succession

Client knowledge and technical calls leave with the partner. The firm keeps the files and loses the reasoning, and clients notice first.

At RiskClient Continuity

Rework

Records in several systems, documents in several places, and a person re-keying between systems that look integrated.

At RiskMargin & Staff Time

Ungoverned AI

AI tools on client work before rules exist on where data may go, which models may be used, or who approves the output.

At RiskData Handling & Liability
~95%

of enterprise AI pilots show no measurable P&L return. The ones that work start narrow, fit the workflow and learn from the firm’s own material.

~22%

of organisations have a visible AI strategy. Those that do are twice as likely to see AI-driven revenue growth.

Sources: MIT NANDA, The GenAI Divide: State of AI in Business 2025; Thomson Reuters, Future of Professionals, June 2025.

Two Routes, One Foundation

Route 1|Synthetic Intelligence Layer
Data Governance

Four Phases

  1. 0Discovery & Assessment
  2. 1Federated Search Layer
  3. 2Governance of New Data
  4. 3Data Warehousing
Memory

Two Memories

  • Firm MemoryPolicies, templates and precedent
  • Leader MemoryEach partner’s second brain
Models

Your Systems and AI Models

ClaudeMicrosoft CopilotAzureMicrosoft 365Microsoft TeamsMicrosoft Excel
Agents

Department by Department

TaxAuditAdvisoryHRFinanceIT& others

Approval Gate
for Reviewer

Review feeds back into memory
Route 2|AI Native Platformization
Sub-Route AYour Prototype
You ≈25%SPNX ≈75%
Sub-Route BYour BRD
SPNX 100%
1Requirements
2Wireframe
3Architecture
4Secure Build
5Test & QA
6Security & Audit
7CI/CD Deploy
8Run & Maintain
Ownership

100% transfer with code. Your team takes over running and maintaining it.

Start with Data Governance

Your data is found, governed and consolidated in four phases, before the two memories and the department agents are built on it.

PHASE 0

Discovery & Assessment

Weeks 1–4

  • Data-source Inventory
  • Ticket-dump Analysis
  • Stakeholder Interviews
  • Roadmap & Pricing Lock
PHASE 1

Federated Search Layer

Weeks 5–16

  • Search Architecture
  • Connector Build
  • Tax-practice Pilot
  • Firm-wide Rollout
PHASE 2

Governance of New Data

Weeks 14–24

  • Classification & Retention
  • Filing Standards
  • Merger Onboarding Playbook
  • Access Rationalization
PHASE 3

Data Warehousing

Months 6–12

  • Data Warehouse Foundation
  • Pipelines
  • Migration & Archive
  • AI-readiness Layer

Organizational Hierarchical Memory

One governed layer across the firm, holding two kinds of memory and putting them to work through department agents.

Firm Memory and AgentsThe Blueprint
What It Is
One firm-wide layer: your policies, methods and precedent in Firm Memory, with agents for tax, audit, advisory and HR.
How It Works
A governed data layer is built in your environment. Each agent passes a quality gate before release.
What Changes
Work starts from your own precedent, and every AI-assisted output is traceable and approved by a person.
Leader MemoryEach Partner’s Second Brain
Vault
What It Is
A private, linked record of each partner’s notes, decisions, meetings and reference material, held in your environment.
How It Works
Each partner chooses the sources. SPNX links them, applies firm governance and keeps versioned copies.
What Changes
A partner’s judgment stays searchable and stays with the firm when they retire.

AI Native Platformization

Start where the need is sharpest, such as a tax workflow or a client portal. Each platform is built on the shared foundation, so it joins the firm-wide layer without a rebuild.

Route A : Hand off your Prototype

The 25% vs 75% Model
You Bring
A prototype, mock-up or working demo built with an AI tool.
We Build
Architecture (L3) to Run & Maintain (L8).
You Own
The platform, its source code and its documentation.
Fits When
You already have a proof-of-concept or an internal product owner.

Route B : Have SPNX Build It All

The 100% Model
You Bring
A business requirements document.
We Build
Requirements (L0) to Run & Maintain (L8).
You Own
The platform, its source code and its documentation.
Fits When
You want a single accountable builder from first screen to live.

Stages of Development

L0L1L2L3L4L5L6L7L8RequirementsWireframeArchitectureSecure BuildTest & QASecurity & AuditCI/CD DeployRun & Maintain

Client Data Stays in Your Environment

Whichever route you take, SPNX works on your IT team’s terms. Both routes run on the same controls.

YOUR ENVIRONMENTIF OUTSIDE USROUTE 1Synthetic IntelligenceLayerROUTE 2AI NativePlatformizationGATEYour IT approvesevery accessHuman control pointCONDITIONOffshore accessconsentCLIENT DATAClient records, documentsFirm & Leader MemoryStays where it livesAI MODELAccounts in your firm’s nameYour data terms governSwitch provider any timeAGENTSAgent outputsPlatform releasesInside your systemsREVIEW GATEReviewer sign-offon every outputand releaseFeeds backinto MemoryLOGGEDAUDIT TRAILEvery action and loginrecorded, readableby your firm

Roadmap and Commercial Plan

You commit only to a fixed-fee discovery. Everything else comes later. Discovery maps your systems, data and owners, recommends the route and returns a plan with a fee for each stage.

Month 0Month 6Month 12Month 18Month 24
Route 1Synthetic Intelligence Layer
Agents, department by department (whole firm in 12–24 months)
Route 2AI Native Platformization
First platform
Next platforms, each joining the firm-wide layer

Indicative timing. Discovery confirms the plan for your firm.

Discovery
Fixed
Maps systems, data and owners.

Recommends the route and returns a plan with a fee for each stage, so you decide the rest with full information.

Build
One Time
A one-time fee

Sized to the scope you approve after discovery. Building begins only once you have approved that scope.

Operate
Usage + Service
Usage actually consumed

The models and infrastructure on your accounts, plus a service fee that steps down as your team takes over.

Transfer
Fixed
Your team takes over running it.

Fixed at contracting, against agreed competency gates, on a date set before the engagement begins.

Put Your Firm’s Knowledge to Work, Securely.