enterprise governed-AI
Download PDF
SOLEUM
STEWARD

Govern how AI operates inside your institution.

Investor Deck · Q3 2026 · Confidential
The problem

Everyone is adopting AI.
Few institutions can explain, verify, or defend how it is being used.

Boards, regulators, funders, and auditors now ask:
What information did the AI use?
Who reviewed or approved it?
Where did the information come from?
Can we verify the source?
Can we defend the decision?
Are we exposing sensitive data, operational risk, or future liability?
Q3 2026 · 02
Why now

AI oversight has become a boardroom
and legal requirement.

AI adoption
Enterprise GenAI spend reached $37B in 2025, growing more than 3x year over year.
Board accountability
Boards and executives are now directly responsible for how AI is used inside the organization.
Regulations
New AI rules are adding penalties, disclosures, and human oversight requirements.
Shadow AI
Employees are already using AI tools without clear oversight, approval, or records.
Data risk
Sensitive data and internal knowledge need protection.
Oversight failures
95% of enterprise AI pilots show no P&L impact, after $30–40B invested.
The need for
AI oversight infrastructure
Q3 2026 · 03
The Solution

STEWARD turns company knowledge into governed AI answers.

AI spend is rising, but ROI and risk remain unclear. STEWARD helps institutions use AI safely, reliably, and affordably without replacing existing systems.

APPROVED COMPANY INFORMATIONDOCUMENTSDRIVESSYSTEMSCODEPOLICIESREPORTSSTEWARDAI Engine·  Searches approved information·  Verifies sources·  Generates source-backed answers·  Logs every AI interactionAccessed through workspace, Slack,or embedded widget.USER-FACING OUTPUTSACTIONABLE ANSWERS FOR EMPLOYEESSOURCES LEADERS CAN REVIEWRECORDS THE INSTITUTION CAN EXPORTDASHBOARDS, REPORTS, AND INSIGHTSPUBLIC WEBSITE / CUSTOMER TOUCHPOINT
Approved company information
DOCUMENTSDRIVESSYSTEMSCODEPOLICIESREPORTS
STEWARD
AI Engine
Searches approved information
Verifies sources
Generates source-backed answers
Logs every AI interaction
Accessed through workspace, Slack, or embedded widget.
User-facing outputs
ACTIONABLE ANSWERS FOR EMPLOYEESSOURCES LEADERS CAN REVIEWRECORDS THE INSTITUTION CAN EXPORTDASHBOARDS, REPORTS, AND INSIGHTSPUBLIC WEBSITE / CUSTOMER TOUCHPOINT
Q3 2026 · 04
360-degree value proposition

Use more AI while carrying less unmanaged risk.

As the company's AI activity moves through our platform, the institution builds stronger records, clearer oversight, and more reviewable AI usage over time.

De-risking
Reduces:
Data exposure
Unsourced answers
Missing audit records
Indefensible decisions
Ungoverned AI usage
Untested high-stakes decisions
Leveraging
Increases:
+Faster answers
+Institutional memory
+Confidence in AI answers
+Audit and compliance readiness
+Verified outcomes
+Better-tested decisions
De-risking
Data exposure
Unsourced answers
Missing audit records
Indefensible decisions
Ungoverned AI usage
Untested high-stakes decisions
Leveraging
+Faster answers
+Institutional memory
+Confidence in AI answers
+Audit and compliance readiness
+Verified outcomes
+Better-tested decisions
Q3 2026 · 05
The security model

No master key. No new attack surface.

STEWARD uses the permissions already in place. It only sees what each user is allowed to see, without creating a master account or copying the data.

The common pattern
STEWARD
Copies data into a vendor’s shared cloud
Runs in a private, isolated environment per customer
Maintains a stale copy of the data
Reads the live system at answer time
Builds a second permission map
Inherits the permissions already in force
Runs on a master service account
Holds no master key
Adds a new attack surface
Adds none
Stalls in security review for months
Maps onto the posture already approved

We have found no competitor offering this model. The architecture is hard for incumbents to retrofit. This is the white space.

Q3 2026 · 06
Real traction

STEWARD has clear use cases, active pilots, and commercial opportunities.

STEWARD + Supply Chain
Active pilot
Hope and Healing
HOPE & HEALING

Tracks medical and aid shipments across 40 countries so every delivery can be verified and reported to donors.

STEWARD + Supply Chain
Q3 pilot
MAP International
MAP INTERNATIONAL

Creates one verified record of where supplies go, with AI insights in the workflow.

STEWARD + Non-Profit
Design partner
TeamImpact
TEAM IMPACT

Connects young patients with athletic teams and records whether the match happened and helped.

STEWARD + Manufacturing
Q3 pilot
Continental Diamond Tool
CONTINENTAL DIAMOND TOOL

Uses approved company knowledge to answer internal questions with sources and a permanent AI interaction record.

STEWARD powers each current use case. The same engine can be pointed at new teams, new data, and new markets without rebuilding the product.

SOLEUM
Q3 2026 · 07
How It Expands

How STEWARD expands inside an organization.

Begin with STEWARD
  • Start with one use case
  • Use existing company information
  • Deploy with low disruption
Platform expansion

STEWARD creates the entry point for other use cases, departments, executive decision-making, and broader AI leverage.

Usage expansion
  • Add more teams and knowledge bases
  • Keep every interaction on the record
STEWARD · Why it wins
·Live in days, not months; multi-model underneath, no lock-in
·Teams get answers in seconds, in plain language, no technical skills needed
·Every answer drawn only from approved company information, with sources shown
·Builds dashboards, artifacts, and visual insights on your data, not just text
·Ends repeated manual work and builds institutional memory with every use
·Leaders see how AI is used, usage and risk in one place, logs exportable
·Permanent records keep you audit-ready and able to defend decisions
·People stay in control, human review built into the workflow
·Sensitive data stays protected inside your walls
01
Begin with STEWARD
  • Start with one use case
  • Use existing company information
  • Deploy with low disruption
02
Usage expansion
  • Add more teams and knowledge bases
  • Keep every interaction on the record
03
Platform expansion
STEWARD creates the entry point for other use cases, departments, executive decision-making, and broader AI leverage.
SOLEUM
Q3 2026 · 08
Business expansion

One governed engine. One market at a time.

The engine never changes. Each new industry applies the same governed STEWARD engine to a new persona and knowledge base. A configuration, not a rebuild, entered in sequence as each market is won. Expansion carries near-zero marginal product cost.

01
Banking & wealth

Product terms, eligibility, rates, and procedure at the advisor’s desk. A persona per product line. Every answer cited for the compliance record.

02
Insurance

Underwriting rules and claims policy, answered in seconds with the governing clause attached. Consistent decisions across every desk.

03
Retail & D2C

Product catalogs, brand voice, and loyalty knowledge served at every customer touchpoint. One persona per brand across clothing, consumer electronics, and luxury, protecting legacy equity and driving long-term engagement.

04
Government

Internal policy and citizen-service knowledge for caseworkers. Every answer on the record, ready for oversight.

05
Asset & fund management

Mandates, restrictions, and reporting rules per fund. STEWARD answers within each manager’s clearance.

06
Legal & professional services

Matter knowledge and precedent inside the firm’s own walls. Nothing leaves. Everything is cited.

07
Manufacturing & energy

Standard procedures, safety rules, and maintenance documentation at the point of work, access-bound.

08
Healthcare administration

Coverage rules, protocols, and internal policy for staff. Administrative knowledge, not clinical judgment, bound to each role’s access.

ONE ENGINE
NEW PERSONA
NEW KNOWLEDGE BASE
NEW VERTICAL
Q3 2026 · 09
Business model

Measured cost. Metered proof. Software margins.

<4¢
Cost per query

Measured blended cost: telemetry on 100% of conversations.

74–75%
Gross margin

Measured from live telemetry, not estimated. Rising across the plan horizon.

$0
Per-seat fees

Fixed annual base + usage meter. A predictable bill, never per seat.

Value vs price

Recovered working time alone returns ~5× the subscription.

The value math, stated honestly

A 1,000-employee institution: 300 daily users saving 30 minutes a week recover ~7,000 hours a year: roughly $385K in time against a $50K subscription, before a single avoided compliance event.

Projected revenue
2026$1.0M
2027$12.0M
2028$38.0M
2029$85.1M
2030$160.1M
Net of discounts. ~$204M ending ARR by 2030; full model in the data room.

STEWARD owns the reasoning loop and rents the models. When model prices fall, the margin widens and no contract changes.

Q3 2026 · 10
Market opportunity

Enterprise governed-AI work is a $225B market by 2032.

TAM
$225B
Convergent enterprise governed-AI work by 2032 (de-duplicated, not a sum).
SAM
$20B
US firms 250+ employees with governance & knowledge-work AI budgets (serviceable category).
SOM
$204M
5-year ending ARR: ~2,700 accounts × ~$75K blended ACV.

Precedent: Glean, $300M ARR in ~4 years ($7.2B valuation). Sierra, $100M ARR in 7 quarters ($10B).

$0B$30B$60B$90B$120B2026202720282029203020312032
AI assistants & copilots$15B $112B
Enterprise agentic AI$10B $93B
AI knowledge management & search$10B $50B
Convergent category (assistants/copilots, agentic AI, and AI-powered KM/search); overlapping, not summed. Sources: MarketsandMarkets, Grand View, Research and Markets (2025–26).
Q3 2026 · 11
Fragmented competitive landscape

Institutions are being forced to choose between speed
and control. STEWARD gives them both.

STEWARD
Enterprise governed-AI
Frictionless installSourced answersInherited permissionsReviewable recordsPredictable pricingNo master keyMulti-modelHuman oversight
Tools that make AI defensible but not useful at the point of work.
IBM watsonx.governance
Credo AI
OneTrust
Holistic AI
ServiceNow
NeMo Guardrails
Lakera
Arize
Fiddler
Arthur
WhyLabs
Galileo
Patronus
ModelOp
Monitaur
Collibra
Truyo
Trustible
Saidot
FairNow
Cranium
HiddenLayer
CalypsoAI
Prompt Security
Governance and Control
Tools that make AI useful but do not govern it, and where usage-based cost climbs fastest.
Glean
Microsoft Copilot
Hebbia
Guru
watsonx Orchestrate
LangChain
LlamaIndex
CrewAI
AutoGen
Semantic Kernel
OpenAI Agents SDK
Google ADK
Palantir AIP
Cohere North
Writer
Credal.ai
Amazon Q
Gemini Enterprise
Dust
Perplexity Enterprise
Sana
Moveworks
Aisera
Salesforce Agentforce
Sierra
Decagon
Cresta
Kore.ai
Assistants and Applied AI
Auth0 / Okta
Microsoft Entra
Ping
CyberArk
SailPoint
Saviynt
OpenAI
Anthropic
Google
Meta
Mistral
Cohere
xAI
AWS Bedrock
Azure
Identity, models, and infrastructure
Q3 2026 · 12
The team

Built by operators with enterprise AI, product, design, compliance, and mission-sector experience.

Leadership

Donny Vanker
Donny Vanker
CEO & Co-Founder

+20 years building enterprise businesses in sports performance, health care & non-profits. Systems thinker who architects institutional infrastructure.

Pablo Stefanini
Pablo Stefanini
Chief Innovation Officer & Co-Founder

17+ years driving digital transformation at Google, Globant (NYSE: GLOB) and Monks (S4 Capital Group). Senior leadership across enterprise services and global markets.

Michael Maccia
Michael Maccia
Outside General Counsel & Business Development

Advisor to sports, entertainment, and tech companies; 16+ years in privacy, security, compliance, contracts, and litigation. SSP Blue, nPLAY Foundation, Templar Titan.

Strategic Advisors

Cristina Paslar
Cristina Paslar
CMO, IBM Regional Markets

MBA from Yale. 20+ years marketing innovation and AI-driven campaigns.

Brian VanHall
Brian VanHall
CEO of CDT

20+ years executive growth. FACHE Fellow, MBA/MHA Baylor.

Bruce Bruinsma
Bruce Bruinsma2× Exit
Investments & Expansion

40+ years in finance. SOLEUM™ investment strategy.

Adrian Breitburd
Adrian Breitburd
Architecture & Technology Advisor

20+ years at Globant, Santander Openbank (top EU digital bank), Google/Meta.

Technical & Operational Orchestration

Anas Karkoukli
Anas Karkoukli
Head of AI Engineering

Builds secure, scalable, mission-critical systems across the SOLEUM™ ecosystem.

Jeshua Nanthakumar
Jeshua Nanthakumar
Head of Experience Design

Shapes the product vision, experience, and design standards.

Damian Burne
Damian Burne
Head of Operations

Optimizes SOLEUM™ digital efficiency / ROI. Leads agile transformations and workflow automation; embeds ownership.

Julian Garibotti
Julian Garibotti
Financial Analyst & Controller

Financial operations, budgeting, and compliance. Background from Huawei; Universidad de San Andrés.

Key Hires in Progress

Head of Growth & Influence
In 2nd-stage interviews

7 shortlisted candidates include SVP-level operators from S4 Capital, Accenture, and Venrock-backed SaaS. Expected close Q3 2026.

Head of Products
Target Q3 hire

Responsible for all product roadmaps.

Q3 2026 · 13
Milestones & use of proceeds

What this capital unlocks over the next 12 months.

Phase 01LaunchSTEWARD goes to market.
Phase 02ConvertConvert pilots into customers, publish outcomes and case studies, and build enterprise pipeline.
Phase 03ExpandThe same engine expands into the next vertical once the first market is won.
Q3 2026
STEWARD launch

First paid STEWARD customers. Harden the solution.

The enterprise wedge goes live.

Funds the platform
Q3 2026
Go-to-Market

STEWARD use cases already live in the field.

Product and engineering hires.

Already in the field
Q4 2026
First conversions & case studies

Pilots convert to paid contracts. Published outcomes build the enterprise pipeline.

Enterprise sales, strategic partnerships, customer success.

Evidence compounds
Q1 2027
Enterprise security readiness

SOC 2 Type II preparation for enterprise customers.

Investor reporting infrastructure.

Compliance readiness
Q2 2027
Series A readiness

The round is earned on customer evidence.

Target: Q2 2027.

The next gate
Q3 2026 · 14
The raise

This is a disciplined commercialization round.SOLEUM is turning built technology into customer evidence, revenue, compliance readiness, and Series A readiness.

$5M
SAFE
$30M
Pre-money valuation cap
20%
Discount
18 mo
Runway from close

Data room available under NDA. Five-year financial model, competitive benchmark, pilot documentation, legal and IP records.

SOLEUM
Q3 2026 · 15

The next era of AI belongs to institutions that can govern what they use.

SOLEUM gives institutions the infrastructure to see, govern, verify, and defend AI-driven activity.

Donny Vanker
CEO & Co-Founder · donny@soleum.io
Pablo Stefanini
CIO & Co-Founder · pablo@soleum.io
SOLEUM
soleum.io