Re. D / 04 · Owned IP
We rewrote our deal-qualification rules in five different vendor consoles. None of them are ours. Our “sales methodology” is now somebody else’s prompt template.
———————/ VP Revenue
B2B SaaS / €40M ARR
B2B SaaS / €40M ARR
01 · ThesisCompany
Company
Operating Systems
for the AI era.
Spark builds owned operating layers where company intelligence sits at the center, and humans, tools, and agents operate around it.
Tier I/€5,000 fixed/1–2 weeks/Spark Map delivered
Ring · 03 / HumansRing · 02 / AgentsRing · 01 / WorkflowsCore · Memory · Goals · Rules↻ Rotation 0.04 rad/s · stable
Built for
Legal, finance, operations & professional-services teams who want to own their intelligence layer — not rent it from SaaS vendors.
Engagement
Start with the €5,000 Foundation Track. Full implementation runs 8–12 weeks at €25k–€40k, scoped.
Focus
Autopilot & self-driving systems. Approvals, evals, memory, schemas — owned end-to-end.
What you get
An operating layer your team can extend, inspect, govern, and run for the next decade.
02 · Diagnosis
Symptoms · n = 7 · observed in the field
D / 07 = forward-looking
Three years in. Why hasn’t AI compounded?
Six symptoms we hear from operators every week, plus one that’s about to land. The headline’s the same everywhere — spend is up, the org feels busier, none of it compounds. These are the load-bearing reasons why.
SymptomConditionSignal
D / 01
AI pilots that never reach production.
Every quarter ships a new prototype. Almost none of it ends up running the business. The half-life of “promising demo” is a board cycle, then it’s replaced by the next promising demo.
POC → PROD< 8% conversion
D / 02
Context that resets every Monday.
Each conversation starts from zero. The model knows your business about as well on day 400 as on day 1, because nothing about your operation is persisted in a place you control.
SYSTEM MEMORY0 days retained
D / 03
Tool sprawl masquerading as transformation.
Every vendor has shipped an AI feature. You’ve bought most of them. Your team now bounces between forty interfaces that each know one-fortieth of the company — and none of them know each other.
STACK SURFACE47 vendors / 1 strategy
D / 04
Vendor prompts own your business logic.
The actual decision rules — how your company qualifies a lead, scopes a deal, escalates a ticket — increasingly live inside a SaaS provider’s prompt template. You don’t see it. You don’t version it. You can’t leave with it.
OWNED IP0% of decision logic
D / 05
Agents you can’t safely give real access.
The interesting work needs write-access to real systems. Without governed scopes, audit, and approvals, every demo dies the moment legal asks what happens if it goes wrong. So agents stay read-only — and useless.
SCOPEread-only, indefinitely
D / 06
No telemetry. Only anecdotes.
Is it working? Nobody knows. There are no production evals, no drift detection, no per-decision audit. ROI is whatever the loudest stakeholder said at the last all-hands.
EVALS IN PROD0 / unmeasured
D / 07
Every team is about to ship its own agent.
Tool sprawl was the warm-up. The next twelve months will multiply it — every department running its own agents, with its own rules, memory, and decision logic. No shared schemas. No shared evals. The fragmentation you have today, replicated and accelerated — except this time the actors run themselves. The structure of your business gets reshaped by whoever shipped fastest, not by what should be true.
AGENT SPRAWL · INCOMING+1 per team / 0 governance
03 · From the field
Intake transcripts · n = 200+ operators
This isn’t theoretical. It’s what operators tell us, weekly.
We’ve shipped twelve copilots in eighteen months. None of them know each other exists. Every one of them learned our company from scratch — and then forgot.
————————/Head of Operations/Professional services/~120 FTE
Re. D / 05 · Scope
Every agent demo dies the moment legal asks what happens if it goes wrong. So our agents stay read-only, and nothing real ever ships.
————————/ CTO
Fintech / ~80 FTE
Fintech / ~80 FTE
Re. D / 06 · Evals
I can’t tell you if any of our AI initiatives are working. I can tell you what the last all-hands deck said. Those are different things.
———————/ COO
Legal tech / ~250 FTE
Legal tech / ~250 FTE
04 · The inversion
Lat. 40.7128 · Lon. −74.0060
For thirty years, tools were at the center. Now intelligence is.
The current model · "Humans around the edge"
SaaS tools at the center.
Your business logic is scattered across 50+ vendors. Humans copy state between them. Every tool is an island with its own permissions, schema, and politics.
Tools (logic)Humans (couriers)
The Spark model · "Intelligence at the center"
Company intelligence at the core. Tools as instruments.
Memory, goals, decision rules, and schemas live in one owned layer. Agents and workflows orbit it. Humans and tools sit on the outer ring — invoked, not central.
Intelligence (core)Agents / workflowsHumans / tools
05 · Primitives
7 components · owned, inspectable, executable
What Spark builds into your company.
P / 01
Memory
A durable, structured record of what your company knows, decides, and remembers. Not a chat log — a system of record for intelligence.
P / 02
Workflows
Explicit, executable representations of how work moves through your business. Versioned. Forkable. Inspectable end-to-end.
P / 03
Agent instructions
The system prompts, tool grants, and policy boundaries that define how each agent behaves on your behalf. Composable. Reviewable.
P / 04
Approvals
Human-in-the-loop gates with clear scope, signers, and timeouts. The governed surface where machines defer to people.
P / 05
Schemas
Typed, validated structures for every business object — customers, contracts, runs, decisions. Truth lives here.
P / 06
Tool connectors
Thin, owned adapters into the SaaS you keep. Tools become invokable instruments — never the source of truth.
P / 07
Evaluations
Continuous measurement of every agent, workflow, and decision. Production telemetry for your operating layer.
P / 08
Governance
Permissions, audit, drift detection. The seams between your company's intelligence and the rest of the world.
06 · Doctrine
More tools.
Use fewer. Make them more agentic.
Why less
Every extra SaaS layer adds intermediation, hidden logic, maintenance burden, permission complexity, and context drift. Each is a place where your company's intelligence leaks out and decays.
Why agentic
The remaining tools should be instruments your intelligence layer can drive end-to-end. Fewer surfaces. More autonomy per surface. Total ownership of the decision logic that runs between them.
07 · Autonomy levels
From assistants → self-driving company systems
Spark builds at level 3 and 4.
Selected
The system runs the work. People review outcomes.
Workflows execute end-to-end against your operating layer. Approvals trigger only when policy says so. Eval suites catch drift in production. This is where Spark spends most of its time.
Spark focus area · L3 → L4
What changes
- — Decisions move from human to system
- — Approvals become exceptions
- — Evals replace manual QA
- — Tools become instruments
08 · SparkOS
Open spec · not a license · v0.4
SparkOS is the kernel we build into every engagement.
N.B./Not a SaaS subscription. Not a seat. An accumulating open spec, deployed and extended on every engagement — yours at handover.
L 04Humans / external
L 03Tool connectors / data sources
L 02Agents / workflows
L 01Approvals / governance
L 00Memory · Goals · Rules · Schemas
An owned, inspectable stack. Every layer is something your team can read, version, and extend. No vendor owns your company's decision logic — you do.
09 · Operating layer
Example deployment · live rotation
Company intelligence at the center. Humans, agents, and tools in orbit.
FIG · 07.1 / Operating layer schematicSCALE 1 : 1ROTATION · ↻ stableSPARK · OS · v0.4
Company intelligencememory · goals · rules · schemas
Core
Company intelligence — memory, goals, decision rules, schemas.
Inner orbit
Agents and workflows that execute against the core.
Outer orbit
Humans, SaaS tools, data sources, external systems.
Connection arcs
Governed context flow and explicit approvals.
10 · Engagement
Path · 4 phases · 12 weeks · then you operate it
How a Spark engagement actually runs.
Not a deck. Not a strategy. A working operating layer your team owns and runs. Four phases, twelve weeks, then we hand you the keys.
W 01 — 02Phase 01 · Map
P / 01
Audit your current operating layer.
We inventory every AI surface in play, locate where your decision logic actually lives, and score each workflow against the autonomy ladder. The map exists before anything is built.
What we do
- Inventory copilots, agents, automations
- Locate decision logic per workflow
- Score current autonomy (L1 → L4)
What you leave with
- A Spark Map — your operating layer, today
- A ranked list of intervention targets
W 03 — 06Phase 02 · Instrument
P / 02
Install SparkOS primitives.
We stand up memory, schemas, and approvals in an environment you own. Tool connectors are wired as instruments — not destinations. Decision rules become versioned, inspectable artifacts.
What we do
- Memory, schemas, approvals in your cloud
- Tool connectors as governed adapters
- Codify goals and decision rules
What you leave with
- An owned intelligence layer
- Logic your team can read, edit, version
W 07 — 10Phase 03 · Operationalize
P / 03
Move workflows to autopilot.
Two or three load-bearing workflows migrate to L3. Eval suites run in production. Approval boundaries are explicit — everything outside them runs unattended. This is where the operating layer starts paying for itself.
What we do
- Migrate 2–3 workflows from L1/L2 → L3
- Stand up production evals + drift detection
- Define approval scopes and timeouts
What you leave with
- Decisions executed by the system
- Approvals as exceptions, not steps
W 11 — 12Phase 04 · Hand off
P / 04
Operate without us.
Runbooks, eval dashboards, and extension patterns transfer to your team. We hand over a 90-day plan to migrate further workflows from L3 → L4. Ongoing support is optional, quarterly, and on your terms.
What we do
- Runbooks + eval dashboards transferred
- L3 → L4 migration plan, 90 days
- Team enablement on extension patterns
What you leave with
- Full ownership. We're optional.
- A path to self-driving, on your timeline
11 · Offering
Two tiers · one direction · priced
Two ways in. One direction.
A fixed-scope first step and a full implementation. Both are services — high-trust, hands-on, owned at handover. SparkOS isn’t a license you buy. It’s what we build into you.
Tier IThe diagnostic
Foundation Track
A short, fixed-scope engagement. We audit your current operating layer, score every surface against the autonomy ladder, and leave you with a Spark Map and a ranked list of intervention targets. Low risk. Self-contained. Concrete.
Investment€5,000· fixed
Duration1–2 wk· scheduled
You leave with
- Your Spark Map — every AI surface, ranked
- Decision-logic inventory
- Ranked intervention targets, scoped
- A clear next step, with or without us
Engagement coverage
P/01 MapP/02 InstrumentP/03 OperationalizeP/04 Hand off
Tier IIThe full build · Spark’s focus
Implementation
The end-to-end engagement. We install primitives, codify your decision logic, move two or three load-bearing workflows into autopilot, and hand you the runbooks and eval suites to keep extending. You leave with an operating layer your team can run for the next decade.
Investment€25k–€40k· scoped
Duration8–12 wk· embedded
You leave with
- Owned memory, schemas, and an approvals model
- 2–3 workflows in production at L3 (autopilot)
- Eval suites, drift detection, audit trail
- Runbooks, extension patterns, an L3 → L4 plan
Engagement coverage
P/01 MapP/02 InstrumentP/03 OperationalizeP/04 Hand off
N / 01 · After delivery
Ongoing review, L3 → L4 migration, and extension work are available on retainer — but only when you ask. The point of handover is that you don’t need us. Optional. On your terms. Quoted case-by-case.
N / 02 · About SparkOS
SparkOS is the productization path — an accumulating open spec we extend with every engagement. It is not a license, a seat, or a subscription. You take it with you when we hand off.
12 · Intake
Stop renting your company's intelligence.
Your business is already a codebase. Spark makes it explicit, executable, and owned. Start with the €5,000 Foundation Track — a fixed-scope diagnostic that ends with your Spark Map and a clear next step. With or without us.