* CLOUDWORKZ DISCOVERY V4 . SERVICE-DESIGN SHOWCASE

Ten components, ten visualisations*

The discovery framework knows that different information types want different presentation shapes. Brief and Scope wants a four-quadrant + anchor timeline. Pain Statement wants salience-ranked cards. Competition wants a 2-axis map. Personas want avatar + day-one journey strips. Architecture wants a layered stack. Each component renders in its native shape.

Components
10 / 10
Substantive V3 drafts
10
Werner gates surfaced
17
Skills built this session
2
Reading shape
Service-design
01
Brief and Scope

What we are building

What's in, what's out, by when, for whom. The scope quadrant + the three anchors that drive every sequencing decision downstream.

Scope quadrant . the canonical Component 1 shape
In
  • BOSS (free shell)
  • CLOUDZ (5 workspaces)
  • Brain Box, Brand Box
  • ACID Sales, ACID Content
  • Roxi (Persona, voice)
  • Signalz, Axiom (subscriptions)
Out
  • Not an enterprise CRM
  • Not agent infrastructure
  • Not consumer-facing
  • Not call-centre infra
  • Not US fin-services V1
  • Not Unlock
Anchors
  • 11 June 2026 . Atomic system
  • 30 April 2027 . £6.5M first close
  • 1 October 2027 . OS commercialises
  • October 2028 . 100 paying tenants
Risks
  • Substrate slip (Werner Lane 2)
  • Atomic system slip
  • Roxi production incident
  • Raise miss / tenant miss
  • Persona-fronting rejection
  • Pricing model collapse
Anchor timeline . the three dates that bind everything
11 Jun '26
Atomic system
30 Apr '27
£6.5M raise close
1 Oct '27
OS commercialises
Open Werner gates . Component 1 Three: (1) Is "What's Out" a customer-facing message or hidden? (2) Atomic-system slip cascade: does Brand Box V1 + Brain Box V1 both shift? (3) ICP precision: target £1M-£3M or £3M-£10M ARR first?
02
Problem Statement

The single pain, six shapes

Founder-led growth-stage businesses hit the £2-5M ARR coordination-tax wall. Six pain-shapes ranked by salience. The first instance of the pain-pattern-aggregator skill surfaced one new shape candidate.

Salience-ranked pain-shape cards . each shape sized by evidence weight
A

Outbound calling is slower than it needs to be

Sales team interacts with the CRM, not the customer. 40% of day in admin.

36
B

Content production bottlenecked on single approver

Founder approval gate. Drafts stall four days. Moments pass.

v3.1
C

Strategy decisions do not propagate

Monday decisions forgotten by Wednesday. No canonical record, no propagation rule.

28
D

Brand voice drifts

Nobody owns voice; no enforcement mechanism. Designer ships off-brand deck.

v3.1
E

AI is a budget line item, not a revenue contributor

Tokens not tied to outcomes. Operator-tooling waste: 122 sessions, $3,573-$9,483 burned.

30
F
New shape candidate

AI assistants produce confident output that is not verified

Trust deficit between operator and tooling. Manifest claims 11 files exist; zero do.

22
"the biggest blocker to me getting on the phone is dealing with and trying to keep up with all of the changes and the decisions and the innovations that are taking place and trying to get a plan so that we're moving forward." Tom King . Werner handover transcript V1 . 2026-05-26 . most-illustrative quote for Shape A founder coordination tax
Open Werner gates . Component 2 Two: (1) Does Shape F (new trust-deficit candidate) warrant canonical pain-shape status? (2) Does Shape E split into E1 operator-tooling-waste and E2 customer-economy-disconnect?
03
Competitive Landscape

Eight clusters, one position

Two axes structure the competitive map. Each cluster has a Cloudworkz positioning statement. V3.1 refreshes per-vendor evidence post Wave-2 competitor-profile agent.

2-axis competitive map . horizontal: general-AI ↔ sales-specific . vertical: platform-level ↔ persona-fronted
PLATFORM-LEVEL PERSONA-FRONTED GENERAL-AI SALES-SPECIFIC C6 General-purpose Lindy / CrewAI C7 Content + brand Jasper / Writer C8 Strategy tooling Notion AI C1 Enterprise CRM Salesforce / HubSpot C2 Sales engagement Outreach / Salesloft C3 Call analytics Gong / Chorus C4 AI voice agents Synthflow / Bland C5 AI cust-support Sierra / Decagon CW Cloudworkz
Per-cluster position cards
1
Enterprise CRM + AI
Salesforce Einstein, HubSpot AI, Microsoft Dynamics Copilot
Cloudworkz position: Sit on top, not replace. CRM = system of record; Signalz = salesperson's surface.
2
Sales engagement
Outreach, Salesloft, Apollo
Cloudworkz position: Calling-first stack. They focus on email; we focus on the call.
3
Call analytics + coaching
Gong, Chorus, Avoma
Cloudworkz position: Gong analyses calls humans place. Roxi places AND analyses. NEO generates coaching paths.
4
AI voice agents
Synthflow, Bland, Air, Retell, Vapi, Vocode, ElevenLabs
Cloudworkz position: Roxi adds the refinement loop other voice agents lack. Analyser-reviewer-actioner discipline.
5
AI customer-service agents
Sierra, Decagon, Ada, Intercom Fin, Forethought
Cloudworkz position: Different lane. They're inbound; Roxi is outbound. Lexi V2 closer to their shape, legal domain.
6
General-purpose AI agent platforms
Lindy, Relevance AI, CrewAI, LangGraph, Vercel AI, Copilot Studio
Cloudworkz position: Agent-as-product, not agent-as-infrastructure. Persona-fronting + ACID discipline + multi-tenant.
7
Content + brand AI
Jasper, Copy.ai, Writer, Claude API, OpenAI API
Cloudworkz position: ACID Content + Brand Box = Writer-class brand voice + approval gate + propagation rules + atomic system.
8
Strategy + decision tooling
Notion AI, ChatGPT Pro, Tegus, manual whiteboards
Cloudworkz position: No category-leader today. Brain Box productises strategy-template + decision canvas + pre-mortem.
Open Tom gate . Component 3 Net positioning statement commitment + geographic phasing (UK then EU, or UK + selective US?).
04
Personas

Four V1 personas

Each persona carries a day-one journey strip showing the surfaces they touch. V3.1 adds four back-office personas (Operator, Marketing-Ops Lead, CS Lead, Architect).

Personas with day-one journey strip . each surface a step in the workflow
SP

The Salesperson

Customer-facing . Cloudz Sales

Goal. Book 8 meetings/week. Sound natural. Learn from her own calls.

Frustration. CRM admin eats 40% of day. Last week's script not the same as this week's. Vague coaching.

BOSS
→
Cloudz Sales
→
ACID Sales
→
Roxi
MM

The Marketing Manager

Customer-facing . Cloudz Marketing

Goal. Ship campaigns on time. Brand consistency across sub-brands. Wait-list to first-call conversion.

Frustration. Brand voice drifts. Drafts stall on founder approval. Off-brand AI copy needs rewriting.

BOSS
→
Cloudz Marketing
→
Brand Box
→
ACID Content
SM

The Sales Manager

Customer-facing . Cloudz Sales oversight

Goal. Team conversion +1pp Q-over-Q. Every rep's skill score trending up. Refined canonical script.

Frustration. Reviewing 200 calls/week impossible. Coaching one-off. Script frozen because changes risk breaking variants.

Cloudz Sales
→
ACID Sales
→
Roxi review
→
NEO coaching
FD

The Founder

Cross-workspace . Tom-class

Goal. Cut coordination-tax 50% → 20%. Back on the phone. Close the next milestone.

Frustration. Every decision waits on the founder. Team can't see what founder sees. Files lost between sessions.

BOSS Home
→
All workspaces
→
Brain Box
→
Approval gates
Open Tom + Werner gates . Component 4 Three: (1) Persona-fronting framing (buying Roxi vs assembling Modules?). (2) NEO name (keep placeholder or rename?). (3) Solopreneur tier V1 or V2?
05
Play to Win

Roger Martin cascade

Five questions cascade top-to-bottom. Six sources of advantage anchor the "how to win" step. Wave 2 grilling agent pressure-tests each source against the 8 competitive clusters for V3.1.

The five-question cascade . each step constrains the next
QUESTION 01

Winning aspiration

100 founder-led growth-stage businesses on Cloudworkz OS by October 2028. £15M ARR. Category-defining.

QUESTION 02

Where to play

UK first . Western Europe second . North America selectively . 4 sub-segments (founder-led services, B2B SaaS post-PMF, professional services, investment firms).

QUESTION 03

How to win

Six sources of advantage below . ACID-grade discipline as brand promise . persona-fronted commercial framing as differentiator.

QUESTION 04

Capabilities required

Multi-tenant substrate . Roxi voice infra . ACID-grade Module library . Brand Box atomic system . BOSS shell + CLOUDZ workspaces . GTM motion replicable.

QUESTION 05

Management systems

Quarterly anchor-check . architect rule-write authority . founder client + operator + admin model . ACID discipline as template . vault-as-canonical-state.

Six sources of advantage
01
ACID-grade discipline

Analyser-reviewer-actioner refinement scarred into substrate. blast_radius + 25%-clamp + sentinel detection + approval gates + dual-write.

02
Persona-fronted commercial framing

Customers buy Roxi, Lexi, NEO as products. Reuse invisible; productisation visible.

03
Multi-tenant substrate from day one

Werner Lane 2 carries the cost upfront. Every tenant on the same substrate.

04
Founder + architect rule-write authority

Customers buy the discipline along with the product. OS does not degrade ad-hoc.

05
Brand-voice enforcement at substrate

Brand Box gates every customer-facing artefact. Multi-brand from day one.

06
Founder dogfooding

Cloudworkz built by a founder-led team using Cloudworkz to sell Cloudworkz.

Open Tom + Werner gates . Component 5 Three: source-of-advantage commitment + persona-fronting acceptance + ACID-grade SLO definition.
06
Information Architecture

Six BOSS slots, five workspaces

BOSS shell six top-level slots, each CLOUDZ workspace standardised to 4 sub-slots + workspace-specifics. Cross-surface workflows traverse the IA via Skills + Connectors.

BOSS sitemap . customer-facing top-level navigation
BOSS shell*
01Home
02Marketplace
03Connectors
04Workspaces
05Skills
06Settings
The five CLOUDZ workspaces . per-seat operating layer
Cloudz Sales
ACID Sales . Roxi . Signalz . Axiom
Daily pipeline snapshot, call queue, canonical script library
Cloudz Marketing
Brand Box . ACID Content
Brand foundation manager, content calendar, campaign approval gate, wait-list dashboard
Cloudz Legal (V2)
Lexi (V2)
Contract review queue, perimeter language gate, compliance posture board
Cloudz Ops
Cross-product oversight
Operator surface, back-office team blueprints, scheduled-task health
Cloudz Finance
Cross-product oversight
Driver model, investor reporting cadence, runway dashboard, revenue recognition
Open Werner gates . Component 6 Six: workspace standardisation depth, settings vs workspaces top-level redundancy, Skills marketplace V1 or V2, per-Product IA standardisation, cross-surface IA enforcement, Module-instance surface visibility.
07
Tech Architecture

Five-layer substrate stack

The OS layered architecture inherited from cw-os-matrix. Each layer carries named inventory. Substrate below carries Supabase + Cloudflare + ElevenLabs + external APIs.

Five-layer stack . customer-facing at top, substrate at bottom
L5
BOSS shell + Renders
Customer-facing surface + hosted HTML team-read
6 live renders + 4 Cowork artefacts
L4
Products + Modules
Commercial layer composed from reusable Modules
9 Products + 7 Modules
L3
Vault
Canonical state. Source of truth for every artefact
17 Relay-shared folders
L2
Skills + Engines + Scheduled Tasks
Vault-internal capabilities + cron-runtime
24+ Skills, 10 Engines
L1
Connectors (MCP)
Bridges to external services
11 MCPs + V1 adds
SUB
Substrate
Multi-tenant Systems database + Cloudflare + ElevenLabs + external APIs
Werner Lane 2
Deployment topology . where each layer lives
Cloudflare

Pages (static), Workers (compute for MCP + scheduled tasks), DNS, Access (auth)

Railway

os-mcp server bridging Claude clients to vault via Relay protocol

Supabase

Systems database, multi-tenant. Werner Lane 2 work. Row-level security per tenant_id.

ElevenLabs

Voice runtime for Roxi. Per-tenant voice IDs. Cloudworkz wraps discipline (analyser-reviewer-actioner)

External APIs

Pipedrive, Slack, Drive, Notion, Fireflies, Gmail, Calendar, Apollo, Aircall, Mailgun wrapped by Connectors

Open Werner gates . Component 7 Five architectural questions: multi-tenant Q1 2027 confirm, Module-instance pattern, Brand-Voice latency budget, Connector auth model, MCP server deployment at scale.
08
Back-Office Blueprint

Five teams, one blueprinted

Five operating teams behind the customer-facing layer. Sales Ops V1 blueprinted via the back-office-process-mapper skill (first instance). Marketing Ops V1.5; Support/Finance/Legal V2.

Five operating teams . current state
Sales Operations
Operator (Roy)
4 today . 4-5 at V1
Blueprinted
Marketing Operations
Marketing Ops Lead (Fiorella)
1+2 contractors
V1.5 target
Customer Success
CS Lead (TBD)
0 (founder-served)
V2 (post-tenants)
Finance + Reporting
Finance Lead (William)
0.3 today
V2 (~Q1 2027)
Legal + Compliance
Legal Lead (Tony advisor)
0.1 today
V2 (post-commercialisation)
Sales Operations service blueprint . swim lanes (V1 instance)
Frontstage
Outbound call placed
WhatsApp follow-up sent
Meeting booked
Backstage
Roy generates 07:00 GMT snapshot
Tom reviews queue
Roy manages contractor outputs
Pipedrive hygiene
EOD inputs staged
Supporting
daily-pipedrive-snapshot SOP
transcript-ingest skill
Pipedrive A1-A6 automations
WhatsApp draft-review-send loop
Evidence
Daily snapshot .md
Aircall recordings
Pipedrive deal records
Vault transcripts + decisions
Open Werner + Tom gates . Component 8 Five: Marketing Ops V1.5 timing, Customer Success V2 trigger threshold, Finance Ops V2 owner, Legal Ops perimeter posture, single-point-of-failure mitigation pattern.
09
Data Architecture

Core entities

Multi-tenant from day one. Row-level security per tenant_id. Module instances carry tenant_id (Werner gate). Audit log immutable. Full V3 draft at sprint-9-data-architecture-V3.md (~740 lines).

Core entities . owning service + tenancy scope
T
Tenant
Identity + provisioning
Global
U
User / Seat
Identity
Per-tenant
W
Workspace
Workspace service
Per-tenant
P
Product instance
Product service
Per-tenant
M
Module instance
Module service
Per-tenant + per-product
D
Deal / Lead
ACID Sales + Pipedrive
Per-tenant
C
Call record
Roxi + Aircall
Per-tenant
F
Content draft
ACID Content
Per-tenant
B
Brand foundation
Brand Box
Per-tenant multi-brand
X
Decision record
Brain Box + Vault
Per-tenant
$
Token wallet entry
Billing
Per-tenant per-user
L
Audit log entry
Substrate audit
Immutable append-only
Open Werner gates . Component 9 Headline question: Module-instance pattern modellability in Systems database. Werner confirms (or pushes back on) Module-state-table-per-instantiation pattern. 11 other schema-shaping questions resolve in full V3 draft.
10
Backlog and Roadmap

Seven quarters to 25 tenants

Three anchors drive sequencing. The roadmap below shows the 18-month path. Risk register heatmap below shows what could derail it.

Quarterly roadmap . Q3 2026 → Q1 2028 . anchors highlighted
Q3 2026
11 Jun: Atomic system
31 Jul: Brain Box V1
31 Aug: Brand Box V1
ACID Sales V1 internal
ACID Content V1 internal
Roxi V1.5
30 Sep: Tom-on-phone 60%
Q4 2026
Cloudworkz-OS internal team
10 design-partner prospects
Signalz V1
Axiom V1
First paid pilot scoped
Customer-interview commission
Q1 2027
3 design partners contracted
V1 product spec frozen
Werner Lane 2 multi-tenant
Lexi V1 briefed
NEO V1 briefed
Investor materials ready
Q2 2027
30 Apr: £6.5M first close
V1 launch ready
First paying tenant 30 Jun
Roxi V2 multi-tenant
Q3 2027
1 Oct: OS commercialises
5 paying tenants
Solopreneur tier briefed
Q4 2027
10 paying tenants
£500K ARR
Lexi + NEO shipped
Q1-Q2 2028
25 paying tenants by 30 Jun
£3M ARR
Marketing Ops V1.5
Customer Success V2
Risk heatmap . likelihood × impact (top risks placed)
Low impact
Medium impact
High impact
High likelihood
R12 Sales Ops single-PoF
Medium likelihood
R7 Pricing collapse
R11 Operator-pain evidence base
R1 Werner substrate slips
R4 Raise miss
R5 Tenant miss
R8 Multi-tenant DB
Low likelihood
R6 Persona-fronting rejection
R2 Atomic system slip
R3 Roxi incident
R9 ACID Content refactor
R14 Werner unavail Wave 3
Open Tom gates . Component 10 Three sequencing decisions: lead product at MVP (Signalz / Brain Box / ACID Sales?) . geographic phasing (UK then EU, or UK + selective US?) . first-tenant sales motion (founder-led V1 or partner-led V1?).