08 — Telecom & Media
Network, service and subscriber economics modelled as one engine — where a truck roll, a dropped call and a save offer all resolve to the same currency.
North Star
Margin per subscriber month
The value engine, end to end
The same six moves we run everywhere, expressed in the physics of this industry.
Where does subscriber margin leak — network, service or billing?
Margin leakage map
What actually causes churn and repeat contact?
Causal churn and service model
What happens if we shift capex or save-offer policy?
Network and subscriber twin
Who resolves and retains at volume?
Service and retention agents
Is every automated decision fair and logged?
Consumer-duty control file
Does ARPU hold after the promotion ends?
Outcome contract + cohort monitor
Where the value is
Before a twin is built, we agree where the money sits and what moves it.
Value pool
Subscriber retention
Driver
Service experience and rate perception
Measured as
Retained margin per save dollar
Value pool
Cost to serve
Driver
Contact volume, repeat contact, truck rolls
Measured as
Cost per resolved contact
Value pool
Network capex
Driver
Where congestion and failure actually occur
Measured as
Return per capex dollar
Value pool
Revenue assurance
Driver
Rating accuracy, unbilled usage, credits
Measured as
Leakage as percent of billed revenue
Use cases
Each one is a number somebody already owns, a twin that models it, and an agent that moves it inside a defined authority. Filter by function.
7 use cases
Infers the likely fault before the customer finishes describing it, resolving with the minimum number of questions and the correct fix first time.
Summarises the call from audio, decides whether the issue was actually resolved, and routes the ones that were not before the customer calls back.
Predicts churn on behaviour rather than tenure and sizes the save offer against the customer's actual margin.
Predicts congestion and equipment failure per cell and per route, so capex and truck rolls follow evidence.
Finds the failure modes in the fleet's telemetry and schedules maintenance where it avoids the most missed appointments.
Connects content spend to watched minutes and retained subscribers, with attribution that survives a channel change.
Finds the revenue that never reached the invoice — rating errors, unbilled usage, misapplied credits — and closes the leak at source.
Outcome twins
Each twin owns a set of levers. Each feature moves one lever. Each agent executes inside a defined authority.
Model each subscriber's margin, churn risk and true save value.
Levers
Features
Multi-agent layer
Retention Agent
Offers only what the subscriber's margin justifies.
Resolve the fault the first time and stop the repeat contact.
Levers
Features
Multi-agent layer
Diagnosis Agent
Infers the fault before the customer finishes describing it.
Resolution Agent
Closes standard cases end to end with audit trail.
Put capex and crews where the evidence says the network will fail.
Levers
Features
Multi-agent layer
Capacity Agent
Flags cells approaching congestion before complaints start.
Find the revenue that never reached the invoice.
Levers
Features
Multi-agent layer
Assurance Agent
Reconciles usage to billing daily and raises the delta.
Unit economics
Set your volume, switch features on or off, and move each impact to your own evidence. Everything is expressed per subscriber month.
Subscribers
2,400,000
Contribution per subscriber month
$9
Monthly subscriber contribution; substitute per-account economics for enterprise segments.
Annual program investment
$3.5m
Service Twin
Unit cost — Cost to serve
6.0% — Fewer questions, fewer repeat contacts, fewer truck rolls.
$4.2m
annual value
Subscriber Twin
Volume — Retained subscribers
2.5% — Retention spend follows margin, not tenure.
$540k
annual value
Revenue Assurance Twin
Leakage — Billing leakage
25.0% — Usage reconciled to invoice every day.
$1.6m
annual value
Network Twin
Unit cost — Capex per served gigabyte
3.0% — Capex lands where congestion is measured, not assumed.
$2.1m
annual value
Subscriber Twin
Unit revenue — ARPU
1.5% — Right plan, offered before the complaint.
$1.4m
annual value
Value by feature
Annual value contributed by each active feature at current settings.
Value and ROI vs volume
How engineered value and ROI move as subscribers change. The marker is your current setting.
Cumulative value vs investment
Where the curve crosses the investment line is payback.
Value mix by lever
Which lever the engineered value is actually coming from.
Baseline contribution
$21.6m
Engineered annual value
$9.8m
ROI on program
179%
Payback
4.3 months
Directional model. In an engagement every number here is replaced by your measured baseline, signed off by finance, and written into the outcome contract before any agent goes live.
Scale
Prove the churn and service model on a bounded subscriber base.
Same twins, new cohorts; guardrails and consent rules carried across.
Shared runtime, per-market regulation, one margin-per-subscriber number.
Integrations
Vendor-agnostic by design. Connectors are added per engagement — no platform lock-in.
BSS / OSS
Service & CRM
Data & cloud
Network analytics
Regulations and standards
Compliance is a design input, not a review gate. Every agent action is logged, attributable and reversible.
Consumer duty / fair treatment
Save offers and pricing must be demonstrably fair, not merely legal.
GDPR / ePrivacy
Consent for behavioural targeting; call recording and transcript handling.
Lawful intercept and retention
Strict boundary between analytics data and regulated retention stores.
EU AI Act
Transparency obligations for automated customer-facing agents.
Accessibility (EAA / WCAG)
Automated service channels must remain usable by everyone.
Standards supported
Next industry
Transport & Logistics
