Industry
OSS, BSS and delivery built past launch night
OSS, BSS and delivery platforms that keep working after launch night.
Context
Technology in Telecom & Media
We work across the operational and business support stack: mediation and rating, order orchestration and fulfilment, service inventory, self-service channels, and content preparation and delivery for media businesses. The common thread is volume with a deadline — mediation that must clear a day's records before the billing run, or a live event where a delivery fault is public within seconds.
Order fallout is the usual reason operators call us, and replacing the charging system is usually the wrong response. Fallout almost always originates in service inventory that disagrees with the network, so we start by measuring where orders actually fail before anyone proposes a platform purchase. That diagnosis takes weeks and often removes the need for the programme that was already budgeted.
- 18% to 3%
- order fallout rate after fulfilment instrumentation and targeted fixes
- 740M
- usage records mediated per day within a four-hour billing window
- 2.3M
- concurrent streams served during a live event without egress overspend
Challenges
What actually keeps this sector awake
Not generic disruption — the specific constraints that shape every technology decision here.
Order fallout across stitched-together stacks
A convergent offer touches five provisioning systems built in different decades. Ten to twenty per cent of orders drop into manual handling, and the cost lands in a back office team nobody counts against the product's margin.
Mediation and rating volume against a fixed window
Usage records arrive in the hundreds of millions per day and must be deduplicated, rated and posted before the billing cycle. A reprocessing run after a tariff error competes with the same window it needs to finish in.
Service inventory that drifts from the network
Records of what is provisioned diverge from what is actually configured, so capacity planning, fault correlation and order validation all inherit the same errors. Reconciliation is manual and therefore done once a year, badly.
Rights, regions and delivery economics in media
Content rights vary by territory and window, delivery costs scale with concurrency, and a rights breach is a legal event rather than a defect. Peak live audiences make egress and encoding decisions financially material.
Approach
How we address them
Fallout measurement before platform change
Instrumentation across the fulfilment path to attribute every failed order to a specific system and cause, ranked by volume and cost. The output is a short list of fixes with expected value, which is often cheaper than the assumed replacement.
Order orchestration on open telecom APIs
Orchestration aligned to TM Forum Open API models, with compensating actions on partial failure and an operator interface that shows exactly where an order stalled and what will happen when it is retried.
Mediation pipelines with reprocessing headroom
Idempotent record handling, deduplication at ingest, and a pipeline sized so a full-day reprocess fits alongside normal load. Rating rules are versioned so a correction can be replayed against the exact rules in force that day.
Delivery and rights enforcement for media
Packaging, DRM and geographic entitlement enforced in the delivery path rather than the application layer, with per-title cost visibility so encoding ladders and CDN choices are decided on measured spend per hour watched.
Services
Relevant practices
Engineering
Microservices
Service boundaries drawn around team ownership and failure domains, not fashion.
Engineering
API Development
Contract-first interfaces with versioning, quotas and deprecation rules agreed up front.
Engineering
Backend Engineering
Services that keep the numbers right when the queue is deep and a dependency is down.
AI & Data
Data Engineering
Pipelines with contracts, tests and an on-call rota, not a folder of scheduled scripts.
Cloud & Platform
Platform Engineering
An internal platform with golden paths, run as a product with users and a roadmap.
Cloud & Platform
DevOps
Delivery pipelines, release engineering and incident practice, measured by DORA metrics.
Products
Platforms used in this sector
Data & AI
Atmora Lens
A governed semantic layer with dashboards attached, so every report agrees on the numbers.
Customer
Customer Portal
Self-service for accounts, documents and requests, built so the answers are correct rather than merely available.
Operations
Atmora Signal
Metrics, logs and traces on one engine, priced per host, with cardinality budgets per team.
Operations
Workflow Automation
Approvals, exceptions and hand-offs modelled as processes you can version, measure and change without a release.
Data & AI
Admin Dashboard
The internal console your operations team lives in: metrics, records, bulk actions and an audit trail behind all of it.