Reference data & reporting platform journey of an operator.

Disclaimer: this is not a named client, but a reference journey built from situations common in the sub-region. It serves to show how we reason, in what order we work and where we focus our attention. As a matter of principle, it contains no outcome figures.

Journey fact sheet

The starting situation.

Context: A telecoms operator or a public water or electricity utility, present across the whole national territory, sometimes in another country of the sub-region. Its data lives in systems that do not talk to each other: billing, network supervision, the CRM, the ERP, meter or traffic readings, branch files. Executive reporting is built by hand, every month, by extracting and consolidating in spreadsheets. Two departments present two figures for the same indicator.
Trigger: An executive committee where revenue by region differs depending on the source, tightening reporting obligations towards the sector regulator, and artificial intelligence projects announced without reliable data to feed them.
AGILICIS role: Data strategy, then architecture, implementation and skills transfer: management questions to serve, inventory of sources, choice of cloud platform and hosting region, data governance and quality, construction of dashboards domain by domain, preparation of the foundation for AI use cases.
Scope: Revenue and billing, customers (acquisition, complaints, churn), quality of service and network, finance and procurement; executive reporting, regulatory reporting, then first AI use cases on the domains whose data has been made reliable.
Solutions: Data platform on AWS or Microsoft Azure, depending on the existing landscape and the chosen hosting region · integration tools and data catalogue · dashboards on Power BI, Tableau or SAP Analytics Cloud, depending on the platforms already in place at the operator · the cloud platform's AI services for the first use cases.
Status: Reference journey, presented as an illustration. Comparable real engagements are presented in a meeting, with the consent of the clients concerned.

The story

From decision to roll-out.

The decision

Management begins by writing down the questions it wants a reliable answer to every week: what revenue by region and by offer, which customers are leaving and why, where the network degrades service, what cost per unit delivered. This short, prioritised list replaces the technical specification. It sets the domains to address and their order.

Scoping

A few weeks to inventory the sources, measure the data quality of each system and recover, indicator by indicator, the definition each department gives it. Scoping also works through the choice of cloud from Dakar: actual latency to the candidate regions, location of personal data in light of Law No. 2008-12, cost in foreign currency, volumes to transfer. It delivers a target architecture and a plan by domain.

The scenario

For this type of operator, the chosen scenario is a layered cloud platform: collection from sources, raw storage, prepared and governed data, exposure to dashboard tools. A single definition per indicator, kept in a common dictionary, with a business owner. The source systems stay in place; the platform reads them, it does not replace them. Data that must stay in the country stays there, by design.

The domains

First domain: revenue and billing, with revenue by region identical for all departments as the success criterion. Second domain: customers, cross-referencing CRM, billing and complaints. Then quality of service and the network, then finance. Once two domains are reliable, the first AI use cases are chosen on that data, not elsewhere. An internal data team is formed from the first domain and delivers the following ones itself, with our support.

Points of attention

What makes the difference in the sub-region.

A single definition per indicator

Disagreement over the figures is almost never technical: it comes from different definitions of the active customer, of revenue or of an outage. The common dictionary is negotiated between departments, with an owner per indicator and a written calculation rule. It precedes the construction of the first dashboard and is maintained as master data, not as a project document.

Latency, location and costs in foreign currency

A cloud platform is chosen from Dakar, not from a catalogue: hosting region measured in actual latency, subscribers' personal data processed under Law No. 2008-12 and declared to the CDP, transfer volumes and data egress costs estimated in foreign currency. Some data stays on site or in the country; the hybrid architecture is planned, not endured.

Quality is corrected at the source

Wrong data corrected in the dashboard stays wrong everywhere else. Anomalies detected by the platform are sent back to the source system and its owner, with follow-up. Quality checks run at every load and their results are visible to the business. Trust in the figures is earned at that price.

AI comes after the foundation

Predicting churn or network load only makes sense on reliable, historised and governed data. AI use cases are chosen on domains already made reliable, with a business result measurable by the operator itself and governance of models and training data. An accurate dashboard is worth more than a model on doubtful data.

Deliverables

What the operator holds in its hands.

The deliverables of a journey of this type, as we produce them. Without quantified indicators: real results are presented in a meeting, with the consent of the clients concerned.

Scoping

Data strategy and target architecture

Prioritised management questions, inventory and quality diagnosis of sources, comparison of cloud regions by latency, location and cost in foreign currency, target architecture, plan by domain with success criteria.

Governance

Common dictionary and quality rules

Indicator dictionary with owners and calculation rules, data model by domain, source catalogue, quality checks defined and monitored, register of personal data processing operations declared to the CDP.

Implementation

A platform that serves the dashboards

Cloud platform deployed in layers, supervised loads, executive and regulatory reporting dashboards delivered domain by domain, a single figure for all departments as the success milestone, first AI use cases scoped on the domains made reliable.

After the project

A data team that delivers on its own

Internal data team trained and equipped, data governance in operation, cloud cost control monitored, support and enhancements provided from Dakar, during the teams' working hours.

Going further

Read, compare, dig deeper.

Our offering Data & Analytics →

Our offering Cloud: AWS, Microsoft Azure, data sovereignty →

Our reading of the sector Telecoms & digital services → · Energy, mining & utilities →

Viewpoint Enterprise AI: the use cases that really create value →

From the illustration to your case

And for your data?

Thirty minutes with an AGILICIS expert to compare this reference journey with your systems, your indicators and your ambitions in AI.

Let's talk →