- Home
- Our services
- AI & automation
Digital technologies and solutions
Useful AI, inside your processes.
No generic AI: use cases that read your documents, understand your SAP and Salesforce master data and act inside your transactions, under supervision proportionate to the risk. And, alongside it, classic process automation, often sufficient and always cheaper. Our data and AI teams in Dakar design, measure, govern and industrialise these use cases with you.
The problem
Prototypes everywhere.
Production, almost nowhere.
Generative AI demos impress, then stop. Without a connection to the core business system, without reliable data or a governance framework, the prototype remains a prototype. The AI that matters reads your real documents, understands your master data and acts inside your transactions, under control. That work is as much engineering as it is models. That is where we put our teams.
The use cases
Four families worth the effort.
Enterprise AI is judged inside a process, never in a tool catalogue. Four families of use cases we know how to examine, measure and integrate into SAP and Salesforce. The use cases that really create value →
Document extraction
Supplier invoices, delivery notes, bills of lading, identity documents, contracts: reading, field extraction, automatic checking and matching, all the way to posting or case creation in SAP or Salesforce.
Assistants
Assistants grounded in your data and your knowledge base: business users query the system in natural language (a figure, a status, a procedure) and get an answer based on your master data, not on a guess.
Agents
Agents that detect an anomaly, prepare the action and trigger it in SAP or Salesforce, according to explicit rules and an appropriate level of supervision: exception handling, incident resolution, qualified reminders.
Forecasting
Forecasting of demand, collections, stock-outs or customer churn, computed on your real data and enriched with external signals when they add something. A figure to discuss, not a black box.
By process
Where AI and automation change the daily routine.
- Finance
Supplier invoice extraction and matching · closing support: detection of unusual entries and variance commentary · assisted matching of receipts, including mobile payments.
- Procurement & supply chain
Compliance checking of purchase requests against the contract · stock-out and supplier delay alerts with recommendations · demand forecasting and stock optimisation by product family.
- Sales & customer service
Qualification and routing of incoming requests in Salesforce · assisted replies based on customer history · detection of at-risk complaints and customers likely to leave.
- Operations & logistics
Reading of transport and customs documents · container and delivery tracking · assisted planning of routes and resources.
- IT & run
Incident analysis and resolution proposals from the knowledge base · assistance with development and code review · monitoring of integration flows.
The prior question
Where an agent is relevant. Where classic automation is enough. Where nothing should be automated.
The useful question is not "can we put AI here?", but "do we need AI here?". We answer it before building.
Relevant
When the situation varies with every case, the decision requires cross-checking several sources and the action remains reversible or supervisable: exception handling, incident resolution, purchasing or replenishment recommendations.
Classic, and sufficient
When the rule is stable and known: a workflow, a validation rule, a software robot (RPA) or a standard integration does better, cheaper and more safely than an agent. Fixed-rule matching or scheduled reminders belong here.
Automate nothing
When the stakes are irreversible, regulatory or human: approval of a payment above threshold, a credit decision, a labour relations decision. AI prepares the file. The decision remains whole.
Process automation
On SAP and Salesforce, before AI and with it.
- Workflows & approval flows
Purchase requests, expense reports, contract approvals, account openings: explicit, traced flows with delegations and reminders, in SAP, in Salesforce or between the two.
- Software robots (RPA)
For repetitive tasks that cross several screens or several systems without APIs: data entry, checks, extracts. A useful bridge, provided it is documented and its end is planned for when integration becomes possible.
- Built-in automation
Native platform automation (SAP Build, Salesforce Flow), API integrations, document generation. What a stable rule can settle does not need a human intervention every time.
- AI-augmented automation
When the rule is not enough (a document to read, a request to qualify, an exception to handle), AI takes over within the same flow, with a human checkpoint where it is needed.
Governance
AI that is accountable.
The level of human supervision is proportionate to the risk, the materiality and the confidence level of the use case: systematic approval on actions that commit the company, sample-based checks where confidence is established, bounded autonomy on reversible tasks. Every agent decision is logged: what was seen, what was proposed, what was done, and by whom. Personal data processed by AI falls under Law No. 2008-12 and the CDP. We design every use case within this framework.
Our approach
From prototype to industrialisation.
Identify
Use cases with measurable value, prioritised with the business from your real processes, and the answer to the prior question: agent, classic automation, or nothing.
Prototype
A framed prototype, on your data, with a success criterion defined in advance and a short deadline. Feasibility is demonstrated in a few weeks.
Measure
Accuracy, straight-through processing rate, time saved, errors avoided: the value produced against the initial criterion, before any extension.
Govern
Level of supervision, traceability, compliance, control of risks and of model drift over time. The framework that makes adoption possible.
Industrialise
Integration with SAP and Salesforce, data security, extension to volumes, entities and languages, team training, model maintenance. A service that holds the load over time.
The technology
The building blocks we put to work.
Use case workshop
What are your two or three highest-value use cases?
A short workshop with our data and AI teams to identify them from your real processes, and to rule out those better left unautomated.
Let's talk →