The verdict in three sentences
In 2026, an SME can deploy an AI assistant for customer service for 3,000,000 to 9,000,000 FCFA, with a RAG system wired to its document base without exposing sensitive information. The system includes guardrails and a human in the loop, and ships in 6 to 12 weeks. Expected outcome: response time cut by 50 to 70% for a run cost of 100,000 to 400,000 FCFA/month.
Use cases ranked by ROI
Not all AI applications are equal. Start with high-volume, low-risk ones before more sensitive cases.
| Use case | Main gain | Risk | Priority |
|---|---|---|---|
| FAQ chatbot on site/WhatsApp | -50% repetitive tickets | Low | Very high |
| Email triage and routing | -3 to -6 h/day | Low | Very high |
| Assisted replies (drafts) | +40% agent speed | Medium | High |
| Call summaries / minutes | Better traceability | Low | Medium |
| Product recommendation | +10 to +20% basket | Medium | Medium |
| Dissatisfaction detection | Improved retention | High | To scope |
The FAQ chatbot and email triage deliver the fastest return, often in the first month.
Build custom or take an off-the-shelf solution
Two approaches. Generic SaaS solutions start fast but adapt poorly to your business and sometimes expose data; custom (RAG on your documents) costs more but stays controlled and confidential.
| Criterion | Off-the-shelf SaaS | Custom AI assistant (RAG) |
|---|---|---|
| Setup cost | 500,000 to 2,000,000 FCFA | 3,000,000 to 9,000,000 FCFA |
| Monthly cost (run) | 150,000 to 500,000 FCFA | 100,000 to 400,000 FCFA |
| Knowledge of your business | Low | High |
| Data confidentiality | Variable | Controlled (dedicated hosting) |
| Guardrails / human in the loop | Limited | Custom |
| Timeline | 2 to 4 weeks | 6 to 12 weeks |
In practice, you wire the AI to a document base (procedures, catalogue, FAQ) for reliable, sourced answers.
Run costs in detail
Monthly cost depends on exchange volume and the model used. Here is a 2026 order of magnitude for an SME.
| Monthly volume | API calls (AI models) | Hosting + supervision | Total run/month |
|---|---|---|---|
| 2,000 conversations | 60,000 FCFA | 60,000 FCFA | ~120,000 FCFA |
| 5,000 conversations | 130,000 FCFA | 90,000 FCFA | ~220,000 FCFA |
| 10,000 conversations | 240,000 FCFA | 120,000 FCFA | ~360,000 FCFA |
| 15,000 conversations | 310,000 FCFA | 140,000 FCFA | ~450,000 FCFA |
A human validates sensitive replies: the AI proposes, the agent decides on high-stakes cases.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Mini case study
Aissatou, director of an 18-person services SME, was buried under repetitive requests on WhatsApp and email. Chosen project: custom AI assistant (RAG) wired to her procedures and catalogue, 5,500,000 FCFA, delivered in 8 weeks, run at 250,000 FCFA/month. Result: 55% of requests handled automatically, average response time down from 6 h to 1 h 45, and 2 agents reassigned to sales. Estimated saving: ~3,200,000 FCFA/year in agent time, excluding additional revenue.
FAQ
Will the AI expose our sensitive data?
No if the project is well designed: dedicated hosting, RAG on your authorised documents only, and isolation of confidential data. Nothing is used to train a public model.
Do we have to replace our agents?
No. The AI absorbs the repetitive volume; your agents focus on complex cases and sales, with a human in the loop on sensitive replies.
How long before we see results?
From go-live, generally 6 to 12 weeks after project kickoff. The gain on response time is visible in the first month.
Does it work on WhatsApp?
Yes, WhatsApp Business is a priority channel; the assistant integrates there alongside the website and email.
What monthly run budget should we plan?
Between 100,000 and 400,000 FCFA/month in 2026 depending on conversation volume and the chosen model.
Let's scope your project. Describe your channels (WhatsApp, email, site) and your monthly request volume, for a quote between 3,000,000 and 9,000,000 FCFA depending on scope. Detailed quote within 48 h. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
