The verdict in three sentences
An insurer receiving 25,000 WhatsApp messages a month can hand 55 to 70% of requests to an AI bot connected to its policy administration system, provided the scope is limited to certificates, claim tracking and premium due dates. The 2026 budget sits between 10 and 25 million FCFA (about $17,500 to $44,000), with a 2 to 3 month timeline, plus a monthly running cost driven by the WhatsApp Business API and the AI model. The figures below come from an insurer case in Abidjan; a Dubai insurer applies the same architecture, with budgets usually quoted in USD or AED and data rules set by the UAE data protection law. ROI comes from redeploying 4 to 5 agents and cutting response times from hours to under a minute.
Which requests a bot can really handle
Not all messages are equal. The first job is to classify the last three months of conversations. For property, casualty and life insurers, the typical split looks like this, with a 2026 order of magnitude for the no-agent resolution rate.
| Request type | Share of messages | Target no-agent resolution |
|---|---|---|
| Motor insurance certificate or regional motor card | 30% | 85 to 95% |
| Claim notification (photos, accident report) | 20% | 40 to 60% (complete pre-notification) |
| Status of an open claim | 18% | 70 to 85% |
| Due dates, premiums and payment link | 15% | 75 to 90% |
| Quotes and new policies | 10% | 25 to 40% |
| Complaints and disputes | 7% | 0 to 10% (direct escalation) |
| Weighted average | 100% | 55 to 70% |
The bot does not make anything up: it queries the policy administration system to retrieve the policy, the due date or the claim status, and answers from that data. For premium payments, it sends a payment link (Wave, Orange Money or MTN MoMo in West Africa; cards or Apple Pay in Dubai), which noticeably shortens renewal collection time.
2026 budget: investment and monthly cost
| Item | 2026 range (FCFA, excl. tax) | Approx. USD |
|---|---|---|
| Scoping, conversation analysis, scenarios | 1,000,000 to 2,000,000 | $1,750 to $3,500 |
| WhatsApp bot and AI engine (French, English and local expressions) | 5,000,000 to 12,000,000 | $8,800 to $21,000 |
| Connector to the policy and claims system | 2,000,000 to 6,000,000 | $3,500 to $10,500 |
| Escalation console for agents with history | 1,000,000 to 3,000,000 | $1,750 to $5,300 |
| Testing, staff training, go-live | 1,000,000 to 2,000,000 | $1,750 to $3,500 |
| Total investment | 10,000,000 to 25,000,000 | $17,500 to $44,000 |
| Monthly operations: WhatsApp API, AI, hosting (estimate) | 400,000 to 1,200,000 a month | $700 to $2,100 a month |
On the WhatsApp Business API, Meta charges for template messages sent at the insurer's initiative (renewal reminders, claim notifications) between $0.02 and $0.08 depending on category and country. Replies to customers who write first, within the 24-hour window, are largely free. The AI model cost depends on volume: a few US cents per complete conversation.
Conditions for success
Three points separate a useful bot from a gimmick. Escalation: as soon as a request falls outside scope or the customer gets impatient, the conversation passes to an agent with full history, so the customer never repeats themselves. Compliance: personal and health data fall under local data protection law (Law 2013-450 and the ARTCI regulator in Côte d'Ivoire), which requires a declaration, controlled hosting and a defined retention period. Monitoring: a weekly dashboard tracks resolution rate, escalation reasons and satisfaction, to keep improving scenarios.
Deployment usually follows this pace: 3 weeks of scoping and analysis, 5 to 7 weeks of development and integration, 2 to 3 weeks of pilot on a single product (often motor), then extension to other lines.
Mini case study
Aminata Koné, head of customer relations at a property and casualty insurer in Abidjan, receives 25,000 WhatsApp messages a month, about 10,000 distinct requests. Her team of 14 agents replies in 4 hours on average, and up to 2 days during renewal season.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
After a project billed 18,000,000 FCFA (about $31,600), the bot resolves 62% of requests, i.e. 6,200 a month. At 7 minutes per request, that is 723 hours, roughly 4.5 agent positions. Calculation: 4.5 × 450,000 FCFA monthly loaded cost = 2,025,000 FCFA a month, i.e. 24.3 million FCFA (about $42,600) a year. With operations at 800,000 FCFA a month (9.6 million a year), the net gain reaches 14.7 million FCFA a year and the project pays back in about 15 months, before counting faster premium collection.
FAQ
Do we need a different WhatsApp number from our current customer service line?
No, the existing number can migrate to the WhatsApp Business API. Migration takes 1 to 3 days, and the history of the standard WhatsApp Business app is not transferred.
Does the bot understand voice messages?
Yes, voice note transcription can be built in. It adds around 10 to 20 FCFA (a few US cents) per voice message and meets a real need: 25 to 35% of messages received by the insurers we observed are voice notes.
What if the AI gives wrong information?
The bot only answers from policy data and an answer base approved by the underwriting department. Complex coverage questions trigger escalation, which in practice concerns 30 to 45% of conversations.
How long before reaching 55% resolution?
The pilot usually starts around 40%. The 55 to 70% threshold is reached after 2 to 3 months of scenario tuning, based on escalation reasons reviewed every week.
Can premiums be collected directly in WhatsApp?
The bot sends a payment link and confirms receipt by message. The certificate is then generated in under a minute after payment.
Let's scope your project. Share your monthly message volume, your top three contact reasons and your policy system: we will price the bot (range 10 to 25 million FCFA, about $17,500 to $44,000) and a 2 to 3 month schedule. 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.
