evidoria

← Back to browse

Good practice Imported

WhatsApp + LLM Chatbot for Teacher Professional Development — Bertoua, Cameroon

Cameroon · Bertoua · See the Cameroon profile

Evidence: Observational / pre–post Top 59% 47/100 · Ask Evidence Copilot about this practice

A WhatsApp chatbot delivering LLM-drafted teacher-training content was piloted with 47 primary teachers in Bertoua, Cameroon. It scored significantly higher on usability than an online-form baseline (p=.044), but connectivity costs and English-only content limited reach.

7.82
Chatbot usability score
7.33
Online-form usability score
0.046 p-value
Usability difference significance
7.41
Combined usability-experience score, chatbot
7.03
Combined usability-experience score, online forms
87.5 %
Teachers owning personal smartphones
77.1 %
Teachers using WhatsApp daily
39.6 %
Teachers preferring French content
WhatsApp + LLM Chatbot for Teacher Professional Development — Bertoua, Cameroon

Details

Maturity
Pilot
Promoter
International NGO teacher-training programme (Bertoua, East Region) with LLM-supported content
Period
Training workshop September 2025; study published on arXiv April 2026
Keywords
teacher professional development, generative AI, WhatsApp chatbot, low-connectivity contexts, primary education

Context

An international NGO piloted a WhatsApp-based chatbot that delivered LLM-drafted teacher-training content to 47 primary teachers (39 female, 8 male) across six private primary schools in Bertoua, in Cameroon's East Region, during a training workshop in September 2025.

Objectives

Test whether a WhatsApp chatbot could be a more usable alternative to traditional online-form delivery of in-service teacher professional development content.

Activities

The chatbot delivered structured pedagogical modules -- Teaching with Love & Logic, Play-Based Learning, and The Stoplight Approach -- with content drafted by an LLM and reviewed by NGO facilitators before release. Teachers engaged in roughly 30 minutes of guided daily practice over one week. Evaluation combined a bilingual pre-test (n=40), WhatsApp interaction logs (n=47), paper post-training surveys (n=47), six school-based group discussions, and 14 follow-up interviews.

Results

The chatbot scored higher than online forms on usability (7.82 vs 7.33, p=.046, d=0.37) and on combined usability-experience (7.41 vs 7.03, p=.044, d=0.32); the difference in learnability (7.00 vs 6.76) was not statistically significant (p=.212). 87.5% of teachers owned personal smartphones and 77.1% used WhatsApp daily, while 39.6% of teachers preferred French even though the content was English-only.

Conclusions

The study measured teacher-reported usability, not verified classroom or student-learning outcomes, and ran for only one week in six private schools with no sustained-use or public-school evidence. Connectivity costs and intermittent network coverage were reported as the main barrier ('the only difficulty is the network'), and English-only content partly excluded French-preferring teachers.

Implementation
Indicative cost
Low (< €50k) — Not disclosed, but likely low given reliance on existing WhatsApp infrastructure and teacher-owned smartphones.
Time to results
Short (< 1 year) — One-week pilot of daily guided practice during a September 2025 training workshop; findings published as an arXiv preprint in April 2026.
Staffing & skills
International NGO facilitators reviewing LLM-drafted content before release, LLM used to draft training modules

Conditions for success

  • Teacher smartphone ownership and daily WhatsApp use
  • Reliable network connectivity
  • Multilingual content to avoid excluding French-preferring teachers

Common failure modes

  • Intermittent network coverage and connectivity costs disrupted use
  • English-only content excluded a sizeable share of French-preferring teachers
  • No classroom/student-learning outcomes measured, only teacher-reported usability
  • Limited to private schools; no public-school evidence

Where it fits

Governance type
NGO-led pilot in private schools
Scale
6 schools, 47 teachers
Income level
low-income (Cameroon)

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful