evidoria

← Back to browse

Good practice Imported

Shikho: A Venture-Funded AI-Personalised Exam-Prep Platform Reaching All 64 Districts of Bangladesh

Bangladesh · Dhaka · See the Bangladesh profile · See the Dhaka profile

Evidence: Descriptive / self-reported Top 89% 27/100 · Ask Evidence Copilot about this practice

Shikho, a Bangladeshi edtech startup, combines recorded lessons, testing analytics and an AI recommendation engine to personalise exam prep for over 4 million registered learners across all 64 districts, backed by $8m+ in venture funding.

4,000,000+ learners
Registered learners
64 districts
Districts reached
8,000,000+ USD
Total venture funding raised
1,300,000 USD
Seed funding round (2021)
90 %
Self-reported subject-identification accuracy
15 %
Self-reported average improvement from personalised recommendations

Details

Maturity
Scaling
Promoter
Shikho
Period
2021-present
Keywords
edtech, private sector, personalized learning, mobile app

Context

Shikho is a Bangladeshi edtech company, founded in Dhaka, that provides recorded lectures, structured study notes and practice tests aligned to the National Curriculum (NCTB) for students preparing for the SSC and HSC secondary exams, alongside a Bangla-language AI assistant that answers curriculum-linked questions.

Objectives

The company states it uses embedded AI — a recommendation engine built on behavioural learning telemetry — to prompt each student toward the specific videos and exercises matched to their test performance, rather than a one-size-fits-all course sequence.

Activities

Independent business press confirms the company's scale and financing: TechCrunch reported a $1.3 million seed round in 2021 co-led by LearnStart and Anchorless Bangladesh, and Bangladeshi business outlets subsequently reported additional rounds bringing Shikho's total funding beyond $8 million.

Results

Company figures cite over 4 million registered learners across all 64 districts of Bangladesh, the majority from outside the major cities. Shikho's own product blog reports a '90% accuracy' in subject identification and an average '15% improvement' from personalised recommendations, with no published methodology, sample size or comparison group for either figure.

Conclusions

The verifiable evidence here is one of substantial commercial scale and investor confidence in an AI-personalised learning model for a large, high-stakes exam market, rather than of independently confirmed learning gains. As a paid subscription product, Shikho's personalisation features are also not free to all students, which limits how much its rural, lower-income reach translates into equitable access to the AI-driven features specifically.

Implementation
Indicative cost
Medium (€50k–€500k) — Venture-funded: $1.3M seed round (2021) growing to $8M+ total raised across subsequent rounds.
Time to results
Long (> 3 years) — Operating since 2021 through at least 2025-2026, an ongoing commercial product rather than a time-boxed pilot.
Staffing & skills
Shikho (Dhaka-based edtech company), Investors: LearnStart and Anchorless Bangladesh (2021 seed round) and subsequent backers

Conditions for success

  • Bangla-language content aligned to the national curriculum (NCTB) and high-stakes exams (SSC/HSC)
  • Sustained venture funding (seed round growing to $8M+ total) enabling national distribution across all 64 districts

Common failure modes

  • Core AI-personalisation features are paid/subscription-gated, limiting equitable access for the free-tier majority of its rural/lower-income user base
  • Key pedagogical claims (subject-identification accuracy, improvement from personalisation) are unverified company statistics with no independent methodology, sample size or comparison group published

Where it fits

Governance type
private, venture-backed company
Scale
national (all 64 districts of Bangladesh)
Income level
lower-middle-income

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.

Similar practices you may find useful