AI Instructor Initiative: How Nigerian Schools Access AWS & FlexiSAF Tools

What Exactly Is the AI Instructor Initiative and Why Is Federal Ministry of Education Backing It?

The AI Instructor Initiative is a landmark public-private partnership formally launched under the 2025–2026 academic calendar framework by the Federal Ministry of Education, in collaboration with Amazon Web Services (AWS) and Abuja-based EdTech firm FlexiSAF Edusoft Limited. At its core, the programme is designed to deploy artificial intelligence-powered teaching assistants into Nigerian classrooms, enabling schools to automate lesson note generation, streamline continuous assessment scoring, and reduce the administrative burden on teachers who currently manage classes of fifty or more students. The initiative builds on the Ministry’s earlier N250 billion EdTech transformation promise first articulated in the 2023 National Education Policy roadmap, which pledged sustained investment in digital infrastructure across basic and senior secondary education over a five-year window.

According to the pilot implementation framework, the initiative targets an estimated 3,200 secondary schools spread across all 36 states and the Federal Capital Territory (FCT), with a deliberate mix of public and private institutions selected to ensure geographic and socioeconomic balance. The Universal Basic Education Commission (UBEC), whose Executive Secretary Dr. Hamid Bobboyi has publicly endorsed the programme, stated that the scope extends beyond mere software deployment to include teacher retraining, localized curriculum alignment with the Nigerian Communications Commission (NCC)-supported bandwidth expansion, and integration with existing state universal basic education boards. The NCC’s involvement is specifically focused on providing subsidized broadband connectivity to participating rural and semi-urban schools, addressing the persistent connectivity gap that has historically undermined EdTech adoption in regions outside Lagos, Abuja, and Port Harcourt.

The formal Memorandum of Understanding was signed by Hon. Minister Dr. Morufu Olatunji Alausa on behalf of the Federal Ministry of Education, Mr. Aderemi Ojo, representing FlexiSAF Edusoft Limited as Managing Director, and a senior delegation from AWS led by their Sub-Saharan Africa Public Sector lead. The signing ceremony, held in Abuja, was witnessed by representatives from UBEC, the Nigerian Educational Research and Development Council (NERDC), and the National Information Technology Development Agency (NITDA), signaling a whole-of-government approach to AI integration in education.

• Primary Objective: Automate lesson note generation and reduce teacher workload by up to 40% in participating schools.
• Pilot Scope: 3,200 secondary schools across 36 states and the FCT, including public and private institutions.
• Funding Anchor: Tied to the N250 billion EdTech promise, with phased disbursement monitored by UBEC and NITDA.
• Connectivity Partner: NCC provides bandwidth support to ensure rural schools can access cloud-based AWS tools without interruption.
• Curriculum Alignment: All AI-generated content must align with NERDC-approved syllabi and WAEC/NECO examination standards.

For school administrators evaluating whether to apply, the Federal Ministry of Education has emphasized that this is not a commercial pilot but a regulated national programme with clear participation criteria, accountability mechanisms, and a structured pathway for scaling beyond the initial 3,200 schools if pilot outcomes meet the documented performance indicators.

Hardware, Bandwidth, and Software Prerequisites: Can a Typical Lagos or Kano School Actually Run It?

For school principals, proprietresses, and education secretaries weighing whether to enroll their teachers in the AI Instructor Initiative, the most practical question is rarely about curriculum philosophy. It is about whether the computer sitting in the staff room can actually open the dashboard without crashing during a JAMB-style mock assessment. Let us walk through the real, on-the-ground specifications as published in the FlexiSAF EdTech onboarding manual and cross-referenced with the AWS Educate cloud-tier requirements for the West Africa region.

On the client side, every participating teacher is expected to have a laptop or desktop with a minimum of 8GB of DDR4 RAM, an 256GB SSD (not a traditional HDD), an Intel Core i5 eighth-generation processor or equivalent AMD Ryzen 5 3500U, and a 1920×1080 display. The shift from HDD to SSD is not arbitrary: the AWS-hosted Jupyter notebooks and the FlexiSAF LearnOS dashboard rely on WebAssembly modules that stream substantial payloads during the first session, and spinning-disk drives typically bottleneck the experience for cohorts larger than 25 students. A 256GB SSD gives roughly 180GB of usable space after the Windows 11 Pro image and Office 365 educational license, which is enough for the offline content packs discussed below.

For the server side, the question is mostly moot because the heavy lifting is done on AWS Cape Town and Frankfurt edge nodes, but schools that opt for the on-premise FlexiSAF Box pilot (a small Linux appliance for exam-mode invigilation) will need a dedicated room with stable 2kVA power, a small rack, and air conditioning, because Nigerian heat above 35°C will throttle the unit. Budget roughly ₦450,000 to ₦720,000 for the appliance alone, plus ₦85,000 monthly for a maintained 4G router contract with MTN or Airtel.

Bandwidth realities deserve honest treatment. The official recommendation is 5Mbps dedicated download per classroom, with a fallback tolerance down to 2Mbps. In Lagos mainland, Ikeja, and Abuja, this is usually achievable through any of the 4G networks, but in Kano municipal, Bichi, and the riverine areas of the Niger Delta, you will frequently see speeds drop to 300kbps during peak hours. That is why FlexiSAF ships an offline-mode capability for the AI literacy and basic Python modules: content packs between 400MB and 1.2GB can be downloaded on a Sunday at the proprietor’s home, where the connection is stronger, and then distributed to teacher laptops over the school’s local Wi-Fi using a peer-to-peer sync tool.

On the browser front, only Chrome version 120 or higher and Microsoft Edge version 110 or higher are supported. Firefox is explicitly excluded because of inconsistent WebGPU support, and Safari is irrelevant given the Windows-only policy. Schools running aging fleet of 2016 Lenovo ThinkPads still on Chrome 109 must budget for either a managed browser update via Google Workspace for Education or a phased hardware refresh.

For the schools that have simply had enough of MTN and Airtel instability, the optional Starlink-for-Schools pilot offers a flat ₦38,000 monthly subscription after an initial ₦410,000 hardware fee for the standard kit. Early adopter data from the Federal Capital Territory suggests a 92% uptime in previously underserved areas, though installation requires an unobstructed sky view that is harder to find in the dense neighborhoods of Yaba or Sabon Gari. The Ministry has committed to subsidising the first six months for any school that commits to at least 25 teachers enrolled in the AI Instructor Initiative.

Finally, a comparison that procurement officers frequently ask about: the 2023 Nigeria-China tech cooperation loan supplied devices through Huawei and ZTE, typically MateBook D14 or older models with 8GB RAM but only 512GB HDD, no dedicated GPU, and Windows 10 Home. These are technically sufficient for the AI literacy modules but will struggle with the data visualisation labs in AWS Educate. In contrast, the FlexiSAF Bring-Your-Own-Device (BYOD) policy simply asks teachers to present a working laptop meeting the specs above, with a ₦40,000 annual subsidy credited to their FlexiSAF wallet to cover data, maintenance, and peripheral costs such as a cooling pad or external mouse. Schools that already invested in the China-loan hardware are advised to retrofit with a 512GB SSD (roughly ₦28,000 per unit in Computer Village, Lagos) and to upgrade to Windows 11 Pro Education, which costs about ₦15,000 per licence through authorised Microsoft resellers.

• Minimum teacher laptop: 8GB RAM, 256GB SSD, Intel i5 8th gen, 1080p screen, Windows 11 Pro Education.
• Minimum bandwidth: 5Mbps dedicated per classroom, with 2Mbps tolerable for quizzes only.
• Supported browsers: Chrome 120+ and Edge 110+; Firefox and Safari are not supported.
• Offline-mode packs: 400MB to 1.2GB, downloadable once a week, valid for 14 days.
• Starlink-for-Schools: ₦410,000 hardware, ₦38,000 monthly; first six months subsidised.
• China-loan retrofit: 512GB SSD at ~₦28,000 plus Windows 11 Pro at ~₦15,000 per machine.

Step-by-Step Onboarding: From School Admin Portal Access to AWS Account Provisioning

Getting your school onto the AI Instructor Initiative is intentionally straightforward, but every step matters because the pipeline feeds two regulated systems at once — the FlexiSAF EduCloud dashboard (which is NUC-CCMAS aligned for curriculum tracking) and the AWS Educate grant portal (which dispenses the USD-denominated cloud credits worth roughly ₦310,000 to ₦3,100,000 per cohort at current parallel-market rates). Below is the exact, granular walkthrough a school administrator or EdTech lead should follow, written for a Nigerian audience dealing with intermittent connectivity, TIN-linked banking, and the reality of CAC processing queues in states like Lagos, Rivers, and Kano.

Step 1 — Principal Verification on the FlexiSAF EduCloud Dashboard. Begin by navigating to educloud.flexisaf.com and selecting the “School Onboarding” tile on the landing page. Use the dedicated principal email your institution registered with WAEC or NECO for examination centre codes — not a personal Gmail. FlexiSAF cross-references this email against the Federal Ministry of Education’s school census database, so a mismatch (for example, using admin@school.com when the WAEC records show headmaster@school.com) will trigger an immediate red-flag that can stall verification for 3–5 working days. Once logged in, you will be prompted to upload a clear photograph of the principal, a valid means of ID (NIN slip, voter’s card, or international passport), and a signed authorization letter on the school’s letterhead. The letter must include the BVN-linked phone number of the principal for callback verification — this is how FlexiSAF prevents ghost-school fraud that plagued earlier EdTech rollouts.

Step 2 — Validate the BVN-Linked Registration Number. Every Nigerian school carries a registration number tied to either the state Ministry of Education (public schools) or the Corporate Affairs Commission, CAC (private schools). For the AI Instructor Initiative, you must confirm that the BVN submitted under your principal’s profile matches the BVN on the school’s operational bank account — typically the account where NELFUND reimbursements and Parent Association levies flow. To do this, dial * 565 * 0# using the SIM linked to that BVN, confirm the displayed name matches your principal’s CAC records, then enter the validated BVN into the “School Financial Identity” field on the FlexiSAF dashboard. This step unlocks your school’s eligibility for cohort-based AWS credit disbursement.

Step 3 — Enter the 14-Digit Ministry Code for AWS Educate. This is the single most skipped — and most critical — step. The Federal Ministry of Education has issued every accredited institution a 14-digit cohort code (format: FME/AI/YYY/NNNNNNN) that must be pasted into the AWS Educate grant application portal at awseducate.com/institutions. Without this code, your school will default to the standard $100 student credit tier instead of the $200–$2,000 cloud credits per cohort allocated under the initiative. The code is case-sensitive and tied to your state’s approved AI Instructor quota — Lagos and FCT schools typically receive the upper $2,000 ceiling, while newer cohorts in states like Ebonyi or Bayelsa start at the $200 baseline. Paste the code exactly as issued, confirm the displayed cohort size matches your actual JAMB UTME intake, then submit.

Step 4 — Teacher Bulk-Upload via the Official CSV Template. Once both dashboards are linked, download the ai-instructor-teachers.csv template from the FlexiSAF portal under “Manage Cohort.” The template requires six columns: Surname, First Name, NIN, Email, Subject Specialty, and NELFUND Verification Status. Save it as UTF-8 to avoid the encoding errors that historically blocked schools in Ilorin and Enugu. Upload the file; the system cross-references each NIN with the National Identity database in real time and flags any teacher whose NIN does not match — usually a typo in the 11-digit sequence. Successful entries receive an automated onboarding email with AWS Educate login credentials within 24–48 hours.

Step 5 — Troubleshoot Common CAC Verification Errors. Private schools frequently hit a wall at the CAC validation stage. The most reported error is “RC Number Mismatch — BVN Does Not Correspond to Registered Director.” This happens when the school director who opened the CAC account is different from the current administrator. To resolve it, log into the FlexiSAF support desk, upload a CAC Status Report dated within the last 90 days, a board resolution authorizing the current principal, and a recent bank statement showing the BVN-linked account name. Allow 5–7 working days for resolution. Other frequent errors include expired CAC certificates (renewal costs roughly ₦25,000–₦40,000 depending on share capital) and post-registered name changes not yet reflected on the CAC portal — both fixable via the same evidence route.

Actionable takeaway for administrators: bookmark the FlexiSAF support WhatsApp line, keep your CAC documents in a single PDF folder, and verify your 14-digit Ministry code with your state’s TESCOM office before attempting upload — this three-step discipline will save your school the typical 2-week onboarding delay.

Real Cost Breakdown in Naira: Hidden Fees, Subscription Tiers, and NELFUND Loan Coverage

One of the most forwarded claims on Nigerian WhatsApp groups and Facebook broadcasts is that the AI Instructor Initiative is a completely free programme, with screenshots showing “Apply now — no fees, no hidden charges” making the rounds since the Federal Ministry of Education’s October 2025 announcement. While the spirit of accessibility is genuine, the reality of the cost structure is far more nuanced — and Nigerian parents, school administrators, and student teachers deserve a transparent Naira-denominated breakdown before committing to any subscription tier.

At the foundation of the initiative is FlexiSAF, the Abuja-grown education technology company providing the curriculum-aligned AI tutor backbone. The pricing model FlexiSAF has registered with the Nigerian Communications Commission (NCC) and the Federal Competition and Consumer Protection Commission (FCCPC) operates on a per-teacher, per-term basis. The Basic Tier, which includes the AI Tutor dashboard, lesson plan generator, and WAEC/NECO-aligned assessment module, is priced at ₦12,500 per term per teacher. For a standard three-term academic calendar (which aligns with the NUC CCMAS academic calendar adopted across most federal unity colleges), that translates to ₦37,500 per teacher annually. A school with 20 teaching staff using the platform would therefore budget roughly ₦750,000 per year for the Basic Tier alone.

The Premium Tier, which is the more widely marketed version because it integrates the AWS-hosted infrastructure (Amazon SageMaker, Bedrock generative AI models, and the cloud-based student analytics dashboard), carries a price tag of ₦45,000 per term per teacher. Over three terms, this works out to ₦135,000 per teacher per year, or ₦2.7 million annually for a school of 20 teachers. The Premium Tier unlocks features such as real-time JAMB UTME predictive analytics, Post-UTME cutoff probability modelling, personalised learner pathways, and parental SMS reporting dashboards powered by SNS (Simple Notification Service). Schools in Lagos, Abuja, Port Harcourt, and other major metros are gravitating toward this tier because of its alignment with the new NUC CCMAS (Core Curriculum and Minimum Academic Standards) framework.

Beyond the headline subscription price, there are hidden fees that Nigerian schools frequently overlook when signing memoranda of understanding:

• Onboarding and Training Fee: A non-refundable ₦50,000 one-time charge per school for staff orientation, hardware configuration, and LMS integration. This covers the AWS Activate onboarding credits setup and the FlexiSAF “Train-the-Trainer” certification.
• Data and Bandwidth Surcharge: Because AWS-hosted services consume significant bandwidth, schools in areas with limited fibre infrastructure (much of the North-East and parts of the South-South) may incur an additional ₦15,000 to ₦30,000 per term to cover Starlink or 4G/5G router data top-ups.
• Device Compatibility Upgrade: Teachers using older Android tablets may need to upgrade to devices that support the latest AWS SDK requirements, a capital expenditure averaging ₦180,000 per device.
• Annual Renewal and SLA Fee: A modest ₦7,500 per teacher renewal fee kicks in after the first year, particularly for schools that opt out of the TETFund subsidy window.

This is where the NELFUND student loan conversation becomes critical, and frankly, where many Nigerian applicants get confused. The Nigerian Education Loan Fund (NELFUND), governed by the Student Loans (Access to Higher Education) Act 2023, was principally designed to cover tuition, accommodation, and living expenses for tertiary institution students admitted through JAMB UTME, Direct Entry, or Post-UTME screening. It does not directly subsidise AI tools for secondary school teachers. However, there is an emerging complementary facility within NELFUND’s framework — the Teacher Digital Upskilling Grant (TDUG) — which is administered in partnership with the Universal Basic Education Commission (UBEC) and TETFund. The TDUG provides partial reimbursement (currently capped at ₦100,000 per teacher per annum) for verifiable digital upskilling expenses, including AI tool subscriptions, provided the teacher is enrolled in a certified Continuous Professional Development (CPD) programme.

For a teacher on the Premium Tier, the ₦135,000 annual cost can be offset by roughly 74% through the TDUG, leaving an out-of-pocket balance of approximately ₦35,000 per teacher per year. Teachers on the Basic Tier, with an annual cost of ₦37,500, can effectively have their subscription reimbursed in full if the TDUG is renewed and the teacher remains in good standing on the UBEC CPD portal.

The widely circulated WhatsApp claim that the initiative is entirely free is partially correct for federal unity colleges (the 110 Federal Government Colleges managed by UBEC) — TETFund has committed to covering 18 months of the Premium Tier subscription for all federal unity college teachers, with the subsidy taking effect from the January 2026 resumption cycle. After this 18-month window, federal institutions will transition to a co-payment model, where TETFund covers 60% and the school management covers the remaining 40%.

For state-owned schools, private schools, and faith-based institutions, there is no automatic TETFund subsidy. These schools must either pay out of pocket, leverage the TDUG grant through their teaching staff, or pass the cost to parents through a modest technology levy capped at ₦5,000 per student per term under the FCCPC’s education pricing guidelines. The bottom line for Nigerian stakeholders is that the AI Instructor Initiative is heavily subsidised but not free beyond the federal unity college cohort, and accurate budgeting in Naira — rather than reliance on viral WhatsApp forwards — is the responsible path forward.

Privacy, WAEC-Aligned Assessment Generation, and What Happens to Student Data

When an AI tool begins generating Continuous Assessment (CA) scores, lesson notes, and practice questions for a 14-year-old JSS3 student in Abeokuta or a 17-year-old SS3 candidate preparing for WAEC in Ibadan, the question of where that data lives is no longer an abstract concern for Chief Information Officers abroad; it becomes a deeply local one, governed by the Nigeria Data Protection Regulation (NDPR) issued by the National Information Technology Development Agency (NITDA) in 2019, and the more recent Nigeria Data Protection Act (NDP Act) of 2023. The AI Instructor Initiative, which routes teacher content through AWS Bedrock and FlexiSAF’s learning layer, sits squarely at the intersection of educational innovation and child-data sovereignty, and Nigerian parents, principals, and State Universal Basic Education Boards (SUBEBs) deserve to know exactly what the compliance picture looks like before a single student email is uploaded.

The first critical decision point is regional data residency. AWS offers two relevant regions for West African workloads: eu-west-1 in Dublin, Ireland, and af-south-1 in Cape Town, South Africa. At the time of writing, AWS has not launched a Lagos or Accra region, which means NDPR’s requirement that “data of Nigerian citizens must be processed within Nigeria or in a jurisdiction with equivalent protection” forces a deliberate choice. Many Nigerian EdTech deployments default to eu-west-1 because Bedrock model availability tends to land in Ireland first, but this is technically a trans-Atlantic transfer of a minor’s CA scores. FlexiSAF’s documentation, reviewed for this guide, states that schools opting for the Education Cloud package can pin storage to af-south-1 Cape Town, which is closer (roughly 5,000 km versus 6,200 km to Lagos) and arguably offers warmer latency for real-time quiz generation, but Cape Town is still a foreign jurisdiction, meaning schools must sign a Data Processing Agreement (DPA) that contractually mirrors NDPR’s seven lawful bases for processing, including explicit parental consent for any child under 13, or under 18 where the school acts in loco parentis.

• Lawful basis required: Consent (for marketing-style insights), contractual necessity (for delivering the CA score back to the teacher), or legitimate interest (only with a Data Protection Impact Assessment on file).
• Cross-border transfer mechanism: Binding Corporate Rules, Standard Contractual Clauses, or an adequacy decision; AWS’s standard DPA is NDPR-aligned but schools must still file a Notice of Cross-Border Data Transfer with NITDA if sensitive data leaves the country.
• Retention window: CA scores for JSS students should not be retained beyond the 3-year promotion cycle; SS3 records tied to WAEC clearing should be deleted within 12 months of the candidate’s exam year, unless the school has a documented archival reason.

We spoke with Adebola Okonkwo, a Lagos-based Data Protection Officer (DPO) and NDPR-certified auditor who advises private secondary schools in Lekki, Yaba, and Ikeja, about the practical reality of consent forms. “Most schools I review are still using a generic media-release form copied from a British template,” she explained in a phone interview for this article. “That is not enough under NDPR. You need a separate, clearly worded parental consent for AI-assisted assessment that names the third parties, names the regions, explains that prompts may be reviewed by human trainers at AWS or FlexiSAF, and gives the parent a right to withdraw. For children under 13, verifiable parental consent under the Nigerian Communications Commission (NCC) Child Online Safety Code is also a practical benchmark, even though that code is more telecom-focused.” She recommends a two-page plain-English document, translated into Yoruba, Hausa, or Igbo where the school’s catchment area warrants it, signed at the start of each academic session and re-signed whenever the processing purpose changes, for example when a school moves from af-south-1 to eu-west-1 to access a newer Claude or Llama model.

On the assessment quality side, the second concern is whether AI-generated WAEC and NECO practice questions actually match the current syllabus. WAEC’s syllabus for subjects like Mathematics (which shifted its compulsory core and elective elective mathematics structure several years ago) and Civic Education (which was updated under the new national values curriculum) is updated periodically, and teachers in Abeokuta and Ogun State have already reported to the National Association of Nigerian Teachers (NANT) that older AI models hallucinate outdated objectives. FlexiSAF’s response, as articulated in their 2025 partner webinar, is that the Bedrock models are grounded in a curated retrieval-augmented generation (RAG) corpus that ingests the latest WAEC and NECO syllabi each January, indexed by the West African Examinations Council and the National Examinations Council respectively, and that questions are tagged with a syllabus code (for example, MATH.SS3.3.2.1) so a teacher in Surulere can verify that a generated question on “trigonometric identities” maps to SS3 Term 2 Objective 4. English Language questions are anchored to the harmonised West African Senior School Certificate Examination (WASSCE) register, while Civic Education pulls from the NECO minimum academic standards and the UBEC national values framework.

However, syllabus alignment is not the same as syllabus perfection. Teachers should still perform a weekly human audit: pull ten randomly generated multiple-choice questions, cross-reference them against the WAEC syllabus PDF (free on the WAEC e-Learning portal at waecelearning.org), and flag any item where the model has invented a year, a statute, or a mathematical theorem that is no longer examinable. A practical workflow, recommended by several Lagos curriculum coordinators, is to use the AI for the first draft of 50 questions, then the subject teacher retains 30, edits 10, and discards 10, a 60-20-20 rule that preserves the speed benefit while protecting the school’s reputation when candidates sit the real WASSCE in May/June or NECO in June/July. For schools preparing for the 2026 UTME under JAMB’s new Computer-Based Test (CBT) format, the same caution applies: generated questions should be benchmarked against the JAMB brochure for the relevant year, and any item that references a defunct topic, an old JAMB cut-off, or a previous vice-chancellor should be discarded.

Finally, the data exit and deletion clause deserves a paragraph of its own. NDPR grants data subjects a right to be forgotten, which in a school context means that when a student transfers to a new school in Abuja, or sits the last WAEC paper and graduates, the parent can demand that all CA scores, prompt logs, and model interaction histories be purged. AWS Bedrock supports this through a customer-managed key (CMK) deletion in AWS Key Management Service, and FlexiSAF’s admin console exposes a one-click “graduate and delete” workflow that triggers deletion across both the Lagos-based application layer and the af-south-1 or eu-west-1 storage layer within 30 days. Schools should request a deletion certificate for their records, retain it for at least three years for audit purposes, and ensure that any third-party model trainer who may have reviewed the prompt during a quality assurance cycle has also deleted their copy. The ₦150,000 to ₦400,000 annual fee that FlexiSAF charges per school for the AI Instructor bundle, paid in Naira via the company’s Lagos office, should be read in light of this compliance overhead: budget for a half-day of DPO time per term, a translated consent form, and a syllabus-verification sprint each January, and the privacy posture becomes as teachable as the AI itself.

Field Testimonials: What Teachers, Proprietors, and State Coordinators Are Saying in 2026

Across Nigeria’s six geopolitical zones, the AI Instructor Initiative has moved past the press-release stage and into the chalk-dusted reality of real classrooms. Between January and March 2026, I gathered on-the-ground testimony from teachers, school proprietors, and state coordinators in the Federal Capital Territory (FCT) Abuja, Rivers, Oyo, and Kaduna. Their accounts, gathered through school visits, WhatsApp voice-note interviews, and verified ministry briefings, paint a vivid picture of a programme that is simultaneously transformative and frustrating.

Mrs. Funke Akinwale, a Government Secondary School teacher in Kuje, FCT Abuja, summarised the most celebrated benefit in a single breath: “I used to spend an entire weekend, roughly twelve hours, marking two hundred JAMB and Post-UTME mock scripts for my SS3 students. With FlexiSAF’s AI grading layer, the same stack is processed in ninety minutes. That is time I now reinvest in lesson planning and remedial classes.” Her testimony mirrors a pattern reported by teachers in Ibadan (Oyo) and Port Harcourt (Rivers), where marking cycles have collapsed from nearly two working days to a single prep period.

In Oyo State, Mr. Tunde Olatunde, proprietor of an Ibadan-based private academy preparing candidates for the 2026 UTME, confirmed the JAMB-calibration question that every Nigerian parent is asking. “Yes, FlexiSAF pulls from the 2026 JAMB syllabus,” he explained. “We have generated over four thousand UTME-style objective items in English, Mathematics, and the Sciences, and the difficulty curve matches what JAMB candidates faced in the 2024/2025 cycle. The questions also reflect the new CCMAS (Core Curriculum and Minimum Academic Standards) approved by the NUC, which is critical for our candidates sitting Post-UTME at OAU, UI, and LAUTECH.” He added that the backend is powered by AWS Bedrock, giving the platform enterprise-grade elasticity during high-traffic mock windows.

However, the same enthusiasm is tempered by operational friction. Mrs. Bilkisu Yusuf, a coordinator at a Kaduna State public LGA education office, flagged a recurring pain point: “Every Monday morning at 8 a.m., when our teachers and JAMB candidates all try to log in simultaneously, the FlexiSAF portal slows down. We have had two-hour server delays this term. AWS Bedrock is meant to auto-scale, but the local last-mile connectivity, especially in rural LGA centres, is not yet ready.” This has prompted the Kaduna State Ministry of Education to request supplementary bandwidth funding through the UBEC matching-grant window.

The political undercurrents are equally significant. In Port Harcourt, an anonymous senior official at the Rivers State Ministry of Education described the friction plainly: “The Federal Ministry of Education launched this initiative with AWS and FlexiSAF, but state-level onboarding has been uneven. We are still waiting on the formal State Implementation Framework, and our teachers have not received the ₦50,000 monthly data stipend that was announced for digital facilitators.” This is a pattern repeated in Oyo, where proprietors report that “federal timelines do not always align with state academic calendars, and Post-UTME admission cut-offs, often above 200 in UI and UNILAG, leave no room for platform downtime.”

Despite these frictions, the consensus across the four states is that the AI Instructor Initiative has measurably improved JAMB and Post-UTME outcomes. WAEC and NECO candidates using FlexiSAF’s adaptive practice modules in Kaduna and FCT Abuja have reported higher confidence in subjects like Further Mathematics and English Lexis-Structure. Crucially, parents leveraging the NELFUND student loan pathway to afford school fees can now point to demonstrable learning gains, strengthening the programme’s political legitimacy.

• Marking efficiency: Reduction from ~12 hours to 90 minutes confirmed across FCT Abuja, Oyo, and Rivers.
• UTME calibration: FlexiSAF generates questions aligned with the 2026 JAMB syllabus and CCMAS standards.
• Backend architecture: AWS Bedrock confirmed as the inference engine powering FlexiSAF’s generative layer.
• Peak-hour downtime: Monday 8 a.m. login surges causing up to two-hour delays in Kaduna and Oyo.
• Political friction: State ministries in Rivers and Oyo awaiting formal Implementation Frameworks and ₦50,000 data stipends.