Intelligence, engineered for Africa.

We design, build and run AI systems end to end: GeoAI and machine learning pipelines, AI agents and secure cloud infrastructure, with the cybersecurity and data privacy layer that lets institutions trust them.

  • AI transformation
  • GeoAI
  • Machine learning
  • Generative AI and agents
  • Cybersecurity and privacy
  • AI skills
Start a project
  • 17open-source projects you can inspect
  • 5public-sector systems built and live online
  • 4 stagesfrom workflow to published, verifiable results

AI scales through trust, not hype.

No single breakthrough changes an institution. AI spreads through diffusion: the patient work of embedding it into real organisations and real accountability structures, until it runs reliably, earns trust and keeps being used. That diffusion stalls without a trust layer.

Across Africa, startups and public institutions now hold growing volumes of sensitive data and run critical digital services, often without security expertise, tools or support built for their context. That isn't a failure of ambition; it's a capability gap. Left open, it closes the hard way: breaches that erode public confidence, failures that set adoption back, and systems that act unpredictably with no one accountable.

TovuTech closes that gap. We help organisations of every size unlock real value with practical AI, robust compliance frameworks and privacy-first data management, so they can adopt new technology with confidence. Whether you're a startup putting AI into your product or an institution strengthening governance, we work as an extension of your team.

We make AI ideas real, secure and trusted.

Abstract neural network of coloured nodes converging on a glowing hub

Practical AI, end to end

We turn complex technical problems into working systems, from use case and data pipeline to deployment, monitoring and published results.

A built-in trust layer

Security, privacy and governance are designed in from day one, aligned with Kenya's Data Protection Act and international standards.

No one left behind

We bring readily available AI tools into everyday work and train people to use them well, with a focus on young people in the Global South building sustainable digital livelihoods.

What we deliver

Use our infrastructure, adopt a ready-made solution, or build something new with us, from the first workshop to a secured, monitored system in production.

AI as a Service

Deploy powerful AI capabilities instantly through our secure cloud infrastructure, without buying or managing hardware.

  • Compute that scales. Managed services grow with demand and shrink when idle, so you pay for use, not idle servers.
  • Managed data pipelines. Ingestion, cleaning and scheduling for spreadsheets, databases, APIs and satellite imagery.
  • Models trained for you. We prepare your data, then train, fine-tune, evaluate and host models on your behalf.
  • Model monitoring. Accuracy, drift, latency and cost tracked, with alerts and a plain-language monthly report.
  • Your data, your region. Google Cloud regions you approve, or hybrid and on-premise when data must stay on site.
Discuss AI as a Service
Your apps, staff and citizensweb, mobile, USSD, WhatsApp
Secure API gatewaykeys, quotas, audit log
Models and agentsGemini, open models, models we train for you
Data pipelinesingest, clean, Earth Engine, BigQuery
Secure cloud infrastructureencrypted, monitored, backed up
You use the top layer. We run, secure and monitor everything beneath it.

AI Marketplace

Ready-to-use models, agents and workflows built for African operating realities, deployable on sovereign or hybrid infrastructure without starting from scratch.

  • Flood and drought early-warning agent
  • Trilingual citizen assistant
  • Bursary and grants workflow
  • Soil spectra analysis
  • Crop leaf disease check
  • Project monitoring map

Each one is adapted from a system we've already built. The marketplace opens in stages; ask about early access.

AI and GeoAI consultancy

Strategy-to-production delivery for teams that need real outcomes: use-case selection, data readiness, deployment, governance and performance.

  • AI readiness assessments and implementation roadmaps
  • AI risk management, ethics and data sovereignty
  • GIS systems with development partners, such as rangeland monitoring and livestock route mapping
  • Compute and cost optimisation

We work in QGIS, ArcGIS Pro and Enterprise, Google Earth Engine, Python (PyTorch, scikit-learn, LightGBM, XGBoost, FastAPI), Vertex AI, BigQuery, Looker Studio, AppSheet and Firebase.

Generative AI and the agentic web

We build with generative AI every day: assistants that answer from your own documents, content tools reviewed by experts, and websites that don't just inform visitors but act for them.

Retrieval-augmented assistants

Chatbots that answer from your policies, manuals and FAQs using retrieval-augmented generation (RAG), and cite the source instead of guessing.

AI agents with tool use

Agents that call your systems through function calling: check an order, open a ticket, book a meeting, or draft a reply for a person to approve.

Content and media generation

Reports, social posts, translations, images and voice-overs with Gemini, Imagen and open models, always checked by a human reviewer.

Fine-tuned open models

Gemma, Llama and other open-weight models fine-tuned on your data and hosted privately when information can't leave your control.

Document intelligence

Multimodal models that turn scanned forms, invoices and certificates into clean, structured data for your systems.

Evaluation and guardrails

Test sets, red-teaming, content filters and usage monitoring that keep generative systems accurate, safe and on budget.

In development

ECD curriculum and learning portal

A generative-AI learning portal for early childhood development, aligned to Kenya's Competency-Based Curriculum for pre-primary (PP1 and PP2).

  • The five pre-primary activity areas: language, mathematical, environmental, psychomotor and creative, and religious education activities
  • Stories, songs and picture activities generated with Gemini and reviewed by ECD educators before children see them
  • Kiswahili and mother-tongue audio for children who can't read yet
  • Teacher and caregiver dashboards that track competencies, not just scores
  • Offline-first and light on data, for shared phones and tablets

Builds on RoboKid, our CBC learning platform for Grades 1 to 3.

Partner with us on the pilot

Agentic websites: we design, build, host and deploy

Fast, secure websites with an AI agent built in, delivered end to end.

  1. DesignUser research, brand system, accessible UX to WCAG 2.2
  2. BuildStatic-first pages, serverless edge functions, structured data for search and AI engines
  3. DeployGit-based CI/CD to a global CDN, TLS, security headers, web application firewall
  4. OperateUptime monitoring, Core Web Vitals, analytics, content updates and the AI agent

This website is an example: static-first pages on Cloudflare's edge network, serverless functions for TovuAgent, strict security headers, and automatic deployment from GitHub on every change.

Smart Bots: turn your website into an agent

We study your current website and upgrade it with a Gemini Pro–powered assistant that answers, qualifies and acts.

  1. Audit your pages, analytics and enquiries to learn what visitors really ask.
  2. Build a knowledge base from your pages, documents and price lists, indexed for retrieval.
  3. Connect tools so the agent can capture leads, send quotes, book meetings in your calendar, and hand over to a person on WhatsApp or email.
  4. Launch and improve with one script tag, then review conversations and refine every week.

What your business gains

  • Always on. Answers at 2 a.m., on weekends and during holidays.
  • More qualified leads. Captures each visitor's need and contact details, then sends them to your inbox or CRM.
  • Meetings booked for you. Offers times from your Google or Microsoft calendar and sends invitations.
  • Less repetitive work. Routine questions are handled, so staff focus on real conversations.
  • Multilingual. English, Kiswahili, Somali, French and more.
  • Demand insight. A dashboard of top questions and unmet needs feeds your next product or campaign.
Live: agent-ready

This website talks to AI agents, too

More and more people ask an AI assistant instead of searching. So tovutech.com runs its own Model Context Protocol (MCP) server: AI assistants such as Claude, ChatGPT, Gemini or Cursor can connect and get accurate, up-to-date answers about our services and work, straight from the source, with no guessing and no scraping.

https://www.tovutech.com/mcp

Add it to your assistant as a custom connector (remote MCP server). It is public and read-only: five tools, no personal data, rate-limited and protected by our security layer.

  • MCP and tool use. The same pattern we build for clients: agents that call your systems to check, book and report.
  • Contextualisation. Answers grounded in your own documents and data (retrieval-augmented generation), not the model's memory.
  • Compute where it fits. Edge functions for speed, cloud GPUs for training, private hosting when data must stay in Kenya.
  • Trust layer built in. Attack detection, automatic blocking and alerts protect every agent we deploy.
Live client website

Just launched: pbagco.com

House lights down, website up. We designed, built and launched the new agentic website for the PBAG Consortium (Peers Best Art Group), Limuru's theatre, comedy and talent movement since 2011.

Data-driven PR and campaign intelligence

Our dedicated, data-focused AI uncovers the insights hidden in each client's brand analytics and turns them into PR and digital campaign strategies that deliver measurable results.

  1. 1

    Collect

    Google Analytics 4, Search Console, and insights from Meta, X, TikTok, LinkedIn and YouTube, plus public conversations through official APIs and listening tools.

  2. 2

    Understand

    Open models such as Gemma and Llama, run privately, classify sentiment and emotion, cluster topics, segment audiences and spot anomalies.

  3. 3

    Strategise

    PR angles, content calendars, channel and budget mix, best posting times, and early warnings when a story starts to turn.

  4. 4

    Measure

    Reach, engagement rate, click-through, conversions, cost per result and sentiment shift, in a monthly Looker Studio dashboard.

See it work on sample data

40 social posts about a fictional grocery delivery brand, Soko Fresh. Or paste your own posts, one per line. The analysis runs entirely in your browser.

Sentiment
    What people talk about
    Engagement by hour of day

    Insights and recommended actions

    • Press "Analyse sample posts" to see the insights.

    This demo uses a lightweight keyword model so it can run in your browser. In client work, Gemma and Llama models do this at scale with much higher accuracy, using only data collected lawfully through official platform APIs or the client's own analytics, in line with Kenya's Data Protection Act.

    Cybersecurity and data privacy

    The trust layer for AI. End-to-end support for innovation-led organisations that handle sensitive data across platforms, teams and technologies. We combine privacy by design, data governance and hands-on security engineering, grounded in Kenya's Data Protection Act and its links to commercial practice, technology contracts and global standards such as the GDPR.

    Glowing shield of hexagons protecting a padlock within a wider network

    Cyber resilience and awareness

    Tailored assessments of your security posture, staff awareness programmes, and a roadmap aligned to standards such as ISO/IEC 27001 and the NIST Cybersecurity Framework.

    Data privacy compliance

    Data protection impact assessments, records of processing, consent and retention design, cross-border transfer reviews, and help for foreign organisations localising their data frameworks for Kenya.

    Technical and operational support

    Hands-on implementation of security controls, incident-response planning and drills, technical office hours, and our AI Security Quick Scan.

    AI, data and intellectual property

    Protect the value in your models, datasets and code: access control, data and model provenance, licensing hygiene and secure training environments.

    Smart security and connectivity

    AI-enhanced CCTV that recognises the events that matter and pushes alerts straight to your phone, IoT sensor integration, and Starlink deployment and configuration that keeps remote sites securely online.

    For formal legal opinions we work alongside partner advocates.

    Where we work

    Choose a sector to see the problems we take on and the work we've already done there.

    How a project runs

    Four stages, the same on every project. You see working output at the end of each one, so you can change direction early.

    In the lab now

    Work in progress in Garissa County and with local development partners. The animations show how each system works; live maps will be published as each one matures.

    Ongoing

    Mapping land degradation with GeoAI

    Machine learning on Sentinel-2 and Landsat time series in Google Earth Engine flags where vegetation and soil condition are declining across Garissa County, and tracks recovery after reseeding or Prosopis clearance.

    • Earth Engine
    • Sentinel-2
    • Random forest
    • Python
    Ongoing

    Water assets, aquifers and borehole safety

    Sensors are expensive, so we use GIS, Earth Engine and Python instead: mapping boreholes, pans and dams, modelling likely aquifer zones from geology, terrain and rainfall, and monitoring boreholes for water-safety risks.

    • QGIS
    • Earth Engine
    • CHIRPS
    • Python
    Ongoing

    Rangeland health and livestock routes

    With local development partners, we combine vegetation indices, water points and community mapping to monitor rangeland condition and map seasonal livestock migration routes, so plans follow how herders actually move.

    • NDVI
    • Participatory GIS
    • QGIS

    Success stories

    Real systems for real problems, each with its own story page. Many were built by our founder for the County Government of Garissa before TovuTech existed; every card says where a project came from and how far along it is.

    Follow new work at github.com/jmsmuigai.

    How our technology works together

    Every TovuTech solution follows the same five layers, from sensing the world to acting on it safely. Choose a layer to see the tools and a real example.

    Colourful orbits of learning nodes around a central core

    TovuTech Academy

    Skills for the AI economy, so no one is left behind.

    Practical courses that integrate readily available AI tools into everyday work. Learn remotely or in person, join live webinars, and study with an AI tutor that generates lessons and quizzes on demand.

    • AI foundations for leaders
    • Practical AI tools at work
    • GeoAI with Earth Engine and QGIS
    • Python for data analysis
    • Cybersecurity and AI security
    • Data protection essentials

    The AI pulse, live

    We track the AI industry daily so our clients don't have to. The latest headlines, new model releases and trending open models are on our AI News page, updated automatically.

    Loading the latest AI headlines…

    Open AI News

    Check where you stand

    Two quick self-assessments. Nothing you choose here leaves your browser.

    Your result appears here as you answer.

    The team

    A founding team that combines AI, data and security, with specialist experts brought into each project as the work requires.

    James M. Mburu

    Founder and Chief Executive Officer

    James is an applied AI specialist, machine learning researcher and GeoAI practitioner with years of public-sector technology leadership. At the Department of Computing, Jomo Kenyatta University of Agriculture and Technology (JKUAT), his work focuses on applied machine learning, geospatial intelligence and multi-sensor data fusion.

    He is PRINCE2 and ITIL certified, and founded TovuTech in October 2026 to connect AI, machine learning and geospatial intelligence to practical problems across Africa.

    • Applied AI and ML
    • GeoAI and Earth observation
    • Digital government
    • JKUAT researcher
    • PRINCE2
    • ITIL

    Kelvin Njiriri

    Co-founder and Director, Cybersecurity and IoT

    Kelvin is a cybersecurity and Internet of Things specialist who secures the networks, devices and data that organisations depend on. He designs surveillance systems that go beyond recording: by integrating AI video analytics into CCTV, they recognise events that matter, such as intrusions or unusual activity, and push timely alerts to clients.

    He also deploys and fine-tunes Starlink connectivity for remote and off-grid sites, so field teams, clinics and offices stay securely online. At TovuTech he leads cybersecurity, smart security and connectivity work.

    • Cybersecurity
    • Network security
    • IoT integration
    • AI video analytics
    • Starlink deployment

    Responsible by design

    The commitments we make on every engagement.

    Data protection
    We work to Kenya's Data Protection Act, 2019. Personal data is collected only when needed, encrypted in transit and at rest, and deleted when the purpose ends.
    Data sovereignty
    You decide where data lives. We write it down before we build, and offer hybrid and on-premise options for sensitive records.
    Secrets stay secret
    API keys and credentials live in encrypted server-side stores, never in web pages, notebooks or code repositories.
    Honest models
    We test on data the model has not seen, report where it fails, and keep a person in the loop for decisions that affect people.
    Local context
    Interfaces in English, Kiswahili, Af-Soomaali and other Kenyan languages, designed for low bandwidth and shared devices.
    You own the result
    Code, models, notebooks and documentation are handed over, so you are never locked in to us.

    Questions people ask

    What does TovuTech do?

    We design, build and run AI systems end to end: we define the workflow with you, design the data and machine learning pipeline, deploy the solution, and publish the analysis as a notebook or web page, with security and privacy built in.

    You're a new company. Why should we trust you?

    TovuTech Limited was incorporated in October 2026. We don't claim a long client list. What we can show is the work: systems in daily use, ongoing projects in the field, and repositories you can inspect before you talk to us.

    Do you offer cybersecurity and data protection support?

    Yes. We assess your security posture, support compliance with Kenya's Data Protection Act and the GDPR, implement security controls, plan incident response and run an AI Security Quick Scan. For formal legal opinions we work alongside partner advocates.

    Can you train AI models for us?

    Yes. We prepare your data and train, fine-tune, evaluate and host models on your behalf, in cloud regions you approve or on your own infrastructure. You own the resulting model.

    Do you train teams?

    Yes. TovuTech Academy runs courses in AI foundations, practical AI tools, GeoAI, Python for data, cybersecurity and data protection, remotely, in person and through webinars.

    What is a Smart Bot, and what does it cost to run?

    A Smart Bot is an AI agent added to your existing website. It answers from your own content, captures leads, books meetings and hands complex cases to your team. Running costs depend on traffic and are usually modest; we estimate them for you during the audit.

    Can you build and host our website?

    Yes. We design, build, host and deploy fast, secure, agentic websites, with Git-based deployment to a global CDN, TLS, security headers, analytics and an optional AI agent.

    Do we need our own servers to use AI?

    No. With AI as a Service we host and run the models, pipelines and monitoring for you. If your data has to stay on site, we deploy on your hardware or in a hybrid setup.

    How long does a first project take?

    A discovery workshop takes a few days. A first working pilot typically takes six to ten weeks, depending on how ready the data is.

    Tell us about your idea

    Describe the problem in a few sentences. We reply within two working days with questions, a suggested first step, or both.

    Prefer to think it through first? can help you shape the brief.

    This opens your email app with the message filled in. We use your details only to reply.