For hall committees
See when your hall is actually used, and fill the quiet hours.
For facilities officers
A maintenance schedule that follows real wear and real bookings.
For municipal leaders and funders
Evidence of how each public asset performs, and a business case when it could do more.
Weekly use by hall
Illustrative scores that change on a timer. Not live data.
Where it applies
Illustrative scenarios for a pilot. They are not client case studies, and they contain no results.
The hall nobody books
Occupancy is low and keys sit with one person. The system suggests open slots for training, meetings and markets.We would measure: hours used per week.
Explore →The hall that costs too much to run
Lights and heating run all day for a few hours of use. Energy follows the booking plan instead.We would measure: energy use per booked hour.
Explore →The hall booked twice
Clashes on busy days cause disputes. Booking rules keep groups apart.We would measure: clashes per month.
Explore →The repair that waits for a budget
Photos and sensors show what is failing first, so limited money fixes the right thing.We would measure: time from fault to repair.
Explore →The hall that could earn income
A draft multi-purpose business case shows what weddings, classes and clinics could bring in.We would measure: income per available hour.
Explore →The shared hall
A clinic, a school and a church share one building. One calendar shows everyone's real needs.We would measure: groups served per month.
Explore →What Hall Optimiser does
Reads occupancy, energy and condition signals; drafts a booking plan, a maintenance schedule and a multi-purpose business case; and explains each draft in plain language through the solution bot.
What it does not do
It does not set fees, cancel bookings, declare a building unsafe or replace the caretaker and committee. A named person decides every time.
AI dependency. This solution requires current AI models to operate. Without them there is no usage analysis, no booking plan, no photo assessment and no business case. See AI helpers.
The problem
Across South Africa, news reports describe community halls that are vandalised, locked, unscheduled or left to decay while the public pays for them. The hall is meant to serve residents, yet no one can see how it is used or what it needs. See Research.
- Hall built or upgraded
- Little scheduling
- Few bookings
- Income lost
- Maintenance deferred
- Vandalism and decay
- Measured and maintained
Underused
Open hours are going unbooked. Suggest uses and open the diary.
Costly to run
Energy use is high for the hours booked. Match power to use.
Needs repair
Condition is slipping. Schedule maintenance before it becomes a closure.
The tool produces leads and draft plans for people to verify. It never declares a facility unsafe or a booking cancelled.
How it works
- Sense. Occupancy, energy and condition sensors record how the hall is used and how it is holding up.
- Analyse. Usage patterns show quiet hours, peak days and wasted energy.
- Look. Photos are analysed to assess the condition of floors, roofs, walls and fittings.
- Schedule. A booking plan and a maintenance schedule are suggested together, so repairs fall in quiet hours.
- Build the case. A draft multi-purpose business case shows how the hall could serve more people and earn income.
- Verify. The caretaker or committee checks the suggestions on site.
- Decide. A named person approves, modifies, rejects, asks for a site check or escalates.
AI and people
Each model has one clear job. Model names are as given in the project brief.
GPT-6 Luna
Analyses how the facility is used.
GPT-6 Sol
Generates booking optimisation and maintenance schedules.
GPT-6 Astra
Produces multi-purpose business cases.
Image 2
Assesses facility condition from photos.
Edge and IoT
NVIDIA edge inference on site, with an IoT platform such as Azure IoT or AWS IoT.
Human reviewer (mandatory)
A hall committee chair or municipal facilities officer makes every final decision. Never skipped.
Model candidates by provider (under evaluation, not endorsements)
- OpenAI GPT family: the project-brief models (GPT-6 Luna, Sol, Astra) for usage analysis, scheduling and business cases.
- Google Cloud: Gemini on Vertex AI as a long-context alternative for pattern analysis across months of hall data.
- Image models: the brief's Image 2, Gemini image understanding, or Amazon Rekognition Custom Labels for condition photos.
- NVIDIA: Jetson edge modules with Triton Inference Server or Metropolis for on-site counting, so images need not leave the building.
- AWS and Azure: Amazon Bedrock, SageMaker and IoT Core, or Azure IoT Hub, for secure ingest and model hosting.
Model identifiers must be checked against each provider's current documentation before any production build. All AI output on this site is simulated.
Indicative running cost
Planning assumption only, to be validated against current provider pricing at pilot: roughly R20 to R80 per facility per month for the cloud AI tier, excluding sensors, connectivity and installation.
This solution requires current AI models to operate; the models listed (or their current equivalents) are essential to the production system.
3D model
Drag to rotate, scroll or pinch to zoom. Move the slider to change how much the hall is booked.
How this model works
The model shows one small community hall, simplified for teaching. The slider sets how much of the hall's available time is booked. As bookings rise, more people appear inside, the door sensor pulses faster, the energy bar climbs and the calendar fills. The round figure is the caretaker, who walks to the calendar when the hall is quiet, the door when it is busy and the meter when use is heavy. The assembly view pulls the parts apart so you can see what each one does.
- Community hall: the building. Dots inside are people present.
- Occupancy sensor: a door frame that counts people. Its ring pulses faster as use rises.
- Energy meter: the bar rises with power use. Green is fine, amber is high, red is very high.
- Booking calendar: filled squares are booked slots in a four-week view.
- Round figure: the caretaker. They go to the calendar when the hall is quiet, the door when it is busy, and the meter when use is heavy.
- Spinning cage: the AI helper in the background.
Simplified for teaching. It is not an engineering drawing.
Why this matters
A note from the founder
I started Kitso AI Solutions because I kept seeing community halls that the public paid for sitting empty, locked or falling apart, and nobody could see how they were used or what they needed. I believe AI should help ordinary people make better decisions, not make the decisions for them. I want every hall we work with to become a place where residents meet, learn, trade and get services, with a caretaker and a committee who can see the evidence and decide for themselves.
Who it is for
Residents
A hall that is open, clean and scheduled for the things the ward needs: training, clinics, markets, youth programmes and celebrations.
Hall committees and caretakers
A plain-language view of when the hall is used, what the next repair is and how to fill quiet hours, without needing a facilities background.
Municipal facilities officers
Evidence to prioritise limited maintenance budgets, and a record that shows why a decision was made.
Ward councillors and community development officers
A fair picture of how each public asset performs, so requests for upgrades can be backed with evidence.
Programme funders
A draft business case and a measured baseline before money is spent, and a way to see whether it worked.
Local groups and small businesses
Easier access to open slots for classes, services and trading.
Why we think it can work
Halls are already run by committees, caretakers and facilities officers who approve bookings and repairs. Hall Optimiser fits that practice: it gathers evidence and drafts plans, and the same people approve, modify, reject, ask for a site check or escalate. Because a human always decides, accountability does not move to software.
Honest scope. These are design intentions for a prototype, not measured results. The demo uses simulated data. We would measure hours used per week, energy per booked hour, clashes per month and time from fault to repair, in a pilot, with the hall committee's agreement.
Research and evidence
Plain-English context: halls that communities paid for are often unscheduled, unrepaired or vandalised, and the people responsible rarely have evidence to act on. Five news reports on community halls and public facilities in South Africa. Open each to read the original. The ten Articles build on them.
GroundUp via allAfrica, October 2025. A hall in Verulam with broken windows and no scheduling of its use.
Statement published on Polity, 2022. Vandalised halls and community facilities, and the income and opportunities lost.
Daily Dispatch, April 2024. A municipality says repairing vandalised facilities draws on service-delivery funds.
Southern Mail, Cape Town. A hall upgraded for about R450 000 was vandalised, with residents and the City disagreeing on who should look after it.
Ridge Times, The Citizen. A township hall decays while a renovation budget is redirected.
Articles
Ten short, educational pieces on the problems community halls face and how an AI plus bot approach could help. Each lists model candidates under evaluation, not endorsements. These are not press releases and contain no results.
Coverage
Target areas with current temperature from Open-Meteo, fetched in your browser and refreshed every 5 minutes. Temperature matters because heating and cooling drive hall energy use. Stage labels are the company's own designations, not from the feed.
About Kitso AI Solutions
Hall Optimiser, a product of Kitso AI Solutions.
Kitso AI Solutions builds AI tools that help South African communities get more from the public buildings they already own. Hall Optimiser is our first product.
Company details
| Company | Kitso AI Solutions |
|---|---|
| Registration no. | 2026/694305/07 |
| Registered with SARS | 8 September 2026 (income tax). Provisional tax from the 2028 tax period, per the SARS notice below. |
| Address | 1234 Anahita Avenue, Centurion, Gauteng, 0157 |
Investment
Hall Optimiser is a product of Kitso AI Solutions. The startup currently has no external funding and requires investment to begin the new project and move from this working prototype into production. Microsoft Azure for Startups, AWS Activate / AWS Startups, NVIDIA Inception, and peer AI and cloud startup programmes are the natural partners for this stage because Hall Optimiser is built on the same stack those programmes exist to accelerate: modern AI models for scoring, advisory generation, long-context pattern analysis and vision verification, plus cloud infrastructure for secure data ingest and human-in-the-loop decision support. Grant or programme support from these partners directly funds the next phase — real backend, real integrations, and pilot deployments with municipalities or enterprise buyers — while the human-always-decides architecture already demonstrated on this site keeps governance and liability clear. Without that support the project remains a high-fidelity prototype; with it, Kitso AI Solutions can deliver a production system that depends on current AI models and that those programmes are designed to help scale.
Our team
General enquiries
info@kitsoaisolutions.co.zaQuotes (RFQ)
rfq@kitsoaisolutions.co.zaDevelopers
developers@kitsoaisolutions.co.zaAI solutions
aisolutions@kitsoaisolutions.co.zaKitso Tshepo Joseph Mofokeng
Director
I started Kitso AI Solutions because I kept seeing community halls that the public paid for sitting empty, locked or falling apart, and nobody could see how they were used or what they needed. I believe AI should help ordinary people make better decisions, not make the decisions for them. I want every hall we work with to become a place where residents meet, learn, trade and get services, with a caretaker and a committee who can see the evidence and decide for themselves. When a building the community already owns starts working for the community again, that is the impact I want to make.
Email: kitso@kitsoaisolutions.co.za
Phone: 072 538 4492
Business documents
SARS Notice of Registration for KITSO AI SOLUTIONS, issued 8 September 2026.
Contact
Choose how you are getting in touch. Submitting opens your email app with the message ready to send to Kitso AI Solutions.
Thank you. Your email app should now be open with your message ready. Press send there to complete it. If nothing opened, email info@kitsoaisolutions.co.za directly.
Nothing is stored on this site.
Legal, privacy and notice
Who we are
Kitso AI Solutions, registration number 2026/694305/07, 1234 Anahita Avenue, Centurion, Gauteng, 0157. Email: info@kitsoaisolutions.co.za.
What this website is
A prototype and demonstration of Hall Optimiser. Everything inside the demo is simulated. The only live data is the temperature on the Coverage page, fetched from Open-Meteo. Coverage stage labels are company designations.
Personal information and the Contact form
The form asks for your name, email, optional phone, optional organisation and your message, only so we can reply to your enquiry. This site is static: it has no server and stores nothing. Pressing send opens your own email app, and the message travels by email from you to us. We keep correspondence only as long as needed to answer you and any follow-up, and delete it on request.
Cookies and local storage
This site sets no cookies and writes nothing to local storage.
Third-party services on this page
- Google Fonts, for typefaces: privacy policy.
- Three.js, loaded from cdnjs (Cloudflare), for the 3D model: privacy policy.
- Open-Meteo, for live temperature: terms and privacy.
Your browser contacts these services directly, so they may see your IP address.
Your rights
Under South Africa's Protection of Personal Information Act (POPIA) you may ask to see, correct or delete personal information we hold about you. Email info@kitsoaisolutions.co.za. You may also complain to the Information Regulator.
Disclaimer
Predictions and AI outputs in the demo are simulated. They are not professional, engineering, legal or financial advice. In any real use, a human always decides.