Bespoke Applications Labs — applications, AI operators and infrastructure
Where we work
Primary service area, in order: Limpopo, Gauteng, North West, Mpumalanga. We work remotely across all of South Africa, and on site where the work needs it. These are coverage areas, not office locations.
Limpopo (LP) — primary service area
Provincial capital: Polokwane.
- Polokwane
- Tzaneen
- Mokopane
- Thohoyandou
- Musina
- Lephalale
- Bela-Bela
- Modimolle
- Phalaborwa
- Makhado
- Giyani
- Burgersfort
- Louis Trichardt
- Hoedspruit
- Groblersdal
Gauteng (GP) — primary service area
Provincial capital: Johannesburg.
- Johannesburg
- Pretoria
- Sandton
- Midrand
- Centurion
- Soweto
- Randburg
- Roodepoort
- Kempton Park
- Benoni
- Boksburg
- Germiston
- Alberton
- Krugersdorp
- Vereeniging
- Vanderbijlpark
- Springs
- Brakpan
North West (NW) — primary service area
Provincial capital: Mahikeng.
- Rustenburg
- Klerksdorp
- Potchefstroom
- Mahikeng
- Brits
- Lichtenburg
- Vryburg
- Hartbeespoort
- Zeerust
- Orkney
- Stilfontein
- Wolmaransstad
Mpumalanga (MP) — primary service area
Provincial capital: Mbombela.
- Mbombela
- eMalahleni
- Middelburg
- Secunda
- Ermelo
- Standerton
- Barberton
- White River
- Sabie
- Bethal
- Lydenburg
- Komatipoort
- eMkhondo
- Hazyview
- Malelane
Rest of South Africa
KwaZulu-Natal (KZN)
- Durban
- Pietermaritzburg
- Richards Bay
- Newcastle
- Ballito
- Umhlanga
- Ladysmith
- Empangeni
Western Cape (WC)
- Cape Town
- Stellenbosch
- Paarl
- George
- Somerset West
- Worcester
- Knysna
- Mossel Bay
Eastern Cape (EC)
- Gqeberha
- East London
- Mthatha
- Bhisho
- King William's Town
- Queenstown
- Jeffreys Bay
Free State (FS)
- Bloemfontein
- Welkom
- Sasolburg
- Bethlehem
- Kroonstad
- Parys
Northern Cape (NC)
- Kimberley
- Upington
- Springbok
- Kuruman
- De Aar
Applications
The product your operation actually runs on.
Web platforms, mobile products and internal tools, shaped around how the work really moves.
Web platforms, mobile products and internal tools shaped around how the work really moves — not around a template. Built to be opened every morning by the people who do the job, and to still make sense in year three.
- Product design and front-end
- Application and API engineering
- Internal tools, admin and reporting
- Deployment, monitoring and handover
Most start as one painful workflow and grow into the system that replaces it. We ship the first useful slice early, then widen it.
Agent Systems
Operators that do the work, with a human on the brake.
AI operators that act inside the business rather than chatting beside it.
AI that acts inside your business rather than chatting beside it: reading the inbox, scoring the lead, drafting the reply, updating the record. Every consequential step keeps an explicit approval point.
- Triage, scoring and routing pipelines
- Drafting and review agent pairs
- Tool use against your real systems
- Approval queues and audit trails
Where a person repeats the same judgement dozens of times a week. We automate the repetition and keep the judgement.
Media Engine
Creative output as a production line, not a scramble.
Creative production as a pipeline, with brand rules living in the system.
Generation, review and publishing wired into one pipeline so campaign work stops depending on who happens to be free. Brand rules live in the system, not in a shared folder nobody opens.
- Generative image and video pipelines
- Brand and template systems
- Review, approval and versioning
- Scheduling and distribution
Teams already producing volume by hand, where the bottleneck is coordination rather than ideas.
Town Square
The infrastructure under local commerce.
Marketplaces, storefronts and the unglamorous machinery beneath them.
Marketplaces, storefronts and the unglamorous machinery beneath them — listings, payments, fulfilment, disputes. Built for merchants who need it to work on a bad-signal phone on a busy Saturday.
- Multi-merchant marketplace platforms
- Payments, payouts and reconciliation
- Catalogue, inventory and fulfilment
- Merchant onboarding and support tools
An operator with real supply and real demand who is currently holding both together with spreadsheets and WhatsApp.
Networks
Connectivity that does not need a hero to stay up.
Connectivity for sites where downtime has a cost, monitored and documented.
Network and infrastructure design for sites where downtime has a cost — offices, schools, venues, multi-branch operations. Monitored continuously, documented properly, recoverable by someone other than us.
- Site survey and network design
- Segmentation, access and policy
- Monitoring and automated response
- Documentation and on-site handover
Multi-site operations where nobody can currently say, with confidence, what is connected to what.
Academy
Capability that stays after we leave.
Practical capability for the people who will run what we build.
Practical training for the people who will run what we build. Not an AI-hype seminar — the specific tools, habits and judgement your team needs to keep the system improving without us.
- Role-specific AI working sessions
- Digital literacy foundations
- Internal enablement and documentation
- Ongoing office hours
Paired with a build, so the team learns on the system they are about to own rather than on a generic example.
The Studio
The Studio
Bespoke Applications Labs engineers the applications, AI operators and infrastructure that turn complex businesses into clear, capable systems. We work across the layers that decide whether a business actually moves — product, intelligence, creative, commerce, infrastructure and capability — and we work on the parts that are hard.
- Small senior team, one system at a time
- Fixed scope per phase, reviewed at each gate
- Your systems, your data, your accounts
- Handover documentation as a deliverable
We do not start with a template, a trend, or an AI feature in search of a problem — and we do not leave behind a system that needs us to stay alive.
Method
Method
Three principles, in order. We map the real hand-offs before choosing any technology. We design applications, agents and infrastructure as one system rather than a pile of tools. And we finish only when your team can run, explain and extend it without us on the call.
- 01 — Start where the work breaks
- 02 — Build the operating advantage
- 03 — Leave a system that holds
Most failed systems were built to a brief that described the symptom. A pile of tools compounds cost; a system compounds leverage. And a system that only survives with its builder attached is not an asset, it is a dependency.
Disciplines
Disciplines
Deliberately narrow. We would rather be genuinely excellent across four disciplines that compose into one system than adequate across twenty that do not.
- /product — Next.js, React, TypeScript
- /agents — Claude, tool use, approval queues
- /infra — Vercel, Postgres, networks
- /media — generative pipelines, brand systems
Because the disciplines cross over. The same team that builds the application wires the operator into it and puts it on infrastructure it understands. Nothing is thrown over a wall.
Operating Rhythm
Operating Rhythm
No discovery theatre and no open-ended retainer. Each phase has a fixed scope, a named output and a gate you can stop at.
- 01 map — where the work actually breaks
- 02 frame — scope, approval points, sequence
- 03 build — application, operators, infrastructure
- 04 hand over — documentation, training, ownership
A systems map, then a build plan, then a running system, then a system your team owns. You can walk away after any of them holding something useful.
Bring us the hard part.
Bring us the hard part.
Tell us what keeps breaking. The first session is a working one: forty minutes on your actual process, ending with an honest read on whether this is worth building — including when the answer is no.
- The process that keeps failing
- Who touches it, and where it stalls
- What it costs you when it does
- Anything already built that has to survive
hello@bespokelabs.dev — or use the form. We reply from a person, not a sequence.