Blueprint diagram: Scope, Build and Embed leading to an AI-first business

Curating Brilliance

We Design AI‑First Businesses

Production agentic systems, working demos, and teams trained to run them — for enterprise, industry and government.

Not a slide factory. Two founders and a hand-picked bench who scope, ship and embed. If it doesn’t work in production, it doesn’t count.

02 — Approach

How We Work

Strategy-firm rigour, an engineer’s bias to ship — and no junior pyramid.

We design and ship: strategy from ex-BCG/Bain consultants, systems from engineers who put them into production, and training that makes them stick. Every engagement is scoped, priced and delivered by the founders you meet on day one.

01. Innovation & Curiosity

Unobvious routes first — new tools, new angles, working prototypes over long decks.

02. Integrity & Trustworthiness

Honest about what AI can and can’t do. Results over hype.

03. Quality & User-Centricity

Built for the people who use it: practical, intuitive, finished properly.

03 — Working Terms

Clear Before We Start

The commitments that sit on every engagement — no fine print, no junior pyramid.

Fixed prices, agreed upfront

Scope and price before we start. No hourly drift, no surprise invoices.

Production systems, not demos

Integrated into your stack, measured on outcomes, ready for real users.

Yours to keep

Everything we build is yours — code, docs, models, and the know-how to run them.

04 — Selected Work

Impact In Action

Client names withheld by design where required. Open builds link to source.

AI.design

AI for a leading SA founder community

Problem
Members needed faster answers from a growing knowledge base; operators spent too long preparing briefings.
Approach
Retrieval-augmented community assistant, member health-scoring engine, and production data pipelines.
Outcome
Preparation time halved; capability productised into an ongoing AI practice for founders and operating teams.
TECH.design

Data platform for a livelihoods programme

Problem
SME growth and job-creation outcomes were hard to see in real time across a national programme.
Approach
Production ETL pipeline feeding interactive dashboards for programme operators and funders.
Outcome
Live visibility into SME growth and job creation across South Africa.
STRATEGY.design

Market sizing: Middle East & South Africa

Problem
Retail and private-equity clients needed defensible entry and growth decisions across regions.
Approach
Market sizing, entry analysis and growth strategy with implementation-ready roadmaps.
Outcome
Board-ready recommendations that moved straight into execution planning.
PRODUCT.design

Hardware for sustainable agriculture

Problem
SprootTech needed low-cost tools so more African growers could learn controlled farming skills.
Approach
Engineered a low-cost LED grow-lighting rig and cloning gel formulation for field use.
Outcome
Practical hardware and materials ready for an agritech startup training programme.
BUILD / LIVE SYSTEMS

SmokeSignal

Problem
Wildfire risk in South Africa needed timely, public-data alerts — not closed commercial feeds alone.
Approach
Live fire tracking on public satellite data; open engineering on GitHub.
Outcome
Working alert map and codebase others can inspect. View on GitHub →
BUILD / PRIVATE AI

Flowga

Problem
Most vision-AI apps send camera data to the cloud — a non-starter when privacy is the product.
Approach
Local-first AI yoga instructor: webcam pose analysis with on-device vision models.
Outcome
No data leaves the machine — a live demo of private, self-hosted AI. View on GitHub →

Also in the wild: SeeMyCity (municipal finance transparency) · 25+ open builds on GitHub →

Facing something similar?

A conversation costs nothing and usually saves months.

05 — Who We Are

Who You Work With

A technical founder and a commercial founder — strategy-firm rigour, operator pragmatism, and a builder’s bias to ship.

Keshav Mohabir, Co-Founder — Technical

Keshav Mohabir

Co-Founder — Technical
  • UCT
  • BCG
  • Bain
  • Stanford ML

Mechatronics engineer (UCT, Honours Distinction) turned strategy consultant turned AI engineer. Stanford Online ML under Andrew Ng (98%). Complex problems at BCG and Bain; co-founded an on-demand grocery startup; steered Bolt Food Cape Town through its COVID grocery pivot. Leads AI.design — agentic systems, data platforms and capability programs.

Shulka Mohabir, Co-Founder

Shulka Mohabir

Co-Founder
  • Citi
  • McKinsey NGWL ’23
  • R100M+
  • 127 jobs

Finance and business development leader. At Citi she managed 21 client financing programs across Sub-Saharan Africa unlocking R100M+ in capital. McKinsey Next Generation Women Leaders ’23. Ran a R6M SME programme that created 127 jobs; founded ventures in sustainable fashion and regenerative agriculture. Runs delivery and the commercial side.

Careers

Join us

Short bench of consultants, engineers, designers and filmmakers — hired for judgment, not headcount.

07 — Contact

Facing an Intricate Challenge?

Goes straight to the founders — we reply within one business day.