Consulting · Business Strategy · AI Systems

Strategy from someone who builds the thing.

Kevin Hokoana advises companies on strategy, architecture, and artificial intelligence — and then goes and ships production systems in the same domains. Hokoana Global is the practice built around that: advice you can act on, from an operator, not an observer.

Start a conversation See the Asher Suite →

The distinction

Anyone can hand you a deck about artificial intelligence. Very few have architected the data layer beneath it, written the inference loop, engineered its cost down, and then had to answer for the result in production.

That gap is the entire value proposition. Recommendations here come with a working reference implementation behind them, because they were learned by building one.

Measured, not asserted

1,653

trades matched

Strategy engine verified trade-for-trade against a reference implementation across three charts — zero mismatches on entry, exit, quantity, or P&L.

266 → 57

candidates pruned

A fitted strategy book cut by walk-forward out-of-sample validation, independence gating, and cost haircuts. Rigor that subtracts.

12×

faster backtests

Sixteen to eighteen seconds down to one and a half, via a content-hashed result cache.

venues ingested

Candle data pulled, normalised, and stored from five trading venues into one first-party store.

locales shipped

The multi-tenant help desk's marketing surface, localised with a documented fallback chain.

20/20

isolation proven

Tenant isolation in the SaaS product verified by an automated row-level-security test, not asserted.

Engineering and validation results. Nothing on this page is a trading-performance claim.

IThe practice

Three practices, one operator.

Engagements are scoped tightly and run personally. There is no bench, no hand-off to a junior team, and no deliverable that ends at the recommendation.

Consulting

Advisory that survives contact with implementation.

Most consulting stops at the recommendation. This does not. Engagements start with the same diligence any firm would run — market, competitive position, unit economics, technical debt — and end with a decision you can act on, written by someone who has had to live with the consequences of decisions like it.

Engagements

Technical and product due diligence
Architecture and build-versus-buy review
Fractional CTO and technical leadership
Vendor, platform, and data-provider selection

Business Strategy

Positioning, product architecture, and the path to revenue.

Strategy work grounded in what can actually be built and sold. Brand architecture, tiering and pricing, go-to-market sequencing, and the unglamorous question of which three things to do this quarter instead of the thirty on the list.

Engagements

Brand and product-family architecture
Pricing, packaging, and entitlement design
Go-to-market sequencing and launch planning
New-venture and market-entry strategy

AI Creation

Agents and AI systems that do real work, in production.

Not pilots and not demos. Agents wired to live data with real tool use, evaluated for correctness and engineered for cost — because an agent that is right but uneconomic is not shipped. Includes the retrieval, caching, and voice layers that turn a model into a product.

Engagements

AI agent design and tool-use architecture
Retrieval, knowledge base, and grounding
Inference cost engineering and caching
Voice interfaces and conversational UX

IIThe proof

The Asher Suite.

A family of production platforms — charting, market data, an AI agent, commerce, and speech infrastructure — architected and built by Kevin Hokoana. It is the reference implementation for everything this practice advises on.

Asher Labs

Live

Trading platform

The flagship platform — terminal, strategy tester, market data, accounts, and billing under one application.

AsherCharts

Live

Consumer charting

A trader-first charting terminal for crypto, equities, and FX — the consumer face of the same engine.

AsherCore

Live

Market data API

The market-data layer, packaged and sold as its own product — data as a service.

AsherSupport

Live

Multi-tenant SaaS

A help desk with tickets, email, scheduling, and per-tenant isolation — sold to other companies as software.

Read how the suite is architected →

IIIHow the work is done

Six operating principles.

These are not values on a wall. They are the rules the engineering actually runs under, and they are why the output holds up.

Own the data

Systems that depend on someone else's feed inherit someone else's outages, licence terms, and pricing power. Where it matters, the data layer gets built and owned.

Verify, never assume

No claim ships without a reproduction behind it. Behaviour is proven against real data in the real environment, not argued from first principles.

Done means deployed

Working on a laptop is not finished. An engagement closes when the thing is running in production and has been checked there.

Build the whole stack

Data, engine, interface, billing, and go-to-market. Owning every layer is what makes the strategy advice worth anything.

Make the numbers worse

The standing rule in the engineering notes: every improvement should make the numbers worse and more trustworthy. Most of what a validation pass produces is deletion.

Report the status, not the story

Built is not live. Dormant is not shipped. Systems get described as what they actually are, including on this website.

IVRange

The method is not sector-specific.

Financial technology is the deepest body of work, not the limit of it. The same approach has produced platforms in healthcare, media, consumer, professional services, and luxury retail.

Healthcare

Caregiver health platform

A medication and vitals tracker for elderly family members living across two countries, with per-patient timezones and a bilingual records portal consolidating scattered reports, labs, and imaging.

Consumer

Bilingual food platform

An installable, mobile-first ordering app for a working chef — Spanish and English from the first screen rather than translated afterwards, with a customer storefront and a kitchen-side dashboard.

Media

News publications

Two publications on a shared engine: feed ingestion, AI-assisted rewriting with attribution rather than republication, and audio narration.

Professional services

Client portal

A membership and administration application for a cross-border services firm, built so a non-technical administrator can run it without an engineer on call.

Luxury retail

Brand & storefront

Brand identity and commerce for a luxury goods label — positioning, storefront, checkout, delivery tracking, and an ambassador programme.

Commerce

Payments & billing

Embedded on-domain checkout, hardened webhook handling with a replay window, usage metering, and entitlement models that gate access to what a customer has actually paid for.

Tell me what you are trying to build.

Engagements begin with a conversation about the actual problem — no discovery invoice, no pitch deck. If it is not a fit, you will be told that quickly.

Start a conversation