What we have built, in enough detail to judge it.

The engagement

Brilliant Perspectives is an established organization with its own history, audience, and revenue. i49 did not found it and does not own it. What i49 did was rebuild its entire technology stack from the ground up and make it AI-native.

This is the Build model at full size: the client keeps the business, and we replace the machine underneath it. The company owns and operates the result today.

What it is

Brilliant Central is the operating system for Brilliant Perspectives. Not a website and not a CRM — the platform the business actually runs on: billing, affiliates, email marketing, community, content, support, finance, and the AI agents that work across all of it, on one governed data plane.

It replaced a stack of bolted-together tools — Zapier, Flight Commerce, Zendesk, ActiveCampaign, ClickFunnels, and PayQuicker — with a single system that owns its own data. Development began in August 2023 and has been continuous since.

Three shipped surfaces

Brilliant Central is the platform and the admin layer. Two member-facing products sit on top of it, both reading and writing through the same core rather than maintaining a second backend of their own.

• The platform — a Next.js modular monolith over Postgres, with 29 production modules, 160+ admin screens, and 74 heartbeat-monitored scheduled jobs
• The website and member portal — public site plus a signed-in portal covering library, discover, journal, inbox, events, music, community Circles, and account
• BrilliantPlus — a native member app built in React Native, live on the iOS App Store and Google Play since June 2026

The AI layer

Eight named agents operate against the same governed data plane the rest of the business uses — among them Ask Brilliant, the member-facing assistant, backed by a help center of more than 300 published articles and a catalog of over 3,100 teachings.

This is the pattern i49 keeps returning to, and the one PromptForge later generalized: agents are only as trustworthy as the data plane and the rules underneath them.

Why it is here

Because it answers the question a product demo cannot: whether this team can carry a system of real size, in production, for years, with somebody else's revenue depending on it every day.

It is also the clearest evidence that we finish. The stack was rebuilt, handed over, and now runs without us.

Who it is for

Brilliant Central is not sold as a product. It is the system of record for one organization, built to its requirements — and the clearest evidence of the scale i49 builds at.

The problem

Firms are putting proprietary knowledge into public models with no control plane. The prompt that encodes how a firm prices a deal, screens a candidate, or reviews a contract is institutional knowledge — and it currently lives in Slack threads, personal notes, and unmanaged pastes into ChatGPT.

Banning the tools does not work; people use them anyway, and the firm loses the upside. What is missing is the layer in between.

The product

PromptForge holds an organization’s prompts and context, routes them to the right model, and enforces policy at the hop. It is not a consumer chatbot and not a replacement for Claude or ChatGPT — those remain the models. PromptForge is the layer that holds the organization’s prompts, context, and rules.

What you get

• A prompt library where prompts are authored, versioned, and shared as organizational assets rather than Slack scraps
• Context management: attach the documents and instructions a model should use, without handing the whole corpus to a vendor to keep
• Many models, one desk — a unified interface over hundreds of models via OpenRouter and direct providers
• Teams and tenants: organization-scoped work, roles, and plan-based access, not a single shared bucket
• Governance that ships — AI governance profiles separating what an agent may say from what it may do, signed policy bundles, and delivery into IDEs and agents over MCP
• Usage analytics and cost visibility, so AI spend stops being a mystery line item

Research and intellectual property

PromptForge is the most research-heavy thing i49 has built. It is accompanied by published work on governed cloud AI as a third path between unmanaged public-model use and the cost of building on-premises, and on delivering curated domain expertise as protected AI context.

That work has produced intellectual property around how organizations govern prompts, context, and AI use.

Status

Live and in market. The product site is mypromptforge.com and the application runs at app.mypromptforge.com. Built and operated by i49 Group.

Who it is for

Organizations whose competitive advantage is encoded in knowledge they cannot afford to leak — professional services firms, and any enterprise that has to show its AI usage is governed.

The problem

Good product. Solid team. Still stuck. Meetings drag, projects stall, and everyone is visibly busy without producing breakthroughs. The instinct is to book an offsite, feel aligned for a fortnight, and return to the same pattern.

The gap is rarely motivation. It is that the company has no repeatable system for how decisions get made, what “done” means, and who is allowed to act without asking.

The system

LeaderForge is 48 short sessions — roughly 15 to 25 minutes each — delivered to phones and laptops so they fit inside a working week rather than replacing one. The curriculum is distilled from Dr. Dionne van Zyl’s doctoral research on high-performing leaders and from the playbooks used to build 27 companies.

Sessions are grouped into four series — Foundation, Productivity, Innovation, and Executive — of twelve sessions each, so a team can start narrow and widen without re-buying the program.

What a team gets

• Core trainings across all four series, with unlimited users on every plan
• Bold-action management, so commitments made in a session survive contact with the week
• Five-minute leader meets at Gold and above
• Learning pathways that sequence the curriculum to the team rather than the catalog
• An AI coach at Platinum, and an AI-infused culture module included on every plan
• Optional add-ons: upload your own content, or run a branded tenant

What LeaderForge claims it changes

These are the outcomes published on the LeaderForge site, stated as the product’s own claims:

• Three to five times faster project delivery, through thin slicing and deep work
• Fifty to eighty percent fewer wasted meetings
• A shared, explicit definition of what “done” means
• Self-managing teams, via the CUE and Buckets models
• A shift from busy work to high-leverage work

Commercials

Silver, Gold, and Platinum plans, sold annually, with unlimited users and the full 48-session library on every tier. The published guarantee is a 30-day refund: if the team is not noticeably better in 30 days, you do not pay.

Status

LeaderForge is the one system on this page we are still incubating. An earlier version is in market today at myleaderforge.com and has paying customers on it.

The next version is built and in testing. It has not been released, and we are not describing it as live until it is. Operated by i49 Group, Inc.

Who it is for

Founders and CEOs of roughly 1–50 person companies, and leadership teams who have had enough of generic training and want a system rather than an event.