Got questions? Let's clear them up.

BusyWork Dispatch is a service that lets teams send real development work to a professional human team directly from the AI assistant they already use, such as ChatGPT, Claude or Copilot. You describe what you need in your AI chat, the request lands in Dispatch, and the BusyWork team scopes it, caps the spend, builds it, tests it and delivers working software back to you. There is no subscription; you buy a Blueprint package and top up Build Credits as needed.

No. BusyWork Dispatch is not a subscription. Customers buy a Blueprint package upfront, which includes onboarding and a starting balance of Build Credits, then top up credits whenever they need more work done. There is no monthly platform fee, no retainer and no pressure to keep spending. Build Credits expire 12 months after purchase, so you are buying capacity to use across the year rather than committing to ongoing payments.

BusyWork Dispatch is designed to work wherever your team already uses AI. It connects through MCP-enabled AI assistants, including ChatGPT, Claude, Microsoft Copilot and coding tools such as Cursor, so requests can be sent without copying and pasting between systems. If your assistant does not support a direct connection yet, work can still be sent through other intake paths, and new connections are added as the product evolves.

You can send BusyWork Dispatch any digital work your AI can start but not finish: website changes, new websites, internal apps, customer portals, dashboards, forms, workflow automations, CRM and reporting integrations, AI agents and assistants, bug fixes, and technical review of AI-generated code or prototypes. Typical requests fall into three groups: building websites, apps and internal tools; automating repetitive admin work; and creating AI tools that understand your business and follow your processes.

A human-led professional development team does the work. BusyWork's team scopes each request, agrees a spend cap, builds, tests and delivers it. AI is used to support delivery where it makes work faster or cheaper, but humans remain responsible for the outcome, and every piece of work is reviewed by a person before it is handed back. That is the difference from pure AI tools: you get tested, integrated software your team can rely on.

A Business Blueprint is a practical map of how your business operates today and how it can be improved with AI, covering your people, processes and tools. It gives you a clear, prioritised list of exactly where AI fits: what to automate, fix or build, and in what order. Every BusyWork Dispatch engagement starts with a Blueprint because it gives the delivery team the context to build properly. Your Blueprint is stored in BusyWorkBook and kept updated as work gets done.

During onboarding, the BusyWork team maps your agreed scope: a workflow, role, department, product or the whole business, depending on your package. We document how work actually happens across your people, processes and tools, then produce your Blueprint: a prioritised plan of what to automate, fix or build. Everything is stored in BusyWorkBook and kept updated over time. Onboarding is mandatory because it gives the delivery team the context needed to build useful work safely.

Because context is what makes build work useful. Mapping first means the delivery team understands your workflows, tools, data and constraints before requests arrive. That leads to better scoping, fewer repeated explanations, automations that fit how your team actually works, and AI agents that follow your real processes. It also produces your Blueprint, a prioritised plan of what to build, and seeds BusyWorkBook so every future request starts from shared understanding rather than a blank page.

BusyWorkBook is BusyWork Dispatch's business memory layer. It stores your business context: workflows, roles, tools, preferences, Blueprints, past requests and decisions. Every new request starts from that shared understanding, which means fewer repeated explanations, faster and more accurate scoping, and build work that fits your systems from the start. BusyWorkBook is created during onboarding and kept updated as work is delivered, so the service gets smarter about your business over time.

The Future Operating Model Blueprint is a strategic engagement for leadership teams, priced from $50,000 AUD. Before scaling AI implementation, we map how work happens today, design your future-state operating model and deliver a practical 12-month roadmap for AI-enabled change. You finish with a clear view of your best opportunities, an agent and automation roadmap, a governance and approval model, a capability and change-readiness plan, and a recommended Dispatch pathway to deliver it all.

Build Credits are the currency you use to get work done through BusyWork Dispatch. They pay for approved delivery work, including websites, apps, automations, AI agents, internal tools, fixes, testing, review and implementation. Every Blueprint package includes a starting credit balance, for example $3,000 with the Starter Blueprint, and you can top up at any time. Before any work starts, each request gets a capped Approval Band, so you always approve the maximum spend first. Credits expire 12 months after purchase.

BusyWork Dispatch pricing starts at $5,000 AUD for the Starter Blueprint, which includes $3,000 in Build Credits. The Product Launch Blueprint starts from $8,500 with $4,000 in credits, the Growth Blueprint is $15,000 with $12,000 in credits, and the Scale Blueprint starts from $35,000 with $30,000 in credits plus ongoing advisory support included. There is no subscription. After your initial package, you simply top up Build Credits as needed, with bonus credits on larger top-ups.

Human-in-the-loop delivery is generally estimated around the time the team spends. Autonomous AI and tool work is calculated from the underlying usage cost plus BusyWork's delivery markup. Whichever mix a request needs, the important part stays the same: every request is estimated and assigned a capped Approval Band before any credits are spent, so you always know the maximum cost upfront and never receive a surprise bill for AI usage.

No. Every BusyWork Dispatch customer starts with a Blueprint package because onboarding is what makes the service work. The Blueprint gives the delivery team the context needed to scope and ship useful work safely, and it seeds BusyWorkBook, the memory layer that makes every future request smarter. Skipping onboarding would mean slower scoping, more back and forth and lower quality work, so Build Credits are only available with, or after, a Blueprint package.

Yes. You can top up Build Credits at any time without changing your package. Larger top-ups earn bonus credits: top-ups of $500 or more receive 5% bonus credits, $2,000 or more receive 10%, and $5,000 or more receive 15%. Credits expire 12 months after purchase. If you already know there is a lot to build, a larger top-up gives you noticeably better value for every dollar you spend.

Yes. Build Credits expire 12 months after purchase. This structure lets you buy build capacity upfront without committing to a monthly subscription or retainer. In practice, most teams use their credits well within the year, because the Blueprint gives them a prioritised list of work worth doing. If you are unsure how quickly you will use credits, start with a smaller package and top up as the work proves its value.

Yes. Build Credits can be used by anyone in your business. To help manage spend, credits can also be allocated by workflow, department, product or business area, and approval rules can differ by team. Combined with Approval Bands, this gives leaders a simple way to let many people send requests through Dispatch while keeping full visibility and control over what is spent and where it goes.

Ongoing AI Strategy & Advisory Support is a 12-month support package priced at $3,400 + GST a year. It includes twelve 60-minute strategy sessions across the year, a regular check-in and action review to help you decide what to build, improve or automate next, and 12 months of WhatsApp build support. It is included in the Scale Blueprint and available as an add-on with any other BusyWork Dispatch package.

Yes. BusyWork Dispatch runs a Give 1,000, Get 1,000 referral programme. When someone you refer buys their first credit pack of $5,000 or more, excluding GST, they receive 1,000 bonus Build Credits and you receive 1,000 credits too. Referral rewards apply to first-time customers only. It is the easiest way to reduce the cost of your next build if you know another business that should be using Dispatch.

Approval Bands are spend caps agreed before any work begins. Every request is placed into one of four bands: Small, up to $100 in Build Credits, for quick fixes and copy changes; Medium, up to $500, for small features and simple integrations; Large, up to $1,000, for dashboards and multi-step automations; and Custom, over $1,000, for complex builds and product work. You approve the maximum spend before work starts, and if a request needs more, BusyWork pauses and asks first.

Before any work starts, the BusyWork team reviews your request and assigns it an Approval Band: Small (up to $100), Medium (up to $500), Large (up to $1,000) or Custom (over $1,000 in Build Credits). You approve the maximum spend for that band before credits are used. If the work turns out to need more than the approved cap, BusyWork pauses and asks before continuing. If BusyWork keeps working inside an approved cap and overruns internally, BusyWork absorbs that cost, not you.

BusyWork pauses the work and asks before continuing. You are never charged beyond the Approval Band you approved. You can then approve a higher cap, reduce the scope, or split the request into smaller stages. If BusyWork chooses to keep working within an approved cap and the job overruns internally, BusyWork absorbs the overrun. This is how Dispatch removes surprise invoices from custom development work.

Yes. You choose how much control you want. Teams can manually review every request, or set rules such as auto-approving all Small requests, auto-approving Small and Medium requests, or always requiring manual approval for Large and Custom work. This means quick fixes can flow through Dispatch without waiting on sign-off, while bigger spends always come back to the right person in your business for review.

No, and that is deliberate. Dispatch work blends human implementation, AI-assisted development, testing, debugging and tool usage, so exact upfront quotes would slow everything down and can be misleading. Instead, every request gets a capped Approval Band before work begins: you approve a maximum Build Credit spend, work is estimated before credits are spent, and BusyWork pauses and asks if anything needs to exceed the approved cap.

Big requests are broken into smaller, useful pieces. The BusyWork team can split the work into staged scopes, estimate each stage separately, or turn the whole thing into a build roadmap inside your Blueprint. Each stage then gets its own Approval Band, so spend stays capped and visible throughout. The aim is always to make the next valuable step clear rather than asking you to approve one large, risky project.

Wherever it is most useful, depending on the request. Smaller changes, such as website fixes and automation tweaks, are shipped directly into your own systems, sites and tools. Larger builds are handed over through the Dispatch app with a review step, so you can check the work before it goes live. Every request gets a delivery summary in Dispatch, and the outcome is recorded in BusyWorkBook so future work builds on it.

You do. Everything BusyWork Dispatch builds for you, including code, websites, automations, AI agents and internal tools, is owned by your business once it is delivered. Full intellectual property ownership transfers to you, so you are never locked in: you can host the work yourself, hand it to another developer, or keep building on it with Dispatch. BusyWork retains no rights over your delivered work.

Yes. Maintenance for delivered work is handled through Build Credits like any other request, with small fixes typically falling into the Small or Medium Approval Bands. You can also request security and technical reviews of existing systems, including AI-generated code your own team has produced. Larger Blueprint packages include faster turnaround priority and more ongoing support, and your Blueprint is kept updated as maintenance work gets done.

Raw, unreviewed AI output often is. Independent testing has repeatedly found that a large share of AI-generated code ships with a known vulnerability, and that iterating on it with more AI tends to add problems rather than remove them. That is the risk with pure vibe coding, where code goes from prompt to production without anyone checking it. It is not how we work. We use AI to move fast, then put deliberate guardrails around what it produces. The output of that process is not the same thing as raw AI code, and it should not be judged as if it were.

We do not rely on the AI getting it right. We assume it will get some things wrong, and we build the system so those mistakes cannot reach you. That means three things: we deliberately reduce how much code needs securing in the first place, we lean on audited specialist services for the dangerous parts, and we test continuously rather than hoping a single review caught everything. Speed comes from the AI. Safety comes from the process wrapped around it.

No, and that is on purpose. A person reading every line by hand is slow, does not scale, and is genuinely poor at catching the repetitive vulnerability classes that AI code tends to fail on. Machines are far better at that. So we point automated and AI-driven security testing at the code continuously, and we reserve human judgement for the two things people are best at: the architecture, and the small number of critical paths where money and personal data live. It is not less oversight. It is oversight aimed where it makes a difference.

It means security checking runs at the same speed the code is written, not as a one-off before launch. Static analysis for code-level flaws, dependency scanning for the libraries, and checks that actively try to break the running application. Because AI can introduce a new weakness in a single change, point-in-time testing is not enough on its own. Continuous testing catches a problem shortly after it appears, instead of months later when a customer finds it.

We do not write our own versions of the risky parts. Payments go through Stripe, so raw card details never touch the software we build and the compliance burden stays at its simplest level. Logins and identity go through dedicated providers like Clerk. Hosting and secrets sit on managed platforms. Each of these is audited far more heavily than any individual firm could manage in-house, and every one of them is a category of bug we cannot write because we did not write that code.

Leaked keys are the most common failure in AI-built apps, so we treat them as a first-order concern. Secrets are kept out of the code and out of version history, stored in proper environment configuration, and anything that has been exposed is rotated. This is basic hygiene that pure vibe coding routinely skips, and it prevents one of the most frequent and most damaging mistakes outright.

This is a real and growing attack. Roughly one in five AI-generated code samples references a package that does not exist, and attackers register those predictable fake names with malicious code inside. We defend against it by verifying that every dependency is real and maintained, pinning versions, and locking the dependency list so nothing unexpected can slip in during a build. Nothing gets installed just because an AI suggested it.

Every change runs through an automated pipeline before it can reach you, so a regression gets caught early rather than in production. We also keep changes small and deliberate, because large sweeping AI edits are exactly where new holes appear. The result is that we keep the speed of AI development without the fragility that usually comes with it.

On the parts that matter most, it is often more secure, for a few reasons. We build on the same audited payment, identity and hosting infrastructure the large firms use. We deliberately keep the surface area small, which means fewer features, fewer endpoints and fewer ways in to defend. And we test continuously rather than at the pace a human team can review. A smaller, well-architected, constantly tested application has less to go wrong than a large, sprawling one that was reviewed once.

The opposite, in practice. Attackers hunt at scale for the well-known failure patterns of unreviewed vibe coded apps: exposed databases, missing authorisation checks, hardcoded keys. Because we design those failure patterns out from the start, the easy ways in are closed. A smaller, tightly scoped, properly configured application is a smaller and less rewarding target than a big surface full of the exact mistakes attackers already scan for.

For enterprise builds, it does not have to leave your systems at all. We can build directly inside your own environment, your own cloud tenant, so your data stays entirely within your security perimeter and never flows out to a shared platform. That is a stronger position than standard multi-tenant SaaS, where your data lives on infrastructure shared with other customers and outside your control. You keep your own encryption, your own network policy, and full visibility, while we do the building. For regulated industries, that difference is often the whole decision.

A human-led team is. AI supports the work, but people remain responsible for scoping, architecture and the final result. Context about your business, tools and past decisions is held in BusyWorkBook, so security-relevant details are not re-explained or lost between requests. You always have a person accountable for what ships, not just a model.

Ongoing. Software is only secure relative to how it changes and how threats evolve, so a one-off audit ages quickly. Maintenance and security reviews are part of the model: larger Blueprint packages include more ongoing support, and reviews can be added to any package through Build Credits. The intent is to keep what we build safe over its life, not just on launch day.

An AI Loop is a repeatable workflow where AI helps move work from request to result. Every loop answers six questions: what the goal is, what triggers the work, what context AI needs, what AI should do, where a human reviews, and where the output goes. Examples include sales follow-up, proposal, reporting and onboarding loops. BusyWork Dispatch builds AI Loops for teams, and runs a free 60-minute AI Loops workshop for leaders scaling AI across their organisation.

Book a short founder call with Ben. Because Dispatch is a new way to get work done, the first conversation is about fit: we look at your team, tools and goals, then recommend whether your first Blueprint should map a workflow, role, department, product or your whole business. From there you choose a Blueprint package, complete onboarding, and start sending requests from your AI assistant with Build Credits ready to spend.