Custom ChatGPT or Claude Project Build for AU Small Business

The right Custom GPT or Claude Project turns your team’s collective knowledge into a 24/7 reference. We build the version your team will actually use – not a research demo.

What a custom build actually does

A Custom GPT (on ChatGPT) or a Claude Project (on Claude) is a private version of the assistant that already knows your business. You load it with your real material, your SOPs, product specs, pricing, brand voice guide, common customer questions, and you write custom instructions that tell it how to behave. From then on, anyone on the team opens it and asks a question in plain English, and gets an answer drawn from your documents in your tone, instead of a generic web answer or a re-typed prompt.

The difference from “just using ChatGPT” is that the knowledge and the instructions are baked in. Your VA does not need to paste in the pricing sheet every time. Your new hire does not need to know the perfect prompt. The build does the remembering, so the output is consistent no matter who is driving.

In 2026 these tools are good enough that a well-built Project reliably handles the structured, repetitive 80% of knowledge work: drafting a reply that matches a past example, looking up the right SOP step, rewriting copy to your style, summarising a long document, turning rough notes into a clean brief. They are still assistants, not autopilots, so a person reviews anything that goes to a customer or commits the business to something.

What it looks like in a real AU business

The same build behaves very differently depending on who uses it:

• The support-drafting assistant. An ecommerce VA pastes in a customer email. The Project, loaded with your returns policy, shipping rules and tone guide, drafts a reply that is already on-brand and policy-correct. The VA reads it, tweaks a line, sends. A 6-minute reply becomes a 90-second one, and a new VA sounds like your most experienced one from day one.
• The internal SOP brain. A 12-person services firm loads 30 SOPs into a Claude Project. Instead of interrupting the founder with “how do we onboard a client again”, the team asks the Project and gets the current step-by-step, with the document it came from. The founder stops being the help desk.
• The proposal and quote drafter. A trades or fit-out business loads its standard scopes, rates and past winning proposals. A salesperson describes the job in a paragraph and gets a structured draft proposal in the company format, ready to check and price.
• The customer-facing product helper. A business with a deep product range loads its spec sheets and FAQs into a Custom GPT and lets customers ask it questions, with clear guard-rails on what it will and won’t answer.

How the engagement runs

It is a fixed-scope project, typically $1,500 to $3,500 depending on how much source material there is, run over 2 to 4 weeks. It starts with a 90-minute discovery and scoping workshop where we map who will use it, for what, and what “good” looks like. Then we curate and ingest your source documents (your SOPs, product docs, FAQs from Notion, Google Drive or PDFs), write and tune the custom instructions to your business, and build out test prompts and edge-case coverage so it behaves under pressure, not just in the demo.

You get a 60-minute training session for your team and VA, 30 days of post-launch tuning and support, and a maintenance handover document so you can keep it current yourself. What we need from you is the source material and a couple of hours of a knowledgeable person’s time in discovery; the build is only as good as the documents you can give it, which is why a business with no written SOPs should document first.

Custom GPT or Claude Project, and where it breaks

Which platform comes down to what your team already pays for. Custom GPTs need ChatGPT Plus or above; Claude Projects need Claude Pro or above. If you have both, we will recommend based on the use case in scoping. Neither is meaningfully “smarter” for most SME knowledge tasks; the deciding factor is usually your existing subscription and where your team already works.

The honest limits matter. A standard build uses static knowledge, your documents as they were on build day, so it does not know about a customer’s live order or a price you changed yesterday unless you update the documents or pay for a proper API integration (real build work, separate scope). It can still occasionally state something confidently that is not in its documents, which is exactly why anything customer-facing or money-related is reviewed by a person before it ships. And it is only as current as you keep it: a Project loaded once and forgotten drifts out of date, so the handover document covers how to refresh it.

A few honest questions people ask

Can you connect it to our database? Read-only API integrations are possible but add scope and cost. Most builds use static knowledge (SOPs, docs), which covers around 80% of real use cases without touching your database.

What about data privacy? ChatGPT Team and Enterprise, and Claude Team and Enterprise, both come with business data-privacy guarantees. By default we do not put customer personal information into source documents unless you explicitly approve it, in line with the Australian Privacy Principles.

What if our use case changes after launch? 30 days of post-launch tuning is included. Beyond that it is a retainer or hourly updates, and the handover document means many clients self-maintain.

Who this is for, and who should wait

This is a strong fit for a business with 20 or more SOPs or product docs to consolidate, a VA who keeps answering the same internal questions, or a support team fielding the same product enquiries on repeat. In all three cases the knowledge already exists; it is just trapped in scattered documents or in one person’s head, and a Custom GPT or Claude Project turns it into something the whole team can query.

It is not the right first step if your SOPs are not written down yet. A build is only as good as the documents it is fed, so a business without documented processes should document first (our SOP template pack is the place to start) and build once there is real material to load. It is also not the answer if your core need is real-time access to live customer data; that is custom API work, a different and larger project, not a static-knowledge build.

What “done” looks like

Every DotVA AI services engagement is fixed-scope, fixed-price, and fixed-timeline. No retainers without a defined deliverable. No “discovery phases” that turn into 6-month projects. The scoping call is free; the written quote that follows is what you sign off on.

Why we say no

We turn down roughly 30% of applications. Usually because the team isn’t ready, the goal isn’t clear, or there’s a simpler path that doesn’t need us. We’d rather tell you that on the discovery call than take a brief that won’t deliver. The shortest path to outcome isn’t always us.

Apply below

Two minutes. We read every application within 1 business day. If we’re a good fit, we book the scoping call. If we’re not, we say so – and usually suggest who is.