Honest answers before the audit

Frequently asked questions.

The questions people ask before booking a Context Audit. Answered plainly, including the objections worth taking seriously.

The main questions.

The things people want answered before committing to a scoping engagement.

The honest objection

Can I not just use Claude Projects myself?

You can. The honest answer is that the product is not the storage. A Claude Project you set up yourself will drift within weeks because nobody has the job of keeping it current, applying a consistent schema, or tuning a skill against real output. The context ages, the structure loosens, and it becomes a folder with an AI label on it.

The value Admin Chief charges for is the discipline, the tuned skill, and the ongoing maintenance: the unglamorous work of inventorying sources, cleaning messy exports, running scheduled freshness sweeps, and checking whether the output is still accurate. That is the part buyers consistently do not do for themselves. If you will reliably do it, you probably do not need this service.

The honest objection

Do I still need this if I already use ChatGPT?

AI tools are powerful, but they start every conversation from zero. They have no memory of your clients, your past decisions, or your way of working, so you re-explain the business each time, and the output is only as good as what you can paste in that moment.

Admin Chief is the memory those tools are missing. We organise your company's context into one structured brain that your tools and our agents read from, so nobody re-explains anything, and the work that comes back is grounded in your real history rather than a generic guess.

Getting started

What is the Context Audit?

The Context Audit is a paid scoping engagement, priced at 500 to 750 pounds, that runs before any brain is built or any skill is stood up. During the audit we inventory your context sources, assess their format and condition, propose a schema and organization rules, identify the single highest-value skill for your situation, and write a fixed build quote with explicit in-scope and out-of-scope items.

The audit exists because the main variable in a brain build is data-loading effort: how many sources, how messy the exports, how much cleanup is required. Without that mapped in advance, any price is a guess that overruns. The audit converts the unknown into a number before you commit to the build. It is the only route to a defensible fixed price, and no build starts without one.

The engine

How do you build and run the brain?

We build and maintain our delivery engine on our own infrastructure. It handles the ingestion pipeline, source organization, retrieval, skill execution, and the scheduled maintenance sweeps that keep the brain current. We run our own knowledge base on it, which is how we know the pattern works in practice rather than only in theory.

The client manages none of it. We operate the infrastructure, assemble the delivery stages, and maintain the brain and skills on your behalf. What that means in practice for a given engagement depends on scope, and we describe it accurately before work begins.

Approval and sending

Do you send anything without my approval?

No. Every output produced by every agent, whether a client brief, a draft quote, an angle pack, a chase draft, or a deal summary, is produced for a human to review. Nothing is sent to a client, customer, prospect, or any third party without an explicit decision from your team. This is a firm operating rule that applies from the first day and does not change as more agents are added.

The approval-first design is not a disclaimer. It is how the skills are built. An output is a candidate artifact, not a sent message.

Pricing

What does it cost?

Three steps, each with its own price agreed before work begins.

Step 1

Context Audit

500 to 750

Paid scoping engagement. Source inventory, schema proposal, skill recommendation, and fixed build quote.

Step 2

Brain Build

2,000 to 4,000

One company brain built from agreed sources, one flagship skill stood up behind an approval gate. Scope capped to what the audit defined.

Step 3

Maintain and Operate

2,000 /mo

A managed-employee retainer: live sync, scheduled freshness sweeps, agent tuning, and one new or upgraded skill per quarter. Ongoing, not open-ended.

All prices are in pounds sterling. Step 2 is tiered by source count and data volume. Step 3 is a flat monthly fee. The audit fixes step 2's price before you commit to it.

More questions.

Other things worth knowing before reaching out.

Which verticals do you work with?

Agencies, coaches and consultants, local service businesses, ecommerce brands, real estate teams, solicitors and conveyancers, and accountants and bookkeepers. Each has a named flagship agent: Scout (Client Context Retriever) for agencies and for solicitors, Echo (Offer Angle Finder) for coaches, Ledger (Quote Prep Assistant) for local service, Vox (Customer Voice Miner) for ecommerce, Atlas (Deal Brief Generator) for real estate, and Badger (Follow-up / Chase) for accountants and bookkeepers. Badger also handles Follow-up / Chase across all verticals. Agencies are the starter vertical where the most detailed work has been done.

What if my context is in a mess?

That is normal and it is exactly what the Context Audit is for. The audit inventories the sources and assesses their condition before any price is set for the build. If cleanup is significant, it is scoped explicitly and either priced into the build or treated as a separate item. You are not quoted a flat price that then blows up on the ingest step.

Do I have to start with a specific agent?

No, but the audit recommends one based on your situation. The single-agent-first approach is intentional: it keeps the scope honest, proves value faster, and avoids building a brain that nobody uses because too many agents were stood up at once with no clear owner for each one.

What sources can be ingested?

This is scoped at the audit and depends on your situation. Common sources include Google Drive documents, Notion workspaces, call transcripts, PDF exports, spreadsheets, and support thread exports. Sources that require live API access, real-time sync, or regulated data handling are discussed in advance and may be out of scope.

Does Admin Chief provide legal, financial, or professional advice?

No. Admin Chief organizes and operates company context for administrative and operational purposes. Skill outputs are drafts and summaries prepared for human review. They are not legal opinions, financial recommendations, tax advice, or any form of professional judgement.

Can I add more agents after the build?

Yes. The build includes one flagship agent. Additional agents from the catalogue are available as paid options once the brain is live and the first agent is producing reliable output. The ongoing maintenance retainer includes one new or upgraded agent per quarter. Extra agents outside that cadence are quoted separately.

What if the brain stops being useful?

The maintenance retainer is renewed based on ongoing output value, not just uptime. If the skill stops producing useful artifacts, that is the right moment to reassess the engagement rather than pay to maintain something that is not working. We track artifact quality as part of the monthly review, not just whether the system is running.

How is the brain different from a regular Notion workspace?

The brain is Notion (or an equivalent structured store) organized to a specific schema with AI-readable formatting, consistent tagging, and organization rules agreed at the audit. The difference is the structure and the maintenance: a standard Notion workspace grows organically and drifts. The brain is shaped to be retrieved accurately by a skill, and it is kept current by scheduled sweeps rather than left to whoever last edited a page.

What this is not.

Honest boundaries are easier to work within than discovered ones.

Not a generic "build you a chatbot" shop. The product is a maintained brain with specific skills, not a conversational UI on top of uploaded documents.
Not a SaaS platform. This is a done-for-you managed service. There is no self-service dashboard, no login portal, and no per-seat pricing.
Not unsupervised. Every skill output requires human review before use. Autonomous sending is not offered.
Not free to try. The Context Audit is a paid engagement. There is no free tier or trial because scoping work takes real time regardless of whether you proceed to a build.
Not a replacement for professional judgement. No output from any skill constitutes legal, financial, tax, HR, or regulated advice.
Paid scoping engagement

Book a Context Audit.

The audit maps your sources, proposes a schema, identifies your highest-value skill, and ends with a fixed build quote. It is how scope gets agreed before anything is built or charged.

Book a Context Audit

After submission we reply with next steps to agree scope and price. Nothing is connected or billed until that is done. Meet the six agents.