Service Description
Effective: 15 September 2026
Last reviewed: 15 September 2026
This document is the Service Description for the Quality Audits Topic, and it forms part of the ApexIQ Companion Service Terms. Where it and those terms conflict about Quality Audits, this document governs. It describes the service precisely, including what it will not do. Where a statement here limits the service, that limit is part of the agreement.
1. What the service is
You send us an artefact. We read it, assess it against a defined internal standard for its type, and return a written report.
The service answers 1 question: is this artefact good enough to send?
It does not answer whether a deal is healthy, whether a strategy is right, or whether a claim is true. Those are different questions and this service does not address them.
2. Who it is for
Sales, Sales Engineering, and Marketing teams who produce written work which goes to a customer, a buyer, an evaluator, or a board, and who want it checked before it goes.
The service is offered to organisations, and to individuals acting for business purposes. It is not offered to consumers.
3. The 6 paths
Every artefact is audited on 1 of 6 paths, chosen by its shape rather than its subject.
| Path | The artefact is | Examples |
|---|---|---|
| Document | A written deliverable in prose | Proposals, statements of work, specifications, runbooks, playbooks, whitepapers, responses to tenders and security questionnaires |
| Presentation | A slide deck | Customer decks, executive briefings, board packs, quarterly reviews, workshops |
| Copy | Public-facing copy | Landing pages, product and pricing pages, campaigns, email campaigns, articles, press releases, adverts |
| Register | A register, log, tracker, or checklist | Risk registers, requirement lists, objection logs, readiness checklists, evidence indexes, content calendars |
| Model | A quantified model | Cost models, pricing models, sizing tools, capacity plans, attribution models |
| Diagram | A diagram | Architecture, integration, process, and data flow diagrams, threat models |
Where an artefact blends 2 shapes, we audit it as its primary shape, name the secondary shape in the report, and check the secondary shape without letting it affect the score.
Where you are unsure which path an artefact takes, tell us what it is and we will say.
4. What each path examines
Broadly, and in the vocabulary of the report:
| Path | What we assess |
|---|---|
| Document | Whether the opening states the point, whether it fits its audience, whether claims are supported and sourced, whether recommendations are specific, and whether the structure survives skimming |
| Presentation | Decision clarity, commercial consequence, problem definition, evidence, storyline, and slide-level quality |
| Copy | Specificity, usefulness, proof, clarity, simplicity, and whether it holds up when skimmed |
| Register | Completeness against its source, ownership, currency, evidence behind each entry, and prioritisation |
| Model | Whether assumptions are disclosed, whether the baseline is sourced, sensitivity, auditability, and whether units and horizons are stated |
| Diagram | Whether it answers 1 question, notation and key, boundaries, flow correctness, and legibility |
Every audit also assesses voice, integrity, evidence, vocabulary, and sentence discipline, and looks at colour and contrast where the artefact carries either.
5. What comes back
One report, delivered as a PDF, named to our file naming standard.
The report contains:
- an executive brief;
- an overall verdict;
- a scorecard, giving a score against each area assessed and an overall score with its band;
- the result of the blocking issue check;
- the issues which most limit the artefact, ordered by impact, each with a severity, a heading naming the problem, why it matters for this reader and this purpose, and the specific fix;
- assessments particular to the path, such as a claim and evidence audit, a row audit, an assumption audit, or an element audit; and
- recommended priority actions.
Every audit is full. There is no light, standard, or abbreviated version, and there is no depth to choose. Every report uses every section, so 2 reports are directly comparable whatever the artefact.
6. How the audit is produced
Stated plainly, because you are entitled to know how your artefact is assessed.
The assessment is drafted with AI assistance and finished by a person. We send the artefact and your brief to Anthropic's Claude API, which produces a first draft of the assessment against our internal standard. A reviewer then checks every finding, corrects what is wrong, removes what does not hold, and takes responsibility for the report.
No report reaches you without that review. The judgement in a report is ours.
Our supplier does not use commercial API inputs or outputs to train its models, and deletes them within 30 days. The Privacy Notice names the supplier, its purpose, and its location. The Confidentiality and Materials document sets out what this means for your material.
7. Score, band, severity, and blocking issues
A report states a score, and a band which the score falls into. The band carries a plain disposition: whether the artefact is ready to send, needs revision before sending, or needs rebuilding.
Each issue carries a severity. A severity is always written as a word. It is never carried by colour alone, so the report still works in print, in mono, and for a reader who does not distinguish colours.
Some defects block. Where 1 of these is present the report says so, and the blocked status overrides the band whatever the score. There is no partial status: an artefact is either blocked or it is not.
We publish the fact that these mechanisms exist. We do not publish the internal standard behind them: what a band is worth, what triggers a blocking issue, or how the areas are weighted. That is our confidential method, and a report states findings rather than the method which found them.
8. Companion inputs
Four artefact types need 1 input beyond the artefact itself, because the assessment is impossible without it.
| The artefact is | We need |
|---|---|
| A response to an issued request | The issuer's question list, the mandated format, and confirmation of who owns each control |
| A register | The source list the register was built from |
| A model | The live workbook, not an exported or flattened copy |
| A diagram | The question the diagram answers, in 1 sentence |
Where a companion input is missing we report the affected area as unverified, and say what we asked for and did not receive. We never infer it, and we never hold its absence against the artefact as though it were a defect of the artefact.
9. What we accept, and how it reaches us
This Topic takes material by email, and by no other route today. The intake form produces a brief file, takes you to Stripe to pay, and opens a pre-addressed email, and you attach the artefact and send it yourself. Where we add another channel, such as a secure upload, this section names it before it is used, and Section 6 of the Confidentiality and Materials document (How it is held) applies to it unchanged.
We accept the common Office and document formats, and images for diagrams. Where we cannot open a file we will tell you within 1 working day and ask for another copy.
For a model, send the live workbook. A PDF of a model cannot be audited for the things a model audit exists to test.
Do not send links to systems which require an account, a login, or a network connection to your environment. We audit artefacts, not live systems.
10. Turnaround
We offer 2 speeds. The Service Terms define both, and the pricing page states what each costs.
| Speed | We deliver |
|---|---|
| Standard | By 18:00 UK time on the 2nd working day after receipt |
| Same day | By 18:00 UK time on the day of receipt, where a complete request arrives before 12:00 UK time, Monday to Friday |
Where we cannot meet a same-day request, we will tell you by 14:00 UK time on the day of receipt and agree an alternative with you. The fee is not refunded, as Section 8 of the Service Terms (Fees and payment) sets out.
Receipt means the complete artefact, the complete brief, and any companion input. The clock starts on the complete package and not on the first email.
11. What we delete, and when
We delete your artefact, your brief, and any companion input when the report is sent.
This is a designed property of the service. It is what we offer, it is not variable by negotiation, and it means we hold your commercial documents for days rather than years.
The consequence is real and you should plan for it. After delivery we cannot re-read the artefact, compare a later draft against it, or answer a question which needs the artefact in front of us.
Where you want the revised artefact assessed, that is a new audit at the published price. Send it to us again, because we no longer hold the original.
We keep the report we wrote, and a record of the engagement. The Privacy Notice states for how long.
12. What the service does not do
A report is our professional opinion of how well an artefact is made. It is not any of the following, and must not be read as though it were.
We do not verify facts. We assess whether a claim is supported, sourced, and bounded. We do not check whether it is true. An artefact can score well and still say something which is wrong.
We do not give legal, regulatory, or tax advice. A report is not a legal review, and nothing in it should be treated as advice on any legal or regulatory question.
We do not validate anything technical. We do not test an architecture, verify a configuration, reproduce a benchmark, or confirm that a described system behaves as described.
We do not check arithmetic or pricing. In a model audit we assess whether assumptions are disclosed, sourced, sensitivity-tested, and traceable. Whether the numbers are correct is outside the audit.
We do not audit systems, environments, or running configurations. We audit artefacts.
We do not audit a single email. An email campaign is audited as copy. A single email is not, whatever it records.
We do not rewrite your artefact. We tell you what to change and why. Producing the revised artefact is your work. We do not offer a rewriting, editing, or remediation service at any price.
We do not guarantee an outcome. No report promises that an artefact will win a deal, pass an evaluation, satisfy a buyer, or produce any commercial result.
13. What we never claim
We never state or imply compliance, conformance, certification, accreditation, or legal assurance against any standard, instrument, framework, or regulation, and we do not name one.
Where a report observes that an element relies on colour alone, that is a finding about whether the artefact communicates: it stops working in print, in mono, and on a poor screen. It is true, and it is sufficient. It is not a statement about anyone's legal obligations, and it must not be presented as one.
If you need an assurance opinion, a certification, or a legal review, engage a firm qualified to give one. This service is not a substitute for any of them.
14. Known limits
Stated plainly, because you should know them before you rely on a report.
- A report reflects the artefact and the brief as they were on the day. A thin brief produces a more general report.
- A report will not find every issue in an artefact. It finds the issues which most limit it.
- Scoring involves judgement. Two competent reviewers may score the same artefact differently at the margin, though the band and the blocking result should agree.
- We assess the artefact in front of us. We cannot know what was said in the meeting, what the buyer privately wants, or what your competitor has proposed, except where you tell us in the brief.
15. The intake form
The intake form at companion.apexedgesalesengineering.com/quality-audits/start/ runs entirely in your browser. It sends none of your answers anywhere, except the email address you gave, which goes to Stripe when you choose to pay.
While you fill it in, your answers are held in your own browser's local storage so that your draft survives a closed tab. That copy stays on your device. It is cleared when you open the email, and when you select Clear form, and held again if you download a corrected brief after the email. Nothing you type reaches us until you send the email yourself, and until you do we do not know you have started.
When the brief downloads, the form gives the request a reference. When you choose a speed, it takes you to Stripe's checkout page to pay, carrying 2 things: that reference, and the email address you gave, so Stripe can fill it in. When you have paid, Stripe brings you back to the form with the payment's identifier, and the form writes the reference, the speed, and that identifier into the email. Your other answers, your brief, and your artefact never go to Stripe. Section 7 of the Privacy Notice (Who else sees it) names Stripe as a recipient.
The form sets no cookies, loads nothing from any other website, and contains no analytics or tracking of any kind. Its fonts and its brand mark are built into the file, so opening it makes no request to any other website.
This is a property of this Topic's intake, which is why it is stated here rather than in the Privacy Notice. Section 2 of the Privacy Notice (The website sets no cookies and runs no analytics) covers the website itself.
16. Changes
We may change this document as the service develops. The version which applies to a request is the version published when we accepted that request. The date at the top is the date it last changed.
Questions about this document: legal@apexedgesalesengineering.com.