Who should not hire Doxia Axis?

Why publish a refusal list?

Because saying no in advance saves both sides time, money, and disappointment.

Most agencies hide their refusal list. They call it "qualification process" and run it inside a sales conversation. The operator finds out they're a bad fit after the discovery call when the proposal arrives at a price that quietly says "please go away." Or worse, the agency takes the work, the engagement stalls at the first ambiguity, and both sides spend three months figuring out what they should have figured out in the intake form.

We publish ours so you can self-screen before you book a call.

Who do we turn down?

Five categories, in order of how often we see them.

1. Pre-PMF teams still hunting for the wedge

If the business hasn't yet found product-market fit, AI is not the highest-leverage problem to solve. The highest-leverage problem is figuring out who actually wants to pay for the thing.

We've sent operators away with a "come back when you have 30 paying customers in the same shape and a repeatable acquisition channel." That's a clean redirect. AI workflows compound on top of a working business. They don't substitute for one.

The full readiness check lives at AI readiness checklist for operators past PMF. If you score below 3 of 8 on that checklist, the audit will tell you the same thing — fix the substrate first.

2. Operators who want a deck for a board slide

If the goal is "we need an AI initiative to show the board" and the actual workflow doesn't matter, we are the wrong agency. We don't produce strategy decks. We don't run multi-month discovery engagements that end in a recommendation roadmap. The deliverable is always something running on your systems.

If the operator's commitment is to the appearance of progress rather than measurable shipped artifacts, the engagement misfires inside the first sprint. The Big-4 firms do this work well; we don't.

3. Operators who can't or won't give direct access

The engagement runs through one operator with direct access to:

• The data — CRM exports, customer support archives, internal documents the workflow needs to ground in
• The people — subject-matter experts on the workflow being absorbed
• The decisions — the operator who can sign off on scope changes, brand voice, deployment risk

Engagements that run through a project manager who runs through a junior staffer fail at the first ambiguity. We've tested this pattern across multiple engagements. The cadence assumes operator-direct access. Without it, the timeline doesn't fit.

If your organization's structure makes operator-direct access impossible (you're a regulated enterprise where data access requires a 60-day compliance review, you're a private-equity portfolio company where the operator can't make decisions without committee approval), we will say so honestly on the intake call and either re-scope to a longer cadence or recommend a different agency.

4. Operators who want zero-error AI

Some operators say "any error is unacceptable" and mean it. AI systems have a non-zero error rate by structural necessity. They are reasoning systems. Reasoning is fallible.

If the operator can't tolerate any error — even with internal-tool human authority gates layered in front of the AI — then AI is the wrong tool for the workflow. Either re-scope to a workflow with bounded failure, accept the error rate, or do not build.

We surface this honestly during scoping. Operators who insist on zero-error AI either accept the calibration after a short conversation or get redirected to alternatives.

5. Operators looking for the cheapest possible vendor

Doxia Axis is priced for the deliverable, not for billable hours. Tier 1 starts at $500 for one specific fix shipped live in seven business days. That's competitive against any agency in the category. But it's not the cheapest possible price — overseas freelance markets and template-driven micro-agencies will quote lower.

If the goal is minimum price for an AI checkbox, we are the wrong vendor. The audit will tell you the same thing — the cheapest possible AI deployment is a worse outcome than no AI deployment, because it produces operational drag without operational lift.

What about edge cases?

Three patterns where the answer is "depends."

Edge case 1 — operator is in a regulated vertical. Healthcare, financial services, legal, education, public sector. Workflows in these verticals can absolutely be good fits, but the engagement timeline expands to accommodate compliance review (HIPAA, SOC 2, attorney-client privilege, FERPA, etc.). We've shipped in regulated verticals; the cadence runs longer than the standard 14-day sprint.

Edge case 2 — operator is outside the US, EU, or India. We work primarily across these three jurisdictions because we know the regulatory surfaces. Operators in other jurisdictions can engage, but we'll be honest about the gaps in our regulatory knowledge for that surface.

Edge case 3 — operator is non-English-language-primary. The audit and engagement run in English. Workflows can absorb non-English content (multilingual customer support, multi-language inbound) but the meta-conversation stays in English. If that's a constraint, surface it on the intake form.

So what's the right move if you might not be a fit?

Three honest paths.

Path 1 — fill out the intake form anyway. /audit intake takes 5 minutes. We screen against the criteria above and send back a yes / no / fix-this-first within 48 hours. The screen costs you nothing and saves both sides time.

Path 2 — read the readiness checklist first. AI readiness checklist for operators past PMF is the same checklist we run during intake. If you score below 5 of 8, fix the substrate first and revisit in two to four weeks.

Path 3 — if you're definitely a wrong fit, here are alternatives. Big-4 advisory if you want a strategy deck. Toptal or Upwork if you want freelance-rate AI work. Glean / Harvey / Microsoft Copilot direct if you want a vendor install with reseller-margin reduced. None of these are wrong; they're just different shapes than what Doxia Axis ships.

What we won't do regardless of fee

Three things, named on the About page and repeated here:

• No deck-led engagements. The deliverable is always a shipped artifact, not a recommendation document.
• No equity in lieu of fees. We don't take operator equity. The fee structure is cash, deliverable-shaped.
• No engagements without direct operator access. This is the load-bearing constraint. Every other refusal traces back to it.

Where to go from here

• The full operator profile: /about.
• The readiness checklist: /answers/ai-readiness-checklist-post-pmf.
• Still want the audit? /audit. The intake form does the screen.
• Or skip the audit if you're sure of the fit: /book — we'll arrive with a draft solution scoped to your business.