All Posts
Services

What an AI Agent Assessment Includes (and Why the Fee Is Credited)

Compare a free AI audit, a $2,500 credited assessment, and a production agent build: what Phi delivers in about one week and how teams should prepare.

6 min read

If you are comparing “AI audits,” workshops, and build quotes, the labels blur. Phi Digital publishes three distinct entry points on phi.llc/services/ai-agents: a free AI audit, a $2,500 AI Assessment (credited against a build), and production builds. This post explains what the paid assessment includes, why the fee is credited, and how it differs from the free audit and from jumping straight into a single-agent project.

The three entry points (do not mix them up)

Free AI Audit — $0

Self-serve at phi.llc/audit. You answer questions about operations; an AI-generated report emails your top automation opportunities and a rough ROI estimate. No obligation.

Use it when: you want a fast, low-friction signal before spending money.

Do not expect: a human systems review, architecture, or a fixed-scope build quote.

AI Assessment — $2,500 (~1 week)

A human-led, async review of workflows, data, and systems. You buy it online; delivery is about one week. Outputs include prioritised agent opportunities, architecture direction, and a fixed-scope quote. The fee is fully credited against a build if you proceed.

Singapore companies may pay S$3,500 by PayNow to Phi Intelligence (email hi@phi.llc)—a payment path for the same assessment offer, not a different product brand for the engagement.

Build — Single Agent from $5,000 / Suite from $12,000

Production software for one workflow or a coordinated multi-agent system, then optional managed retainer from $2,000/month. Builds are billed 50% upfront and 50% on delivery; you own the code built for you. Model/API usage bills to your provider accounts.

What the assessment week actually covers

Expect a structured pass over four questions:

  1. Where does repetitive work live? Tickets, inboxes, documents, dashboards, CRM chores, internal search.
  2. What systems hold the truth? APIs, databases, auth boundaries, rate limits, who can write.
  3. What is the action risk? Read/draft vs preview-then-apply vs autonomous writes (agent vs chatbot).
  4. What would ROI look like under conservative shares? Hours × fully loaded cost × automation share, minus build and usage—see ROI guide and the ROI calculator (sources checked 2026-10-25).

You should leave with prioritised opportunities and a quote you can accept or decline—not a vague “AI roadmap” PDF.

Why the fee is credited

Uncredited discovery workshops create misaligned incentives: vendors prolong “strategy,” buyers stall. Crediting $2,500 against the build means:

  • Phi is paid to do real scoping work up front
  • You are not punished twice if you continue
  • The quote has to survive contact with your systems—because it becomes the build contract

If you stop after the assessment, you still keep the analysis and quote. If you continue, the assessment cash applies to the project.

How assessment connects to the full delivery process

Published process on services:

  1. AI Assessment — ~1 week
  2. Architecture — 1 week
  3. Build & Test — 2–4 weeks (weekly demos)
  4. Deploy & Monitor — 1 week (monitoring, alerting, 30-day warranty on single-agent builds; suite warranty 60 days)

Assessment is step one, not a substitute for architecture or evals.

When to skip straight to Website Agent

If the only job is answering visitors from your public pages, you may not need a $2,500 assessment. The Website Agent is a packaged product: $990 setup, then $299/month (first payment $1,289), 1,000 conversations/month, quality test before launch, one-line install. Read A Website AI Chat Agent for Small Businesses.

Assessment shines when workflows span internal systems, approvals, and custom tools.

What Phi brings that a generic RFP response does not

Phi builds and runs its own products with AI agent fleets. Measured 2026-10-24: Dayze 3,568 merged PRs (60%+ from agent-named branches); CFA Mock Exam 124 merged PRs (290 of 412 main commits with an AI agent co-author); Coincess 30 merged PRs. See fleet lessons and the case studies. That experience informs assessment judgment—especially around CI gates, preview flows, and evals—without pretending your metrics will match ours.

How to prepare so the week is useful

Send (even roughly):

  • Top 3 painful workflows and weekly volume
  • Systems of record and whether APIs exist
  • Compliance constraints (PII, payments, healthcare, etc.)
  • Who can approve autonomous actions
  • Any failed chatbot/automation attempts and why they stalled

Optional: run hours through the ROI calculator so the assessment starts from shared assumptions.

FAQ

Is the free audit required before the assessment?

No. Useful, not mandatory.

Is the $2,500 refundable if I do not build?

It is a paid assessment deliverable. The credit applies if you proceed to a build—it is not framed as a free option after delivery. Confirm current terms on services and terms of engagement.

Can the assessment recommend “do not build yet”?

Yes. A good assessment sometimes says the data model, permissions, or process definition is not ready—and points you to cleanup work or a Website Agent instead.

Who should attend from our side?

Someone who owns the workflow and someone who can authorize system access. Async delivery means fewer meetings, not zero accountability.

Next step

Buy the assessment from AI agent services, start with the free audit if you want a zero-cost pass first, or contact with your workflow list. If you already know you only need site chat, skip ahead to the Website Agent packaging and pricing.

Sample agenda (async, not a week of workshops)

Day 1–2: intake questionnaire, access list, workflow samples (redacted tickets, screenshots, SOP links).
Day 3–4: systems map, risk classification, draft opportunity ranking with rough ROI assumptions.
Day 5–7: architecture sketch, open questions, fixed-scope quote (Single Agent vs Suite vs “not yet”), and a walkthrough recording or written brief.

You should be able to answer: Which workflow first? What tools? What must stay human? What does success look like in 30 days after deploy?

What “fixed-scope quote” means

A fixed-scope quote names the workflow, integrations, eval approach, warranty window, and price—not an open-ended staff augmentation bag. Published floors still apply (Single Agent from $5,000; Suite from $12,000; retainer from $2,000/month). Your quote may be higher once scope is real; starting prices are anchors, not coupons.

Compare that honesty to chatbot marketplace listings that hide integration work. If your real need is public FAQ only, the assessment may recommend the Website Agent and stop—saving you from buying the wrong thing.

Relating assessment to ROI homework

Bring a draft of hours × cost × share from the ROI post. The assessment will challenge optimistic shares and missing usage lines. Crediting the fee only helps if the eventual build is the right build; the worksheet keeps both sides honest.

If you want a zero-cost first pass, run the free audit the same day you book the assessment so the human review starts with an AI-generated opportunity list to confirm or reject.

Topics

ai agent assessmentai automation auditai agent consulting

Related insights

Get AI insights delivered to your inbox

Weekly perspectives on AI agents, automation, and building systems that work in production.

No spam. Unsubscribe anytime.