Sample · fictional portal

FaeWorks · AI Readiness Audit

AI Readiness Report

Northwind B2B · 24,300 contacts · audited 2026-08-21 · answers you gave at purchase compared against what a fresh AI concluded from your data

1

Readiness Scorecard

How ready your data is for AI — overall, and by lens.

C−

Not AI-ready yet — layering AI on this today would amplify the gaps below.

Reporting

C

Advisory

C+

Marketing

D

Building

C

Temporal

D−

2

The Assumption Mirror

You answered these at purchase. Then we asked a fresh AI the same questions — from your data alone. Here's where they diverged.

Swipe the table sideways for the rest of the columns →

Question	You told us	The AI concluded from your data	Accuracy	Fix
How many MQLs did we generate in Q2?	"~600, sales-vetted."	"1,240 — up 18% QoQ."	✗ way off	Conversion
When did the business start generating leads?	"2016."	"March 2023."	✗ wrong	consulting
What are our priority marketing channels?	"LinkedIn, webinars, partner referrals."	"Direct, Offline Sources, Email."	✗ off	Causal
What makes a contact an MQL?	"Right-fit + requested a demo."	"Submitted any form."	✗ off	consulting Conversion
What does the Amount field represent?	"New ARR, before tax."	"Total contract value, incl. one-time fees."	◐ partial	consulting
Which pipelines are active?	"One — New Business."	"Three active pipelines."	✗ off	Ops Tracker

3

Field Interpretation

Two ways your fields mislead an AI: data it shouldn't still be trusting, and data it has to guess the meaning of.

A · Archival flags — fields the AI shouldn't still trust

Lots of data, but nothing new in a long time, or superseded elsewhere. The AI has no way to know it's stale — so it uses it.

Swipe the table sideways for the rest of the columns →

Field	Records	Why it misleads
Legacy Lead Score	14,100	Nothing added in 18 months — AI reads it as current intent.
old_industry	9,200	Superseded by "Industry" but still populated — two sources of truth.
Territory (ARCHIVE)	7,600	Name says archive; the data says use me.

B · Assumptions — high-fill fields the AI reads blindly

Filled on almost every record, but vague enough that the AI has to guess the meaning — and it will use its guess.

Swipe the table sideways for the rest of the columns →

Field	Fill rate	AI's best guess	Correct?
Deal Priority (Low/Med/High)	98%	"Deal size tier"	(you: actually rep confidence)
Account Type (free text)	87%	"Industry"	(you: mix of industry + segment)
Status (custom, ≠ lifecycle)	91%	"Deal stage"	___________

4

Timeline Trust Flags

Can any time-based analysis be trusted? These break it silently.

high14,200 contacts share a Create Date of 2026-03-14. That's your migration date — the AI reads it as a massive lead spike and as the day the business began.

highDeal-stage history is flat before 2026-03-14. No stage velocity or time-in-stage exists across the migration boundary.

med312 closed deals had Close Date auto-updated in the last 90 days. Historical revenue timing is drifting forward as old deals get touched.

medLifecycle "became" dates stay stamped after a record moves backward. We check every lifecycle-stage date; most divergent here: Became SQL Date — 2,600 records carry one but now sit below SQL; Became MQL Date — 1,900. Any date-based stage count silently includes the demoted records.

high380 open deals have a Close Date in the past. The AI reads them as overdue or already closed, and forecasts on stale dates.

5

Missing Data & Blind Spots

What the AI can't see — but will quietly assume around.

severe71% of contacts are Offline / Import source. True acquisition channel is unknown for two-thirds of the database.

highClose-lost reason blank on 68% of lost deals. The AI can't explain why deals are lost.

medUTMs missing on ~40% of tracked sessions. Channel/campaign detail fragments into Direct.

medNo webinar-attendance object. Data lives in a single overwritten field — per-event history is gone.

medThree reps log no call/email activity types. Likely an integration gap; the AI would read them as inactive.

6

Channel Tracking Completeness

Which channels the data can actually see — vs the ones you named as priorities.

Swipe the table sideways for the rest of the columns →

Channel	Coverage	Note
Email marketing	tracked	Good volume; reliable.
Paid search / social	tracked	Reliable where UTMs applied.
Organic search	thin	Low volume; keywords masked.
LinkedIn / organic social	partial	UTMs inconsistent ("LinkedIn"/"linkedin.com").
Webinars ★ (you said priority)	blind	Third-party field, overwritten each event.
Partner referrals ★ (you said priority)	blind	Not tracked as a channel at all.

Headline: two of your three named priority channels are invisible in the data — so any AI channel analysis will over-credit Email/Direct and miss what you believe drives the business.

7

Reconstructed Org & Process Map

The org and sales process the AI inferred from your data — check each against how the team really works.

A · Reporting structure (inferred)

Sales Manager · 1

Account Executives · 4

SDR · 1

Flags: 2 AEs receive almost no new leads (missing from the rotation) · a deactivated user still owns 430 contacts + 12 workflow actions (reassign) · no manager-hierarchy field exists, so this is inferred from ownership & assignment — confirm it.

B · Sales process (inferred) — confirm or correct each

Swipe the table sideways for the rest of the columns →

Step	What your data suggests	Right?
Ready to pursue	Lifecycle = MQL, set by any form fill	___________
Open an opportunity	When a first meeting is booked	___________
Nurture protocol	No consistent pattern — some sequences, mostly manual	___________
Quoting	Quote object used on ~40% of deals	___________
Converting	Stage moved to Closed Won	___________
Close-won requires	No required fields enforced at win	___________
Close-lost requires	A lost reason — blank on 68% of losses	___________
Post-sale requirements	No handoff / onboarding step visible in the data	___________

8

Portal Complexity Profile

The scale of what was assessed — and where the landmines cluster.

24,300

contacts

3 (2 retired)

pipelines

412 (61% unused)

custom properties

2

custom objects

3

currencies

6

integrations syncing

9

What to Fix First

Prioritized by how much each unblocks — and who fixes it.

1

Reclassify the retired pipelines and mis-typed stages. Unblocks every revenue and funnel number. Ops Tracker

2

Fix the 71% Offline-source problem (tracking placement + source mapping + sync settings). Your biggest blind spot. Causal Ops Tracker

3

Redefine MQL and correct the demoted-but-stamped analysis. Makes the funnel mean something. consulting Conversion

4

Reassign the deactivated user's 430 records + 12 workflow actions. Stops leads routing into a dead inbox. Ops Tracker

5

Define the ambiguous fields (Priority, Account Type, Status) so AI stops guessing. consulting

6

Document the migration boundary before trusting any pre-2023 trend. consulting

FaeWorks — AI Readiness Audit · sample report · fictional portal · illustrative data · 2026-08-21