Overall capability grade
—
Stricter 67-capability evidence standard. Grade 3 requires a controlled execution and destination receipt.
First real organization audit · evidence-led
Ability Commerce
full-utilization audit
A menu-mounted review of every graph8 capability: what exists, what is actually proven, what is unsafe or missing, and the exact implementation sequence required to reach 5.0.
Capabilities assessed
—
Every visible menu item has its own decision and next action.
Grade 3 or better
—
Controlled end-to-end proof, not configuration alone.
Critical P0 gaps
—
Must close before any scaled launch.
Qualified held meetings
0
The operating north star; current 30-day baseline.
01 · Relationship
Client context and operating relationship
What the program is expected to deliver, what has been approved, where client and graph8 expectations diverge, and which decisions must close next.
Relationship evidence informs the audit; it does not prove execution.
Loading source coverage…
Loading source coverage…
Loading the sanitized client-context chapter…
02 · Boundary
What is confirmed—and what is not
This report refuses to turn a visible menu, connected object or source-side event into proof of client value.
Fresh commercial identity is confirmed; commercial standing is still provisional.
03 · Capacity
TAM coverage must be mathematically safe
Mailbox count is not capacity. Only authenticated, warmed, healthy and destination-tested resources count.
Current 90-day plan
1,361 contacts · 3 emails each
1,361 × 3 ÷ 65 = 62.8 sends/day
At 12 planned sends per healthy mailbox: 6 active, then add the operating buffer for 8 healthy mailboxes total.
Current safely usable evidence: 0 of 8. Twenty-four connected resources do not satisfy the requirement.
Full approved-criteria proxy
14
11 active + 3 buffer
1,308 companies × two contacts × three emails over 65 weekdays. This is the minimum full-ICP plan after approval, verification and suppression.Currently observed capped pool
10
8 active + 2 buffer
1,910 recent buyer work-email records after a two-per-company cap. Presence is not verification; safe proven capacity remains zero.04 · Market scope
Full product TAM is not the same as approved ICP
graph8 Mashup now supplies a bottom-up market view. The layers remain separate so an expansion opportunity is never mistaken for client-approved targeting or campaign-ready contacts.
Research result—not an activation list.
The deterministic 40-company sample exposed definite false positives and identity/vertical ambiguities, so the saved query did not pass the 90% activation gate. No Mashup candidate was added to Ability, enrolled or contacted.
The deterministic 40-company sample exposed definite false positives and identity/vertical ambiguities, so the saved query did not pass the 90% activation gate. No Mashup candidate was added to Ability, enrolled or contacted.
Full capability TAM proxy
21,383
Companies · known $5M+ revenue · physical-product commerce · at least one observed channel. This is the broader product expansion universe.
Approved-criteria ICP proxy
1,308
Companies · approved vertical signals · at least two observed channels. Candidate search result; client review still required.
Existing approved pilot
731
Companies · A+B+C research grades. The domains were unavailable in this evidence bundle, so overlap with the 1,308 is not claimed.
Verified campaign emails
595
Emails in the existing 1,361-contact cold cohort. This is 43.7% of that cohort and is not an account count.
| Layer | What is now measured | Status | Exact next action |
|---|---|---|---|
| Full product capability | 21,383 company candidates across Ability's broader commerce-operation use cases. | Expansion research | Ability approves or rejects the additional capability lanes; graph8 then refines one lane at a time and re-samples before activation. |
| Client-approved ICP criteria | 1,308 company candidates with known $5M+ revenue; 1,070 have at least one recent buyer work email and 840 have at least two. | Candidate, not approved | Review a new deterministic 100-company sample, correct source/entity mismatches, reach ≥90% clear fit, then save the approved query version. |
| Existing pilot audience | 731 approved companies, 1,361 contacts and 595 verified-good emails from the prior build. | Approved, unreconciled | Export the 731 domains, normalize and compare them with the 1,308. Publish overlap, net-new, excluded and stale counts—never infer coverage from 731 ÷ 1,308. |
| Campaign-ready audience | Mashup contains 6,779 recent buyer work-email records; capped at two per company, 1,910 records cover 1,070 companies. They are not deliverability-verified. | Not ready | Verify employment and email, cap by persona/account, deduplicate, apply customer/competitor/unsubscribe/bounce/DNC suppressions and read the exact eligible cohort back from the destination list. |
Method boundary: US; source employee bands 11–50 through 501–1000 approximate Ability's 15–700 requirement and necessarily include some out-of-band companies. Known revenue of $5M+ is required for headline counts. Query receipt abfefc4fd4ff · validated August 5, 2026.
05 · Priority
The four blockers that control everything else
Close these first. More campaigns or automation before these controls would add volume without a reliable revenue system.
Suppression control is unproven
Zero rows can be valid. Ability's risk is that historical sources are unreconciled, known bounce evidence exists, future hard-bounce capture is disabled and execution-time stop tests are absent.
Safe email capacity is zero
All 24 mailboxes are Healing/Watch, 12 show error/no-score and four domains were independently blocklisted.
No reply-to-revenue chain
Replies exist, but accountable response, Appointments, held/qualified meeting, CRM Deal and revenue readback are not proven.
Commercial scope is unresolved
The subscription object is Active, while current price/payment, entitlements and managed-GTM service scope need Finance/CS reconciliation.
06 · Scorecard
Every graph8 menu item, independently graded
Use the filters to isolate a menu, P0/P1 work or capabilities below grade 3. Each row states what is done and the exact next implementation action.
| Capability | Grade | State / evidence | What is done | Why this grade | Exact next implementation | Proof for next level | Owner |
|---|
No capabilities match these filters.
07 · Expansion
The implementation sequence to 5.0
The goal is not 67 simultaneous projects. It is a safe, owned signal-to-revenue loop, followed by portfolio completion and measured optimization.
0–24 hoursAuthority and safety: reconcile Billing/service scope; implement central suppressions and Compliance; quarantine unhealthy senders; name owners; prove Roles and tenant isolation.
Days 1–5Minimum complete loop: reconcile the existing 731 domains against the 1,308 ICP proxy; pass the search precision gate; publish canonical TAM and Master Lists; recover eight healthy mailboxes for the current cohort; then prove controlled ICP/intent/competitor/visitor/reactivation journeys through Inbox, Appointments, CRM and Reports.
Days 5–15Portfolio completion: Nurture, Web Chat, human SDR/Dialer, Work/Tasks, Agents/Skills/Workflows, Content/Brand and every Active/Ready/Exempt decision.
Days 15–60Operating proof and optimization: two 30-day periods at grade-4 standards; weekly reconciliation; improve audience, message, signal, routing, timing, page, coaching or channel mix from qualified outcomes.
Evidence limitation
Capability evidence is the direct Ability-specific August 4 baseline unless a row is marked Fresh. Fresh August 5 evidence confirms organization identity and the Active Platform subscription, while the new TAM chapter is a read-only Mashup research result. The admin grid and customer modal report conflicting credit balances, so credit state remains unresolved. No grade relies on the CIENCE aggregate dashboard. No email, call, candidate, campaign, list, suppression, billing, CRM or tenant record was changed.
The 21,383 and 1,308 market counts are provisional search estimates, not approved audiences. The source bands cannot represent 15–700 employees exactly, the sample did not pass the precision gate, work-email presence is not deliverability verification, and the existing 731 domains were unavailable for overlap readback. All grades remain provisional until controlled acceptance refreshes execution, negative tests and destination receipts in the real Ability organization. This revision was validated against portable-skills digest 7fd2f8e85d3c and OKF digest 78e608f52d6b.