01-projects/phdata

2026 review-cycle goals - draft v1 (to align with Taylon)

2026-10-21·status: draft - awaiting founder's previous draft + CAF numbers·source: founder iMessage 2026-10-21 09:33 + 09:46 ET (Taylon's goal list screenshot + 4 goal descriptions)
phdatagoalsai-workforcereview-cycle

2026 review-cycle goals - draft v1

Context (founder, 10-21): Taylon is his manager (promoted from IC when AI Workforce moved under Applied AI; Andrew Kim went to the Anthropic tiger team). Corporate parents: Run phData on AI / Deliver with AI / Grow through AI / Drive Delivery Model Excellence. Founder dislikes the utilization goal (staffing is outside his control). Completed goals are fine to include; he wants some Q4 stretch. Hired as Snowflake/Cortex specialist, never staffed on Cortex; bench time filled with Anthropic enablement engagements.

Evidence he named: internal Glean CAF workflow (sales request → CAF report, 100+ runs); 3 client Copilot agents (Lionsgate Music Rights Legal Review, Lionsgate AR Processor, Kwik Trip Sprocket); Kwik Trip AI SDLC workflow bundled into an agent factory = technical core of "codify a repeatable playbook".

Ray's proposed set (5)

# Goal (Taylon's) Already banked Q4 stretch Parent (Ray's guess)
1 Internal AIWF ops agents CAF Glean workflow, 100+ reports Ship 1 more team workflow (scoping helper or maturity summarizer); measure ≥3 regular users + hours/week for both Run phData on AI
2 HITL quality + evals Review stage inside the agent factory Test set + acceptance criteria + ≥10 logged eval runs for CAF and the new agent from #1 Deliver with AI
3 Codify delivery playbook Agent factory (Kwik Trip) Write the non-technical half (cadence, roles, risks, change mgmt/adoption) across Kwik Trip + Lionsgate; get reuse on 1 more project Drive Delivery Model Excellence
4 Anthropic certification CCA-F passed 2026-08-17 CCDV-F by 11-20, CCAR-P by 12-16 (tracked in [[cert-progress]]; see [[2026-10-20-ccdv-f-readiness]] for the CCDV-F readiness read) Grow through AI
5 Patterns → offers + stories (Ray to confirm) Playbooks = #1-#3 outputs; 2 external assets (blog/webinar/case study) Grow through AI

Hours-returned (#3 of his list) folded into #1 as the measure rather than a standalone goal: the ≥20%-gain-on-2-3-engagements version has the same staffing dependency he objected to in utilization. Include standalone only if Lionsgate before/after numbers exist.

Open asks to founder

v2 (founder input 2026-10-21 09:55 ET)

Founder rulings: drop the "Deliver with AI" responsiveness/peer-review goals (baseline practice); KT Phase 2 extension was out of his control, drop; adopt Taylon's generic playbook language over "Productize Copilot Studio practice"; keep the case study but generic (Anthropic partnership needs 3 published case studies by EOY, has 1; could come from enablements or a new staffing); add a sales-support goal.

CAF facts (founder): greenfield 2026; 5-12 requests/week; tagged → Glean processes in <5 min; end-to-end CAF prep 4 business days → <1 business day with better QC; previously requests were lost/untracked; 100+ reports.

Parent Goal Banked Q4 stretch
Run phData on AI Internal ops agents CAF: 100+ reports, 4d → <1d 2nd team workflow (scoping helper) w/ ≥3 users
Deliver with AI Evals + human review - test set + acceptance criteria + ≥10 runs for CAF + new agent
Drive Delivery Model Excellence Codify repeatable delivery playbook (Taylon's language) KT agent factory, Lionsgate x2, KT Sprocket people half written + reuse on 1 more project
Grow through AI Anthropic cert CCA-F CCDV-F 11-20, CCAR-P 12-16
Grow through AI Published case study (generic) - contribute ≥1 to the Anthropic 3-by-EOY count
Grow through AI Support sales CAF is sales-support infra already SME on ≥3 pursuits, track influenced pipeline

v3 - form-ready objectives + key results (sent 2026-10-21 ~10:05 ET)

Form fields (Lattice-style): Objective = title, description, parent goal, cycle 2026 (Mar 2 - Dec 31). KR = title, description, measure (default Percent 0-100; Number assumed available), start, target, parent.

O1 Run AI Workforce operations on internal agents | parent: Internal AIWF "ops agents" that actually run the team

O2 Make evals + human review standard before an agent ships | parent: Human-in-the-loop quality + evals are not optional

O3 Codify a repeatable AI Workforce delivery playbook | parent: same-named Taylon goal

O4 Earn Anthropic + Snowflake AI certifications | parent: Complete Anthropic Certification

O5 Turn delivered work into published stories | parent: Turn patterns into reusable offers + stories

O6 Support sales on AI pursuits | parent: Grow through AI

SUBMITTED (founder, iMessage 2026-10-21 15:56 ET)

Canonical export: [[2026-10-21-lattice-goals-submitted.csv]] (Lattice, 28 rows). Supersedes the drafts above and [[2026-goals-ben-wilson]] where they differ.

Founder changes vs Ray's v4: sales objective held off ("not sure I can get involved in those deals at this time"); certs rolled into "Get key partner certifications" (Grow through AI) as 5 binary KRs; CAF turnaround KR reframed as "Meet or exceed CAF Report Generation SLA" 4→2; internal case study = "Present 1 internal case study"; 2nd internal workflow KR and Royal Caribbean KR dropped; utilization objective still present (Lattice-assigned). All progress = 0 at export.

Already earned, progress not yet entered: CCA-F, SnowPro GenAI, 100+ CAF requests, CAF SLA, 3 Copilot agents, agent factory, 15+ of 24 enablement sessions. Open for Q4: CCDV-F 11-10, AWS AI Practitioner 12-04, CCAR-P 12-22, primary facilitator, eval test sets x2 + 20 runs, playbook publish + 1 reuse, 1 internal case study, 2 external assets, 9 more enablement sessions.