Migrate Genesys CX Cloud TEI study to the pattern; retire Streamlit app

studies/202512_GenesysCX -> studies/202512_TEI_Genesys_CX_Cloud,
rebuilt as pattern Variant 4 (TEI composite reproduction):

- teicalc/ self-contained engine: Forrester's tables as the never-edited
  verbatim anchor (incl. the p.14 typo note and the $0 AI-token line),
  generic model/scenarios/staging carried over from the Amazon Connect
  study, ClientDrivers overlay (agents / weekly interactions / revenue,
  flat composite so no growth re-base) with ai_tokens_annual as a direct
  input for the token line the published study models at $0
- one deliverable notebook (business_case.ipynb): widget-pair sidebar
  drivers incl. the AI-token price, published-vs-overlay KPI columns,
  cash-flow/waterfall/scenario charts, verification gate, backstage JSON
  data appendix
- gate + tests reproduce the published totals within $2: NPV $10.8M /
  ROI 266% (engine $10,783,466 / 265.79%; payback 3.3 months, not
  headlined in the PDF); 29 study tests, headless nbconvert green,
  stage simulation leak-free, exports carry the appendix
- old Athena workflow (00_provision..04_export, config.py, seed_data.py,
  PALLADIUM_GENESYSCX_* keys, ATHENA_EXPECTED reconciliation) deleted;
  git history preserves it

With the last legacy study migrated, the retirement lands too:
- app/ (Streamlit UI) and core/notebook_helpers deleted; nothing else
  imported them
- streamlit stripped from pyproject extras, requirements.txt, Makefile;
  .env.example reduced to the Athena keys; 00_setup.ipynb and
  core/bootstrap.py repointed at the pattern studies
- root README reworked: self-contained studies + slim core/ Athena
  toolkit (tei_client, calculations, export, cli)

All suites green: Genesys 29, Amazon Connect 27, CTM 55, template 7,
root 58.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

studies/202512_TEI_Genesys_CX_Cloud/teicalc/anchor.py

@@ -0,0 +1,197 @@
+"""
+The verbatim anchor — Forrester *The Total Economic Impact™ Of CX Cloud —
+Cost Savings And Business Benefits Enabled By Genesys And Salesforce*
+(December 2025, commissioned by Genesys and Salesforce).
+
+VERBATIM, do not edit. These are Forrester's published composite-organization
+tables and financial summary, transplanted unchanged from the study PDF
+(``docs/The-Total-Economic-Impact-Of-CX-Cloud.pdf``). Client personalization
+lives in :mod:`teicalc.overlay`; scenario stress lives in
+:mod:`teicalc.scenarios` — both deep-copy, neither mutates this record.
+
+Rows keep Forrester's own year-index keys (``"1"``/``"2"``/``"3"``);
+:mod:`teicalc.model` maps them to calendar years (2026–2028). Values are
+*nominal* (pre-risk-adjustment); the risk factor is stored per row and
+applied by the model (benefits ×(1−rf), costs ×(1+rf), per the TEI
+methodology).
+
+Two study-specific footnotes, preserved from the source review:
+
+* The published Total Costs table (p.14) prints the implementation initial
+  as $1,304,600, but the detail table, the cash-flow analysis, and the math
+  (1,190,000 × 1.10) all give **$1,309,000** — the p.14 figure is a typo in
+  the study.
+* ``genesys_ai_tokens`` is **not in the published study** — Forrester
+  modeled $0 AI consumption even though benefits B (self-service uplift),
+  C (agent efficiency), and D (agent assist upsell) all depend on AI
+  capabilities that Genesys bills via AI Experience tokens. The row is
+  anchored at $0 so the reproduction matches the published totals; client
+  cases price it via the overlay's ``ai_tokens_annual`` driver.
+"""
+
+from __future__ import annotations
+
+#: 3-year nominal benefit cashflows — 🟢 published.
+BENEFITS_VERBATIM: list[dict] = [
+    {
+        "field_key": "legacy_retirement",
+        "table": "benefits",
+        "label": "Retirement of legacy systems with CX Cloud adoption",
+        "category": "Cost Savings",
+        "year_values": {"1": 680_000, "2": 930_000, "3": 930_000},
+        "risk_adjustment": 0.05,
+        "notes": (
+            "PDF A1–A4. Telephony $250k Y1 ramping to $500k (legacy "
+            "sunset completes mid-Y1) + WFM/recording/transcription apps "
+            "$100k + reduced dev effort $230k (2,400 hrs @ $94) + reduced "
+            "platform mgmt $100k (1,500 hrs @ $65). Risk adj 5%."
+        ),
+    },
+    {
+        "field_key": "self_service_savings",
+        "table": "benefits",
+        "label": (
+            "Cost savings from reallocated workers and avoided seasonal "
+            "hires with increased customer self-service"
+        ),
+        "category": "Productivity",
+        "year_values": {"1": 2_329_600, "2": 2_329_600, "3": 2_329_600},
+        "risk_adjustment": 0.15,
+        "notes": (
+            "PDF B1–B8. Self-service completion 15%→25% on 80k weekly "
+            "interactions → 8,000 deflected/week → 40 FTEs @ $58,240 "
+            "fully burdened. Risk adj 15%. (PDF B7 formula cites B2 where "
+            "the 12-min interaction length is meant; 40 FTEs is correct.)"
+        ),
+    },
+    {
+        "field_key": "agent_efficiency",
+        "table": "benefits",
+        "label": "CX agent efficiency gains",
+        "category": "Productivity",
+        "year_values": {"1": 2_912_000, "2": 2_912_000, "3": 2_912_000},
+        "risk_adjustment": 0.10,
+        "notes": (
+            "PDF C1–C6. MTTR 12→10 min on 60k agent-handled interactions "
+            "per week → 104,000 hrs/yr @ $28 fully burdened. Risk adj 10%."
+        ),
+    },
+    {
+        "field_key": "agent_assist_sales",
+        "table": "benefits",
+        "label": "Incremental sales from agent assist capabilities",
+        "category": "Revenue",
+        "year_values": {"1": 600_000, "2": 600_000, "3": 600_000},
+        "risk_adjustment": 0.05,
+        "notes": (
+            "PDF D1–D3. $500M revenue impacted (20% of $2.5B) × 1.5% lift "
+            "× 8% gross margin. Risk adj 5%."
+        ),
+    },
+]
+
+
+#: Costs include an ``initial`` (year-0, undiscounted) component for
+#: implementation. Cost risk adjustments are applied *upward*. 🟢 published
+#: (except the ``genesys_ai_tokens`` line — see the module docstring).
+COSTS_VERBATIM: list[dict] = [
+    {
+        "field_key": "cx_cloud_licenses",
+        "table": "costs",
+        "label": "CX Cloud solution costs (licenses)",
+        "category": "Subscription",
+        "initial": 0,
+        "year_values": {"1": 840_000, "2": 840_000, "3": 840_000},
+        "risk_adjustment": 0.05,
+        "notes": (
+            "PDF E1–E3. Genesys Cloud CX 2 $170/user/mo + Salesforce "
+            "Voice $25/user/mo + connector $25/user/mo, 400 concurrent "
+            "users, 20% contractual discount → $650k + $95k + $95k. "
+            "Risk adj +5%. Seat licenses ONLY — AI consumption is a "
+            "separate line (genesys_ai_tokens)."
+        ),
+    },
+    {
+        "field_key": "implementation",
+        "table": "costs",
+        "label": "Implementation and deployment cost",
+        "category": "Implementation",
+        "initial": 1_190_000,
+        "year_values": {"1": 0, "2": 0, "3": 0},
+        "risk_adjustment": 0.10,
+        "notes": (
+            "PDF F1–F5. 10-week implementation: 20 FTEs @ $80/hr fully "
+            "burdened ($640k) + $550k professional services. Risk adj "
+            "+10% → $1,309,000 (the p.14 Total Costs table's $1,304,600 "
+            "is a typo in the study)."
+        ),
+    },
+    {
+        "field_key": "ongoing_management",
+        "table": "costs",
+        "label": "Ongoing management costs",
+        "category": "Operations",
+        "initial": 0,
+        "year_values": {"1": 202_800, "2": 202_800, "3": 202_800},
+        "risk_adjustment": 0.10,
+        "notes": (
+            "PDF G1–G3. 5 people @ 30% time (12 hrs/wk) @ $65/hr. "
+            "Risk adj +10%."
+        ),
+    },
+    {
+        "field_key": "genesys_ai_tokens",
+        "table": "costs",
+        "label": "Genesys AI Experience token consumption",
+        "category": "Subscription",
+        "initial": 0,
+        "year_values": {"1": 0, "2": 0, "3": 0},
+        "risk_adjustment": 0.0,
+        "notes": (
+            "NOT in the published study — Forrester modeled $0 AI "
+            "consumption even though benefits B (self-service uplift), "
+            "C (AI coaching/assist), and D (agent assist upsell) all "
+            "depend on AI capabilities that Genesys bills via AI "
+            "Experience tokens. Anchored at $0 to reproduce the published "
+            "totals. For client cases, enter the negotiated annual token "
+            "cost from the Genesys quote (the overlay's ai_tokens_annual "
+            "driver) and document the quote details (token volume, unit "
+            "price, tier)."
+        ),
+    },
+]
+
+
+#: Composite-organization drivers — 🟢 published (PDF "Composite Organization").
+ASSUMPTIONS: dict = {
+    "annual_revenue": 2_500_000_000,
+    "employees": 10_000,
+    "agents_fte": 600,
+    "concurrent_licenses": 400,
+    "weekly_interactions": 80_000,
+    "interaction_minutes": 12,
+    "self_service_rate_before": 0.15,
+    "self_service_rate_after": 0.25,
+    "mttr_saved_minutes": 2,
+    "agent_hourly_rate": 28,
+    "agent_annual_salary": 58_240,
+    "revenue_impacted": 500_000_000,
+    "revenue_lift": 0.015,
+    "gross_margin": 0.08,
+    "discount_rate": 0.10,
+    "analysis_years": 3,
+}
+
+
+#: The PDF's Financial Summary — the gate's reproduction target. 🟢 published.
+#: The engine reproduces these to within $2 (Forrester's own rounding).
+#: Forrester does not headline a payback for this study; the engine computes
+#: 3.3 months from the cash-flow table.
+PUBLISHED: dict = {
+    "benefits_pv": 14_840_638,
+    "costs_pv": 4_057_170,
+    "npv": 10_783_468,
+    "roi_pct": 266,
+    "discount_rate": 0.10,
+    "analysis_years": 3,
+}