wulf-pulse/LCI_FEATURE_DESIGN.md

Lifecycle Insights Feature Replacement — Pulse Design Doc

Purpose: Document how Lifecycle Insights (LCI) works across the four target features — Warranty, Assessments, Budget Forecast, and QBR/Business Reviews — so we can build equivalent functionality in Pulse, using Autotask (and Datto RMM) as the primary data source.

Reference client: Hynes Industries (321 assets, multiple locations) Source tool: app.lifecycleinsights.io Date captured: 2026-03-16

---

1. Asset List & Warranty

What LCI Shows

Summary banner (top of Asset List page):

• Total Assets: 321
• EOL Status breakdown:
  • 104 Good Standing
  • 60 Approaching EOL
  • 6 Past EOL
  • 151 Unknown
• OS Status breakdown:
  • 177 Supported
  • 27 Not Supported
  • 117 Unknown
• Warranty Status breakdown:
  • 91 Under Warranty
  • 20 Approaching Warranty Expiration (within ~90 days)
  • 97 Expired Warranty
  • 113 Unknown

Per-asset table columns:

Column	Example
Device Name	YNGHYNLT030
Status	Active / Retired / Hidden
Last User (RMM)	HYNES\shopfloor
Location	Youngstown (Main) / Painesville (Branch) / Kokomo (Branch)
Type	Workstation / Hypervisor Host / Network Device
Make	Dell / Lenovo
Model	Latitude 5591 / ThinkSystem SR630
Serial Number	2KNXWT2
Operating System	Microsoft Windows 10 Pro 10.0.19044
Purchase Date	04-23-2019
Warranty Expiration	04-24-2023
End of Life	04-23-2024
Replacement Cost	$3,000
Age (years)	6.9

Table features:

• Sortable columns (click header)
• Filter buttons: Show Only NON Hidden / Duplicate / Hidden / All Assets
• Duplicate detection badge
• Pagination (321 rows, 25/50/75/100/200 per page)
• Search box

Warranty status logic (inferred):

• Under Warranty: warranty expiration date > today
• Approaching: warranty expiration within ~90 days
• Expired: warranty expiration date < today
• Unknown: no warranty expiration date on record

EOL logic (inferred):

• EOL date is OS/hardware specific (LCI maintains a lookup table — shown on dashboard)
• Good Standing: EOL date > today + threshold (~1 year)
• Approaching EOL: EOL date within ~1 year
• Past EOL: EOL date < today
• Unknown: no EOL date derivable

Data Sources

• Device inventory: comes from RMM (Datto RMM in Hynes's case) — device name, make, model, serial, OS, last user, location
• Purchase/warranty dates: can come from RMM or be manually entered
• EOL dates: LCI maintains a global EOL lookup table by OS/software version (visible on dashboard — e.g., "Windows 10 Pro: 10-14-2025")
• Replacement cost: manually set per asset type, or inherited from category defaults
• Locations: mapped from RMM site groupings

Pulse Implementation Notes

Data model (new migration needed):

-- assets table (core)
asset_id, autotask_id, datto_rmm_uid,
device_name, status, last_user,
location_id, asset_type,
make, model, serial_number,
os_name, os_version,
purchase_date, warranty_expiry_date, eol_date,
replacement_cost, age_years,
company_id (FK → companies),
is_hidden, is_duplicate,
created_at, updated_at, last_synced_at

-- eol_reference table (global)
software_name, eol_date

-- asset_locations table
location_id, company_id, location_name

Warranty status computed field:

function warrantyStatus(expiryDate: Date | null): 'under' | 'approaching' | 'expired' | 'unknown' {
  if (!expiryDate) return 'unknown';
  const daysUntil = differenceInDays(expiryDate, new Date());
  if (daysUntil < 0) return 'expired';
  if (daysUntil <= 90) return 'approaching';
  return 'under';
}

Autotask as source: Autotask has a Configuration Items module that maps to devices. Datto RMM syncs device details into Autotask CIs. Fields available via Autotask API: warrantyExpirationDate, purchaseOrderNumber, purchaseDate, installDate, serialNumber, referenceNumber, manufacturer, product (model), location, deviceType.

---

2. Budget Forecast

What LCI Shows

Purpose: Project technology spend for the next 6–8 quarters, broken into categories.

Views:

1. Detail View — collapsible sections per time period, category breakdown
2. Spreadsheet View — bar chart + grid with rows per category, columns per quarter
3. Budget Board — (not fully captured; likely a Kanban/card layout)

Filters available:

• Location (site filter)
• Asset Type
• Asset Status
• "Generate Budget Forecast as of" date (defaults to today; can run future projections)

Time buckets:

• Overdue (past-due items)
• Recommendations Not Scheduled (approved recommendations with no date)
• Then rolling quarters: e.g., Jan-Mar 2026, Apr-Jun 2026, Jul-Sep 2026, etc. (6–8 quarters shown)

Categories:

Category	What it represents
Workstation	Assets reaching EOL scheduled for replacement
Hypervisor Host	Server/hypervisor assets reaching EOL
Network Device	Network hardware replacements
Contracts/Subscriptions	Contract renewals (from Autotask contracts)
Recommendations	Open recommendations with cost + target date

Hynes Industries example data (Spreadsheet View):

Category	Overdue	Not Sched	Apr-Jun 26	Jul-Sep 26	Oct-Dec 26	Jan-Mar 27	Apr-Jun 27	Total
Hypervisor Host	$22,500	—	—	—	—	—	—	$22,500
Workstation	—	—	—	—	—	$18,000	$21,000	$39,000
Contracts/Subscriptions	—	—	$37,231	$43,590	$164,414	$27,737	$37,231	$313,549
Recommendations	$254,754	$107,000	$31,245	$15,045	$29,795	—	—	$437,839
TOTAL	$277,254	$107,000	$68,476	$58,635	$194,209	$45,737	$58,231	$812,888

How Each Category is Populated

Workstation / Hypervisor / Network Device:

• Source: assets with EOL date falling within the quarter
• Cost: replacement_cost per asset
• Scheduled quarter = quarter containing the asset's eol_date

Contracts/Subscriptions:

• Source: Autotask Contracts with end_date or renewal dates
• Cost: contract value
• Quarter = quarter of contract end/renewal date

Recommendations:

• Source: open Recommendations (see Recommendations module)
• Cost: estimated_cost on the recommendation
• Quarter = recommendation's planned_date (or "Not Scheduled" if no date)
• "Overdue" = planned_date < today

Pulse Implementation Notes

Budget Forecast is computed at query time — no separate stored table needed beyond assets + contracts + recommendations with dates and costs.

-- Pseudo-query for budget forecast
SELECT
  quarter_bucket,
  category,
  COUNT(*) AS quantity,
  SUM(cost) AS budget_cost
FROM (
  -- EOL assets
  SELECT date_trunc('quarter', eol_date) AS quarter_bucket,
         asset_type AS category,
         replacement_cost AS cost
  FROM assets WHERE company_id = $1 AND eol_date IS NOT NULL AND status = 'active'

  UNION ALL

  -- Contracts
  SELECT date_trunc('quarter', end_date), 'Contracts/Subscriptions', contract_value
  FROM contracts WHERE company_id = $1

  UNION ALL

  -- Recommendations
  SELECT date_trunc('quarter', COALESCE(planned_date, NULL)), 'Recommendations', estimated_cost
  FROM recommendations WHERE company_id = $1 AND status NOT IN ('Completed', 'Declined')
) combined
GROUP BY quarter_bucket, category
ORDER BY quarter_bucket;

API route: GET /api/budget-forecast/[companyId]?asOf=DATE&location=X&assetType=Y

---

3. Assessments

What LCI Shows

Structure (from Assessments Menu page):

1. Assessment Events — each time you run an assessment for a client = one "event"
2. Assessment Templates — define what you're measuring (items + categories)
3. Assessment Items — individual questions/checklist items
4. Event Types — classify events (e.g., baseline, annual review, gap assessment)
5. Analytics by Customer — score history, trends, category breakdowns
6. Analytics All Customers — cross-client comparison

Wulf Consulting uses: "Wulf Consulting: Default Assessment" template

Hynes Industries history (from dashboard):

• Latest: completed 01-29-2025 (411 days ago) by Jake Hammel
• Score type: "baseline assessment with a score of --" (baseline = no numeric score, just captures state)
• Assessment Event Score History chart shows one data point at 2025-01

Data Quality Scores (dashboard, related to assessments):

• EOL: 53/100
• Warranty: 65/100
• Make: 92/100
• Serial: 83/100
• Model: 95/100
• Overall: 67/100

Assessment Data Model (inferred)

AssessmentTemplate
  ├── id, name, company_id (MSP-level)
  ├── items[] → AssessmentItem
  │     ├── id, category, question_text, weight, scoring_type
  │     └── possible_scores: [0,1,2,3,4,5] or pass/fail
  └── event_type_config

AssessmentEvent
  ├── id, template_id, client_id (company)
  ├── event_type (baseline / annual / gap)
  ├── completed_date, completed_by (technician)
  ├── overall_score (numeric or null for baseline)
  └── responses[] → AssessmentResponse
        ├── item_id, score, notes
        └── linked_recommendation_id (optional)

AssessmentEventScore (for history/chart)
  ├── event_id, category, score, max_score

Key behaviors:

• Templates are MSP-defined (not per-client)
• One client can have many events over time (score history chart)
• Assessment items can generate Recommendations (linked)
• Scores tracked by category (enables radar/bar chart per category)
• "Baseline" events don't produce a numeric score — they capture current state only

Pulse Implementation Notes

Migration: Create assessment_templates, assessment_items, assessment_events, assessment_responses tables.

Admin UI: Template builder (drag-drop items into categories), Event runner (per-client form), Analytics view (score history chart per client).

Autotask link: Assessment items that result in action → create Autotask ticket or recommendation.

---

4. Business Reviews / QBR

What LCI Shows — FULLY CAPTURED

Page: Business Review Meetings for Hynes Industries

Meeting table columns:

Column	Values
Meeting Date	e.g., 03-03-2026
Meeting Type	Business Review / Virtual Meeting / (blank)
Primary Contact	Ryan Day
Account Manager	Jake Hammel
Status	Completed / Scheduled / Auto-Scheduled / Preparation / Skipped/Cxl
Notes	Rich text notes (Pre-Meeting / During Meeting / Post-Meeting sections)
Recommendations	(linked recommendations column)
Reports / Artifacts	(PDF download column)
Delete	delete action

Status types explained:

• Completed — meeting happened and notes are finalized
• Scheduled — manually scheduled upcoming meeting
• Auto-Scheduled — system automatically created this slot based on cadence
• Preparation — notes are being drafted before meeting
• Skipped/Cxl — meeting was skipped or cancelled

Cadence for Hynes: Bi-weekly (every 2 weeks), Ryan Day as primary contact, Jake Hammel as account manager. Running since September 2023.

Notes structure (rich text, per meeting):

Pre-Meeting:
  [prep notes, project status, talking points]

Meeting / During Meeting:
  [live notes taken during call]

Post-Meeting:
  [follow-up actions, decisions made]

Real example notes (from Hynes 09-19-2023):

"Follow up with Ryan on 2024 CAPX Budget... Updated Ryan that moving forward new workstations won't have the BitLocker Password Prompt... Amy Young - Executive Admin - Get her to view details on the Hynes conference room calendars..."

Scale: Hynes has ~56 business review records from 09-2023 to 03-2026 (bi-weekly cadence = ~2/month × 30 months). The "47 overdue" from the dashboard is because many Auto-Scheduled meetings were never formally completed.

Filters on page:

• "Show Meetings with Date in Past" toggle

Actions:

• ADD NEW — create a new meeting record
• Each row has: delete, reports/artifacts (PDF generation)

Business Reviews Data Model (confirmed)

business_reviews
  id, company_id,
  meeting_date DATE,
  meeting_type VARCHAR,           -- 'Business Review', 'Virtual Meeting', etc.
  primary_contact_id INT,         -- FK → contacts
  account_manager_id INT,         -- FK → users (MSP staff)
  status VARCHAR,                 -- Completed/Scheduled/Auto-Scheduled/Preparation/Skipped-Cxl
  notes_pre_meeting TEXT,
  notes_during_meeting TEXT,
  notes_post_meeting TEXT,
  created_at, updated_at

business_review_artifacts          -- linked PDFs / reports
  id, business_review_id, artifact_type, file_url, created_at

business_review_recommendations    -- join table
  business_review_id, recommendation_id

Key behaviors:

• System auto-generates meeting slots based on cadence (configurable per client)
• Each meeting has pre/during/post note sections
• Meetings can link to specific recommendations (show progress at meeting)
• PDF reports can be attached or generated per meeting
• "47 overdue" = count of Auto-Scheduled meetings with date in past and no Completed status
• Agenda links (from dashboard widget) are separate from meeting notes — persistent links per client

Pulse Implementation Notes

QBR = structured meeting log with recurring cadence + notes + linked recommendations.

• Auto-generate meeting slots via a cron job based on meeting_cadence_days per company
• The "Overdue" count = SELECT COUNT(*) FROM business_reviews WHERE company_id = $1 AND meeting_date < NOW() AND status != 'Completed'
• PDF generation: render current EOL health + budget + recommendations at time of export
• Agenda links widget: simple per-client URL bookmarks (business_review_links table)

---

5. Recommendations — FULLY CAPTURED

What LCI Shows

Page: Recommendations for Hynes Industries — 26 total records

Table columns:

Column	Notes
Name	Free-text project/recommendation name
Location	Youngstown (Main) / Painesville (Branch) / Kokomo (Branch) / blank (all sites)
Start	Target start date (optional)
End	Target end date (optional)
Status	New / Proposed / Planned / In Progress / On Hold / Completed / Declined
Tags	(free-form tags)
Priority	Low / Medium / High
Description	Rich text description of scope
Labor Cost	$
Materials Cost	$
Replacement Cost	$
Recurring Cost	Monthly $X → shown as "Monthly / $X"
Total Cost	Sum of all cost components
Delete	delete action

Filters:

• Search box
• Include Completed toggle
• Include Declined toggle
• "SHOW AUTOTASK OPPS" button — links to Autotask opportunities

Views: List (default) — table view

Hynes Industries full recommendation list (real data):

Name	Status	Priority	Total Cost
2024 - Fish Bowl Renovation	Completed	Medium	$11,000
2024 - Pepwave Licensing Renewal	Completed	Medium	$0
2024 - Planned Multiple Switch Lifecycle	Completed	Medium	$24,000
2024 - Planned Workstation Lifecycle	In Progress	Medium	$9,800
2024 - VLAN Cleanup	Completed	Medium	$10,000
2024: O365 Backup	Completed	Medium	$3,560 (monthly)
2024: Password Management	Completed	Medium	$8,340 (monthly)
2024: SEIM Tool	Completed	High	$20,020 (monthly)
2024: ZTNA - Phase 1	Completed	High	$16,350 (monthly)
2025 - Painesville Zoom Room	Declined	Low	$11,000
2025 - Planned Infrastructure Lifecycle	Completed	High	$6,830 (monthly)
2025 - Planned Workstation Lifecycle	Completed	High	$3,000
2025 - PV Planned Multiple Switch Lifecycle	In Progress	High	$5,000
2025 - YNG Planned Multiple Switch Lifecycle	Completed	High	$2,900
2025: Windows 10 to 11 Upgrade	Completed	High	$40,400
2026 - Exterior Cameras - Partial Build	On Hold	Low	$57,000
2026 - PEN Test	On Hold	High	$20,000
2026 - Pepwave Licensing Renewal	Planned	High	$350
2026 - Planned Host/SAN Lifecycle	Proposed	High	$120,000
2026 - Planned Multiple Switch Lifecycle	Proposed	High	$31,000
2026 - Planned Workstation Budget	Proposed	Medium	$86,600
2026 - Zebra Printers Lifecycle	Proposed	High	$80,534 (monthly)
2026: ZTNA: Phase 2	On Hold	High	$30,000
Auvik SaaS	Declined	Medium	$1,760 (monthly)
Physical Switch & IDF Cleanup	On Hold	Medium	$0
Receptionist Hardware	Completed	Medium	$1,800

How recommendations feed Budget Forecast:

• Status Proposed, Planned, In Progress, On Hold → included in budget forecast
• start date (or end date) → determines which quarter they appear in
• No date → "Not Scheduled" bucket
• Past date → "Overdue" bucket
• Completed / Declined → excluded from forecast

Recommendations Data Model

recommendations
  id, company_id,
  name VARCHAR,
  location_id INT (nullable),
  start_date DATE (nullable),
  end_date DATE (nullable),
  status VARCHAR,              -- New/Proposed/Planned/In Progress/On Hold/Completed/Declined
  tags TEXT[],
  priority VARCHAR,            -- Low/Medium/High
  description TEXT,
  labor_cost DECIMAL,
  materials_cost DECIMAL,
  replacement_cost DECIMAL,
  recurring_cost DECIMAL,      -- monthly amount
  recurring_period VARCHAR,    -- 'Monthly'
  total_cost DECIMAL,          -- computed or stored
  autotask_opportunity_id INT (nullable),
  created_at, updated_at

recommendation_assets          -- assets linked to this recommendation
  recommendation_id, asset_id

Pulse Implementation Notes

• Recommendations are Pulse-native (not pulled from Autotask) but can push/link to Autotask Opportunities
• total_cost = labor_cost + materials_cost + replacement_cost + (recurring_cost × months_in_period)
• Budget Forecast query filters: status NOT IN ('Completed', 'Declined')
• Dashboard widget shows count + total cost grouped by status
• The SHOW AUTOTASK OPPS button = link to Autotask opportunities filtered to this company

---

6. Dashboard Summary

The LCI dashboard aggregates all four features for a client into one view:

Widget	Data
CSAT Score	From Crewhu integration (external)
Budget Forecast chart	6-quarter bar chart
EOL Health	% OK / Past EOL / Near EOL donut
Data Quality Score	Per-field completeness scores
Microsoft License Summary	From M365 / CSP data
EOL Replacement Cost	Bar chart by status
Latest Assessment	Score + date
Recommendation Status	Count + cost per status
Assessment Score History	Line/bar chart
Software/OS EOL Table	Global reference table
Tech Debt by Asset Type	Stacked bar
Missing Data Fields	Counts of missing fields
Total Technology Debt	$ value of overdue/EOL assets
Headlines	Alerts: overdue contracts, overdue QBRs, latest assessment age

---

6. Integrations & Data Flow

Datto RMM ──────────────────┐
                             ▼
Autotask CIs ───────────► LCI Asset Sync
                             │
Autotask Contracts ──────────┤
                             │
Autotask Recommendations ────┤
  (or LCI-native)            │
                             ▼
                    LCI PostgreSQL
                             │
                    ┌────────┼────────┐
                    ▼        ▼        ▼
               Budget    Assessments  QBR
               Forecast   Events     Reports

For Pulse:

• Assets: Autotask Configuration Items API → assets table
• Contracts: Autotask Contracts API → contracts table (already partially done)
• Recommendations: Either Autotask Opportunities/Quotes or Pulse-native table
• CSAT: Not currently integrated (Crewhoo is external) — can be skipped initially
• EOL Reference: Static lookup table (maintained manually or seeded from LCI's public list)
• Warranty dates: Come from Autotask CI warrantyExpirationDate field

---

7. Build Priority Order

1. Asset List + Warranty — highest value, data already in Autotask CIs
2. Budget Forecast — derived from assets + contracts (both already syncing)
3. QBR / Business Reviews — report generation + scheduling
4. Assessments — most complex, requires template builder + event runner

---

9. Pages Still to Explore

Fully captured: Asset List, Budget Forecast (detail + spreadsheet), Recommendations (all 26), Business Reviews (all ~56 meetings), Dashboard.

Still needed for complete picture:

• app.lifecycleinsights.io/assessmentevents — see an actual scored assessment event (items, categories, scores per item). The Hynes event from 01-29-2025 was a baseline with no numeric score.
• app.lifecycleinsights.io/assessmentanalytics — analytics/trends view per client
• app.lifecycleinsights.io/managetemplates — what a template looks like with items + categories
• MDS client — verify data model generalizes (different asset mix, different contact structure)

These can be explored via Claude_in_Chrome (already connected to local Chrome where you're logged in) without needing another MFA login.