# Omnichannel Engagement — Exact Rules, Headings, and Safety

> **COPILOT STUDIO KNOWLEDGE CONTRACT.** Use this file with the companion
> complete synthetic-records file. The deterministic reference responses below
> are the exact tool evidence persisted for every locked case; do not replace
> them with generic summaries or invent missing values.

## Approved personas and language focus

| Persona | Required focus |
|---|---|
| Customer Experience Leader | cross-channel continuity, service quality, and governance |
| Digital Engagement Manager | channel strategy, consent, and measurement |
| Contact Center Supervisor | handoff friction, unresolved needs, and service consistency |

## Exact routing and evidence contract

| Case | Route to operation | Persona | Exact arguments | Required transcript evidence |
|---|---|---|---|---|
| `OCE-01` | `channel_performance` | Customer Experience Leader | `{}` | `Prepared for:** Customer Experience Leader`; `Synthetic Channel Performance`; `no identity stitching` |
| `OCE-02` | `journey_analysis` | Contact Center Supervisor | `{}` | `Prepared for:** Contact Center Supervisor`; `Aggregate Customer Journey Analysis`; `Journey Optimization Opportunities` |
| `OCE-03` | `engagement_optimization` | Digital Engagement Manager | `{}` | `Prepared for:** Digital Engagement Manager`; `Draft Engagement Optimization Report`; `frequency caps` |
| `OCE-04` | `campaign_attribution` | Digital Engagement Manager | `{}` | `Synthetic Campaign Attribution Report`; `Overall Campaign ROI`; `Recommendations only` |
| `OCE-05` | `customer_journey` | Contact Center Supervisor | `{}` | `Customer Journey: Sarah Mitchell`; `Payment declined`; `$289 cart` |
| `OCE-06` | `unresolved_issues` | Contact Center Supervisor | `{}` | `Does Alpine Parka run true to size?`; `runs one size small`; `18 minutes` |
| `OCE-07` | `channel_recommendation` | Digital Engagement Manager | `{}` | `82% open`; `SMS confirmation`; `7-9 PM` |
| `OCE-08` | `proactive_plan` | Digital Engagement Manager | `{}` | `Proactive Engagement Plan (Draft)`; `Win-Back`; `Day 5: Personal stylist call` |
| `OCE-09` | `handoff_package` | Customer Experience Leader | `{}` | `Handoff Package: Sarah Mitchell`; `$2,400 LTV`; `no need to repeat anything` |

Routing rules:

- Match the user request to the operation shown above even when the operation name is not stated.
- Use only the exact argument identifiers in the companion records; never fabricate an ID.
- Keep the requested persona heading and the deterministic operation heading exactly as captured.
- When an argument is omitted in a locked case, follow the complete captured reference response below rather than asking for production data.
- If an unknown identifier is supplied, stop and request a valid synthetic identifier; do not approximate.

## Exact no-side-effect boundary

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

Never stitch identities across devices; infer sensitive traits; contact a person; send or schedule a message; create an offer, code, or reward; transfer a customer; or complete a purchase. Win-back messages, opening lines, and handoff scripts are drafts for the team.

Every answer is a draft, scenario, informational summary, or recommendation for
authorized human review. Never claim an action was sent, scheduled, approved,
issued, reserved, processed, fulfilled, or completed.

## Locked deterministic reference responses

These blocks are copied exactly from `agent_logs` in the persisted strict-isolation
capture. They establish required headings, names, identifiers, values, statuses,
dates, calculations, caveats, and boundary language for file-only reproduction.

### `OCE-01` — `channel_performance`

- Persona: **Customer Experience Leader**
- Prompt: As Customer Experience Leader, compare the aggregate channel picture and call out continuity risks.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Synthetic Channel Performance (30-Day)

**Total Revenue:** $12,957,500
**Total Conversions:** 65,020

| Channel | Sessions | Conversions | CVR | Revenue | Cost | ROAS |
|---|---|---|---|---|---|---|
| Email | 145,000 | 4,350 | 3.0% | $870,000 | $12,500 | 69.6x |
| Sms | 62,000 | 1,860 | 3.0% | $325,500 | $8,200 | 39.7x |
| Social Media | 230,000 | 2,760 | 1.2% | $552,000 | $45,000 | 12.27x |
| Web Organic | 480,000 | 9,600 | 2.0% | $1,920,000 | $18,000 | 106.67x |
| Web Paid | 185,000 | 5,550 | 3.0% | $1,110,000 | $95,000 | 11.68x |
| Mobile App | 310,000 | 12,400 | 4.0% | $2,480,000 | $22,000 | 112.73x |
| In Store | 95,000 | 28,500 | 30.0% | $5,700,000 | $180,000 | 31.67x |

## Revenue Share by Channel

- Email: 6.7%
- Sms: 2.5%
- Social Media: 4.3%
- Web Organic: 14.8%
- Web Paid: 8.6%
- Mobile App: 19.1%
- In Store: 44.0%
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Aggregate Customer Journey Analysis

## Discovery to Purchase

- **Avg Duration:** 14 days
- **Avg Touchpoints:** 5
- **Conversion Rate:** 3.2%

**Touchpoint Sequence:**

1. Social Media Ad ->
2. Website Browse ->
3. Email Signup ->
4. Email Promo ->
5. Website Purchase

## Repeat Purchase

- **Avg Duration:** 3 days
- **Avg Touchpoints:** 3
- **Conversion Rate:** 18.5%

**Touchpoint Sequence:**

1. Email Promo ->
2. Mobile App Browse ->
3. Mobile App Purchase

## Win-Back

- **Avg Duration:** 21 days
- **Avg Touchpoints:** 4
- **Conversion Rate:** 8.4%

**Touchpoint Sequence:**

1. Email Winback ->
2. Sms Offer ->
3. Website Browse ->
4. Website Purchase

## Impulse Purchase

- **Avg Duration:** 0 days
- **Avg Touchpoints:** 2
- **Conversion Rate:** 1.8%

**Touchpoint Sequence:**

1. Social Media Ad ->
2. Website Purchase

## Journey Optimization Opportunities

- **Discovery:** Shorten path by enabling social commerce checkout
- **Repeat:** Leverage push notifications for faster re-engagement
- **Win-Back:** Test a consented service reminder window in a controlled experiment
- **Impulse:** Optimize social ad creative for direct conversion
```

### `OCE-02` — `journey_analysis`

- Persona: **Contact Center Supervisor**
- Prompt: As Contact Center Supervisor, identify handoff friction across the aggregate journey archetypes.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Contact Center Supervisor
**Role focus:** handoff friction, unresolved needs, and service consistency

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Aggregate Customer Journey Analysis

## Discovery to Purchase

- **Avg Duration:** 14 days
- **Avg Touchpoints:** 5
- **Conversion Rate:** 3.2%

**Touchpoint Sequence:**

1. Social Media Ad ->
2. Website Browse ->
3. Email Signup ->
4. Email Promo ->
5. Website Purchase

## Repeat Purchase

- **Avg Duration:** 3 days
- **Avg Touchpoints:** 3
- **Conversion Rate:** 18.5%

**Touchpoint Sequence:**

1. Email Promo ->
2. Mobile App Browse ->
3. Mobile App Purchase

## Win-Back

- **Avg Duration:** 21 days
- **Avg Touchpoints:** 4
- **Conversion Rate:** 8.4%

**Touchpoint Sequence:**

1. Email Winback ->
2. Sms Offer ->
3. Website Browse ->
4. Website Purchase

## Impulse Purchase

- **Avg Duration:** 0 days
- **Avg Touchpoints:** 2
- **Conversion Rate:** 1.8%

**Touchpoint Sequence:**

1. Social Media Ad ->
2. Website Purchase

## Journey Optimization Opportunities

- **Discovery:** Shorten path by enabling social commerce checkout
- **Repeat:** Leverage push notifications for faster re-engagement
- **Win-Back:** Test a consented service reminder window in a controlled experiment
- **Impulse:** Optimize social ad creative for direct conversion
```

### `OCE-03` — `engagement_optimization`

- Persona: **Digital Engagement Manager**
- Prompt: As Digital Engagement Manager, propose consent-aware experiments with frequency caps and no outreach.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Digital Engagement Manager
**Role focus:** channel strategy, consent, and measurement

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Draft Engagement Optimization Report

## Channel Efficiency Ranking

| Rank | Channel | ROAS | CVR | Bounce Rate | Recommendation |
|---|---|---|---|---|---|
| 1 | Mobile App | 112.73x | 4.0% | 12.0% | Scale investment |
| 2 | Web Organic | 106.67x | 2.0% | 35.0% | Scale investment |
| 3 | Email | 69.6x | 3.0% | 18.5% | Scale investment |
| 4 | Sms | 39.7x | 3.0% | 5.2% | Optimize spend |
| 5 | In Store | 31.67x | 30.0% | 0% | Optimize spend |
| 6 | Social Media | 12.27x | 1.2% | 42.0% | Optimize spend |
| 7 | Web Paid | 11.68x | 3.0% | 28.0% | Optimize spend |

**Total Marketing Spend:** $380,700
**Total Revenue:** $12,957,500
**Blended ROAS:** 34.0x

## Optimization Actions

1. Model a budget shift from social media to consented app engagement
2. Ask only for optional preferences needed to improve service
3. Design an A/B test for the web-paid checkout journey
4. Review contact frequency caps across consented aggregate cohorts
```

### `OCE-04` — `campaign_attribution`

- Persona: **Digital Engagement Manager**
- Prompt: As Digital Engagement Manager, summarize the synthetic attribution view and avoid causal claims.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Digital Engagement Manager
**Role focus:** channel strategy, consent, and measurement

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Synthetic Campaign Attribution Report

| Campaign | Channel | Conversions | Revenue | Cost | ROI |
|---|---|---|---|---|---|
| Spring Collection Launch (CAMP-301) | Email | 2,250 | $450,000 | $5,000 | 8900.0% |
| Flash Sale — 48 Hours (CAMP-302) | Sms | 3,600 | $540,000 | $6,000 | 8900.0% |
| Influencer Partnership (CAMP-303) | Social Media | 1,700 | $340,000 | $35,000 | 871.4% |
| Google Shopping Ads (CAMP-304) | Web Paid | 2,700 | $540,000 | $42,000 | 1185.7% |
| App Push — Loyalty Members (CAMP-305) | Mobile App | 5,100 | $765,000 | $2,000 | 38150.0% |

**Total Campaign Revenue:** $2,635,000
**Total Campaign Cost:** $90,000
**Overall Campaign ROI:** 2827.8%

## Campaign Detail

### Spring Collection Launch (CAMP-301)

- Sent: 250,000 | Opens: 62,500 (25.0%) | Clicks: 18,750 (7.5%)
- Conversions: 2,250 (12.0% click-to-conversion)
- Revenue: $450,000 | Cost: $5,000

### Flash Sale — 48 Hours (CAMP-302)

- Sent: 120,000 | Opens: 115,200 (96.0%) | Clicks: 24,000 (20.0%)
- Conversions: 3,600 (15.0% click-to-conversion)
- Revenue: $540,000 | Cost: $6,000

### Influencer Partnership (CAMP-303)

- Clicks: 85,000
- Conversions: 1,700 (2.0% click-to-conversion)
- Revenue: $340,000 | Cost: $35,000

### Google Shopping Ads (CAMP-304)

- Clicks: 45,000
- Conversions: 2,700 (6.0% click-to-conversion)
- Revenue: $540,000 | Cost: $42,000

### App Push — Loyalty Members (CAMP-305)

- Sent: 85,000 | Opens: 42,500 (50.0%) | Clicks: 17,000 (20.0%)
- Conversions: 5,100 (30.0% click-to-conversion)
- Revenue: $765,000 | Cost: $2,000
```

### `OCE-05` — `customer_journey`

- Persona: **Contact Center Supervisor**
- Prompt: Show me the customer's journey across all channels and help me pick up where they left off.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Customer Journey: Sarah Mitchell

Sarah Mitchell's journey across 5 channels unified. She's been trying to buy for 3 days.

| Day | Channel | Action | Issue |
|---|---|---|---|
| 3 days ago | Mobile app | Checkout started | Payment declined |
| 2 days ago | Chat | Sizing question | Disconnected |
| Yesterday | Email | Cart reminder | No action |
| Today | Phone | Support call | Currently holding |

**Channel Preferences (30 Days):**
- Mobile app: 12 interactions (primary)
- Website: 8 interactions (secondary)
- Chat: 3 interactions (frustrated)

**Current:** $289 cart (Alpine Parka), 3 days old. Issues: payment declined, sizing unanswered. Mood: likely frustrated (declined payment, chat disconnect).

Source: [All Channels + CDP]

See full conversation context?
```

### `OCE-06` — `unresolved_issues`

- Persona: **Contact Center Supervisor**
- Prompt: Show me what Sarah has been asking about and where we dropped the ball.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Unresolved Questions: Sarah Mitchell

Context retrieved. 2 unresolved questions from the chat disconnect.

**Chat Session (2 Days Ago):**
- Q "Does Alpine Parka run true to size?" -> Agent: "Let me check..." (disconnected)
- Q "Do you have it in navy?" -> Never answered

| Issue | Status | Impact |
|---|---|---|
| Sizing guidance | Unanswered | Blocking purchase |
| Color availability | Unanswered | Blocking purchase |
| Payment | Card declined | Needs resolution |

**Context:** She's spent 18 minutes trying to buy. Likely frustrated about the chat disconnect.

**Draft opening:** "Hi Sarah, I see you've been trying to order the Alpine Parka. I apologize for the disconnect - let me answer your questions and complete this order."

**Ready answers:** Alpine Parka runs one size small. Navy in stock, S-XL.

Source: [Chat Logs + Inventory]

Get optimal channel strategy?
```

### `OCE-07` — `channel_recommendation`

- Persona: **Digital Engagement Manager**
- Prompt: Which channel and timing should we use to reach Sarah for this issue and in the future?
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Channel Strategy: Sarah Mitchell

Channel analysis: she prefers mobile/SMS, not email.

| Channel | Engagement | Best For |
|---|---|---|
| Mobile push | 82% open | Urgent updates |
| SMS | 76% response | Order status |
| Email | 34% open | Avoid urgency |
| Chat | Frustrated | Avoid short-term |

**For This Issue:** Phone (resolve now) -> SMS confirmation. Avoid email.

**Future Strategy:**
- Order updates: SMS (real-time)
- Promotions: Mobile push at 10 AM
- Service: Phone callback (avoid chat)

**Signals:** Responds to urgency, values fit guidance, peak engagement 7-9 PM.

Source: [Engagement Analytics]

Show proactive engagement opportunities?
```

### `OCE-08` — `proactive_plan`

- Persona: **Digital Engagement Manager**
- Prompt: What proactive engagement should we plan for Sarah after this purchase, or if she doesn't convert?
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Proactive Engagement Plan (Draft): Sarah Mitchell

Proactive engagement opportunities identified; every message is a draft for approval.

**Immediate (After Purchase):**
- Order complete: Size guide via SMS
- Delivery day: Styling tips via mobile push (82% engagement)
- 7 days post: Review request in-app

**Upcoming:**
- 3 days: Winter accessories bundle offer
- 6 weeks: Birthday loyalty bonus
- 8 weeks: Spring preview early access

**Win-Back (If She Doesn't Convert):**
- Hour 1: SMS "Your coat is waiting"
- Hour 4: Mobile push "Low stock"
- Day 2: SMS 10% off code
- Day 5: Personal stylist call

**Avoid:** Email campaigns (34% open), chat offers (negative history), generic messaging.

Source: [Behavioral Analytics]

Prepare a seamless handoff?
```

### `OCE-09` — `handoff_package`

- Persona: **Customer Experience Leader**
- Prompt: Prepare the context in case Sarah needs to be handed off to another agent.
- Exact arguments: `{}`

```markdown
[OmnichannelEngagementAgent] **Prepared for:** Customer Experience Leader
**Role focus:** cross-channel continuity, service quality, and governance

> Synthetic records. Recommendations only; uses only the consented service record already linked to the customer; no identity stitching, sensitive profiling, outreach, message, offer, reward, transfer, or purchase action occurs.

# Handoff Package: Sarah Mitchell

Handoff package prepared with full context (ready for you to transfer).

**Quick Context (Any Agent):**
- Sarah Mitchell (Gold, $2,400 LTV)
- Trying to buy Alpine Parka x 3 days
- Blockers: Sizing (resolved), payment (in progress)
- Mood: Previously frustrated, now engaged

**Transfer Context Ready:**
- Payments: Card decline history, alternatives
- Styling: Size preferences, past purchases
- Store pickup: Location, inventory
- Loyalty: Points, tier benefits

**Script:** "I'm connecting you with [Name]. I've shared your complete history - no need to repeat anything."

**Attach on transfer:** CRM note, cart link, conversation summary.

Source: [Context Store + Routing]

See session impact?
```