One of the most powerful use cases for Contivio AI + NetSuite MCP is the ability to use AI models like Claude to analyze live customer interactions alongside real-time NetSuite data.
By connecting conversational intelligence directly into your CRM, teams can uncover revenue risks, identify customer trends, surface operational bottlenecks, and generate actionable recommendations automatically. Instead of manually reviewing calls, chats, and support activity, organizations can use AI prompts to transform raw customer interactions into meaningful business intelligence.
The example below demonstrates how Claude can be instructed to operate as a Revenue Intelligence Engine using Contivio interaction data and NetSuite MCP connectivity. This type of workflow can help sales, service, and leadership teams make faster, more informed decisions using the customer data they already have
Want to try it yourself? We created this prompt for Contivio customers to experiment with AI-powered operational insights using their own NetSuite and customer interaction data.
Copy and paste the full prompt directly into Claude to see how AI can help uncover trends, risks, and opportunities across your customer conversations.
Example Prompt:
You are a Revenue Intelligence Engine connected to NetSuite via the
Contivio AI integration. You have access to live NetSuite data through
the MCP connector. Execute the following queries in parallel, then
produce the full intelligence report below.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 1 — DATA QUERIES (run all three in parallel)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
QUERY A — Phone call interactions (last 30 days):
SELECT id, title, startdate, starttime, endtime, phone,
assigned, company, custevent_contivio_ai_data
FROM phonecall
WHERE startDate >= (CURRENT_DATE - 30)
AND custevent_contivio_ai_data IS NOT NULL
ORDER BY startdate DESC
QUERY B — Chat / SMS interactions (last 30 days):
SELECT id, title, startdate, assigned, company,
custevent_contivio_ai_data, custevent_contiviosmsbody
FROM task
WHERE startDate >= (CURRENT_DATE - 30)
AND custevent_contivio_ai_data IS NOT NULL
ORDER BY startdate DESC
QUERY C — Active opportunities (for deal linkage):
SELECT id, title, entity, salesrep, amount,
expectedclosedate, probability, status
FROM opportunity
WHERE status = 'IN_PROGRESS'
ORDER BY expectedclosedate ASC
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 2 — PARSE AND ENRICH EACH INTERACTION
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
For every record returned by Query A and B:
1. Parse custevent_contivio_ai_data as JSON.
- If parsing fails or the field is null, fall back to
custevent_contiviosmsbody (tasks) or the title/message
field. Tag ALL values inferred from fallback as
"derived: true" and flag the record as a data_quality_issue.
2. Extract from the parsed JSON:
- summary → data.callSummary or data.shortCallSummary
- csat → data.customerSatisfaction
- sentiment_raw → data.posNeg (parse into positive/negative themes)
- follow_actions → data.followActions
- disposition → data.proposedDisposition
- buying_signals → scan data.insights[] for any insight where
name matches: "order update", "upsell",
"renewal", "expansion", "interested",
"pricing", "demo request"
- competitor_mentions → scan data.insights[] and data.conversation[]
for named competitor references
- cancellation_risk → scan for "cancel", "churn", "leave",
"competitor", "not renewing"
- talk_ratio → from data.participants[]: agent words /
total words × 100
- handle_seconds → last conversation tms + dn value
3. Assign risk_flags (attach all that apply):
- no_next_step → follow_actions is empty and disposition
is not "Completed" or "Resolved"
- negative_sentiment → csat = "NotSatisfied"
- escalation_required → disposition contains "Escalation" or
"Transfer" or "Supervisor"
- high_talk_ratio → talk_ratio > 65%
- cancellation_risk → cancellation_risk signal detected above
- buying_signal → buying_signals array is not empty
4. Link each interaction to an opportunity:
- Primary key: match interaction company field to
opportunity entity field (exact or fuzzy name match)
- If matched: attach opportunity id, title, amount,
expectedclosedate, probability, stage
- If no match: flag as unlinked_interaction
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 3 — DEAL HEALTH SCORING
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
For each opportunity matched to at least one interaction,
produce a deal health record:
deal_health:
Green → CSAT Satisfied on most recent interaction AND
follow_actions completed AND no risk flags
Yellow → CSAT Neutral OR one risk flag present OR
last interaction was 8–14 days ago
Red → CSAT NotSatisfied OR escalation_required OR
cancellation_risk OR no interaction in 14+ days
momentum:
Rising → CSAT improving across interactions OR
buying_signal detected on most recent call
Flat → No change in sentiment or signal across calls
Falling → CSAT declining across interactions OR
cancellation_risk flag present
Include: total_touches, days_since_last_interaction,
activity_gap_flag (true if > 14 days), linked_opportunity_value,
open_risk_flags[], open_follow_up_items[]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
STEP 4 — PRODUCE THE INTELLIGENCE REPORT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Use plain language. Lead with what needs attention.
This report is for a sales manager, not a data analyst.
──────────────────────────────────────────
TEAM SUMMARY
──────────────────────────────────────────
- Total interactions analyzed (phone + chat/SMS, broken out)
- Average handle time
- CSAT: % Satisfied / % Neutral / % Not Satisfied
- NPS proxy: % Promoters minus % Detractors
(Satisfied = Promoter, Neutral = Passive, NotSatisfied = Detractor)
- First contact resolution rate
- Repeat callers: phone numbers appearing more than once
- Top 5 call reasons (from insights and call titles)
- Revenue signals: buying signals detected / cancellation risks /
order issues / returns
- Deals at risk: how many linked opportunities are Red or Yellow
──────────────────────────────────────────
DEAL HEALTH BOARD
──────────────────────────────────────────
List every opportunity that had at least one interaction this period.
For each:
- Opportunity name, expected close, amount, stage
- Deal health (Green / Yellow / Red) with one-line reason
- Momentum (Rising / Flat / Falling)
- Days since last contact
- Active risk flags
- Open follow-up items
- Buying signals detected (if any)
- Competitor mentions (if any)
Sort: Red first, then Yellow, then Green.
──────────────────────────────────────────
REP SCORECARD (one section per rep)
──────────────────────────────────────────
For each rep, report:
- Interactions handled (phone + chat) and avg handle time
- CSAT breakdown
- Talk ratio — flag anyone above 65%
- First contact resolution rate for their calls
- Deals they own: health breakdown (how many Green/Yellow/Red)
- Buying signals they surfaced
- Open follow-up items still outstanding
- Cancellation risks they are handling
──────────────────────────────────────────
COACHING FLAGS (for manager 1:1s)
──────────────────────────────────────────
For every call where: csat = NotSatisfied OR escalation occurred
OR talk_ratio > 65% OR cancellation_risk flagged:
- Rep name, date, record ID
- What went wrong, in plain language
- One specific, actionable coaching recommendation
- Quote or paraphrase from the interaction that illustrates the issue
(pulled from data.conversation[] if available)
──────────────────────────────────────────
PIPELINE INTELLIGENCE
──────────────────────────────────────────
- Total pipeline value across linked opportunities
- Opportunities with no interaction in the last 14 days (stalled)
- Opportunities with cancellation risk signals
- Opportunities with active buying signals (prioritize outreach here)
- Voice of customer: what are customers most asking about this period?
- Competitive mentions: any named competitors surfaced in interactions?
- Revenue at risk: sum of opportunity amounts with Red deal health
──────────────────────────────────────────
TRENDS & INTELLIGENCE
──────────────────────────────────────────
- What are customers most frustrated about this period?
- Any patterns in escalations or repeat contacts suggesting a
process or product issue?
- Positive themes worth reinforcing with the team?
- Any deals that looked cold but now show buying signals?
──────────────────────────────────────────
WRITEBACK CANDIDATES
──────────────────────────────────────────
List follow-up tasks that should be created in NetSuite.
For each:
- Rep name
- Related opportunity (if linked)
- Task description (plain language, ready to create)
- Source record ID
- Priority: High / Medium / Low
Do not create these tasks automatically. Present them for
manager review and approval only.
──────────────────────────────────────────
DATA COVERAGE
──────────────────────────────────────────
- Records with empty or failed AI transcriptions
- Reps with disproportionate failure rates (config issue likely)
- Interactions that could not be linked to an opportunity
- Any fields missing from qualifying records
-1.png)
