Insight DURUM
DURUM.ai continuously analyzes your data and automatically generates actionable recommendations to improve your marketing, sales, and operations performance.
Permission Required
The dedicated /ai-insights page requires the ai_insights permission or a plan that includes it. Recommendations that appear directly in dashboard tabs are visible to all users.
Recommendation Types
Each recommendation is categorized by the domain it concerns.
| Category | Example Recommendations |
|---|---|
| Strategy | Budget adjustments, redistribution across campaigns, priority changes |
| Spend | Waste identification, ads to cut, spend optimization |
| Creative | Ads to refresh, fatigued creatives, top-performing visuals |
| Funnel | Bottlenecks, stages to improve, abnormal conversion rates |
| Sales | Rep coaching, prospect distribution, no-show follow-up |
| Scheduling | Best days and times to schedule sales meetings |
Where to See Recommendations
Recommendations are accessible in three places in DURUM.ai:
1. In Each Dashboard Tab
Each dashboard tab (CEO View, Marketing, Sales Reps, Sales, Finance) includes a Recommendations panel adapted to the tab's context.
| Tab | Displayed Recommendations |
|---|---|
| CEO View | Strategic overview, priority alerts |
| Marketing | Budget, creatives, campaign performance |
| Sales Reps | Coaching, workload distribution, no-shows |
| Sales | Pipeline, conversion, at-risk opportunities |
| Finance | Profitability, cost optimization, projections |
Contextual Recommendations
Dashboard recommendations change automatically when you modify the period or filters. If you select a specific campaign, recommendations will focus on that particular campaign.
2. Dedicated /ai-insights Page
The dedicated page groups all active recommendations in one place, across all categories. It's the ideal spot for a complete view of what DURUM.ai suggests optimizing.
3. Slack Reports
If you've enabled Slack notifications, the most important recommendations are included in your daily, weekly, and monthly reports. See the Notification Configuration section to enable this feature.
Anatomy of a Recommendation
Each recommendation generated by DURUM.ai contains the following elements:
| Element | Description |
|---|---|
| Title | Short summary of the recommendation |
| Category | Relevant domain (strategy, spend, creative, etc.) |
| Description | Detailed explanation of the observation and suggestion |
| Confidence level | Degree of certainty of the recommendation |
| Underlying data | Numbers and metrics supporting the recommendation |
Confidence Levels
| Level | Meaning |
|---|---|
| High | Recommendation based on solid data and sufficient volume |
| Medium | Trend observed, but data volume is still limited |
| Low | Signal detected, but data is insufficient to conclude with certainty |
INFO
A high-confidence recommendation doesn't mean you must apply it. It means the data clearly supports the observation. The final decision is always yours.
Recommendations & Actions
Recommendations generated by DURUM.ai are informational. They identify trends, anomalies, and opportunities in your data, but it's up to you and your account manager to decide if and how to act.
Examples of recommendations and possible actions:
| Recommendation | Possible Action |
|---|---|
| "Campaign X has a CPA 3x above average" | Review targeting or creative for this campaign |
| "Rep Y's no-show rate is 40%" | Set up automated reminders or review the confirmation process |
| "Creatives launched more than 30 days ago show fatigue" | Prepare new visuals to refresh ads |
| "Booking rate increases 25% on Tuesdays" | Concentrate appointment slots on Tuesdays |
Getting the Most from Recommendations
Check recommendations at least once a week. Recommendations evolve with your data: a problem flagged Monday may be resolved by Friday if metrics improve. Recommendations disappear automatically when the observation is no longer relevant.