The TeamPulse recommendations engine for SuiteCRM transforms your team’s behavioral signals into a structured coaching action inbox. Instead of manually comparing dashboards to figure out who needs help and what you should do, every coaching need is surfaced as a prioritized card – with the evidence already attached. Managers spend less time diagnosing and more time actually coaching.
The Recommendations view is located in the Analytics group of the TeamPulse sidebar, identified by the lightbulb icon.
[Image Placeholder: TeamPulse Recommendations view showing priority-tiered coaching cards with evidence and action buttons]
How the Recommendations Engine Works
The engine derives coaching priorities directly from the same data that powers the rest of TeamPulse – active signals, current health scores, and momentum direction. It does not require a separate API call or real-time AI request to generate recommendations. The InsightEngine processes signal patterns and score trends locally, then outputs a ranked list of coaching actions.
If you have BYOK AI configured (OpenAI, Azure OpenAI, or a compatible endpoint), the engine can supplement recommendations with richer natural-language descriptions. Without BYOK, recommendations are still fully functional – the priority, evidence, and suggested action are all generated from structured signal data.
The engine evaluates three inputs for each rep:
- Current health score – a weighted composite of activity volume, deal hygiene, follow-up rate, engagement quality, and close rate consistency
- Active signals – specific behavioral patterns such as stale deals, missed follow-ups, or declining call volume
- Momentum direction – whether the rep’s score is trending up, stable, or declining over the rolling window
The Four Priority Tiers
Every recommendation is assigned to one of four tiers based on the severity of the underlying pattern.
| Tier | Trigger Condition | Recommended Action |
|---|---|---|
| Critical | Health score below 60, or significant declining momentum detected | Schedule an immediate 1:1 coaching session |
| High | Recurring deal hygiene gaps or persistent activity shortfalls | Assign deal cleanup tasks; review pipeline together |
| Medium / Opportunity | Minor process gaps or one-off signal triggers | Standard coaching touchpoint; process optimization discussion |
| Recognize | Top performer maintaining a high score streak | Publicly praise in the next team meeting; use as a peer model |
The Recognize tier is positive reinforcement, not a problem flag. Acknowledging high performers consistently is as important as addressing underperformers, and the engine makes sure neither gets overlooked.
Anatomy of a Recommendation Card
Each card in the inbox contains everything a manager needs to act – without clicking through to another view.
- Rep name and current score – identifies who the recommendation is about and their present health score
- Priority badge – color-coded tier label: Critical, High, Medium, Opportunity, or Recognize
- Recommended action – a plain-English coaching instruction, for example: “Schedule a 1:1 to review pipeline and address stale deals”
- Evidence – the specific data point that triggered this recommendation, for example: “6 of 12 deals are stale” or “Call volume dropped 40% over the last two weeks”
- Action buttons – three options to resolve or dismiss the card (detailed below)
The evidence field is the most actionable part of the card. It removes ambiguity from the coaching conversation – you know exactly what to discuss before the 1:1 begins.
The PageInsight Accordion
At the top of the Recommendations view, the PageInsight accordion provides an AI-generated or InsightEngine summary of your team’s overall coaching situation for the current period. It answers questions like: which reps need the most immediate attention, whether a pattern is systemic across the team, and what the dominant signal type is driving the inbox today.
Expand the accordion to read the full situational summary, then collapse it to work through individual cards. The PageInsight updates when the underlying data refreshes – it reflects the same snapshot as the cards below it.
Working the Coaching Inbox
A recommended workflow for processing the inbox efficiently:
- Filter by Critical – use the priority filter at the top of the view to show only Critical-tier cards first. These are the reps who need immediate attention.
- Review the evidence – read the evidence field on each Critical card before acting. The specific data point tells you what to address in the coaching session.
- Send a Coaching Note or Assign a Task – click Send Coaching Note to open the Actions Hub with the rep pre-selected, or assign a deal cleanup task directly from the card.
- Mark Done – once you have acted on a recommendation, click Mark Done to clear it from the inbox. This tracks that coaching action as complete for the current cycle.
- Repeat for High-tier cards, then review Medium and Opportunity cards for lighter-touch coaching.
- Do not skip Recognize cards – acknowledging top performers takes thirty seconds and has a measurable effect on retention.
Mark Done vs. Dismiss
The two removal actions on each card have distinct meanings:
- Mark Done – you have taken the recommended coaching action. The card is removed from the inbox, and the coaching action is recorded as complete for this cycle. If the underlying issue persists into the next scoring window, the engine may regenerate a recommendation.
- Dismiss – the recommendation is not relevant this cycle. The card is hidden without recording a coaching action. Dismiss is appropriate when you have context the engine does not – for example, a rep is on planned leave, or you already handled this in a meeting before the recommendation was generated.
Dismiss does not permanently remove a recommendation type. If the signal pattern recurs in the next evaluation window, the engine will surface a new recommendation for that rep.
Next Steps: Document the Session in the Journal
After acting on a recommendation, use the Journal to record notes from the coaching session. The Journal is accessible from the Actions Hub and allows managers to log what was discussed, what was agreed, and what follow-up is expected. Linking journal entries to specific reps creates a longitudinal coaching record that supports performance reviews and tracks whether interventions are working over time.
Frequently Asked Questions
What triggers a Critical recommendation?
A Critical recommendation is generated when a rep’s current health score falls below 60, or when the engine detects a significant declining momentum trend – meaning the score has dropped meaningfully over the recent rolling window, even if it has not yet crossed the 60-point threshold. Both conditions indicate a rep who needs immediate attention rather than a routine check-in.
Does Dismiss permanently remove a recommendation?
No. Dismiss hides the current card for this cycle. If the same rep continues to trigger the same signal pattern in the next scoring window, the engine will generate a new recommendation. Dismiss is cycle-scoped, not permanent. Use Mark Done when you have taken action; use Dismiss only when the recommendation is genuinely not applicable this period.
Do I need BYOK AI configured to use Recommendations?
No. The Recommendations Engine runs entirely on structured signal and score data – no external AI call is required. BYOK AI enhances the natural-language quality of recommendation descriptions and the PageInsight summary, but the priority tiers, evidence, and suggested actions are all available without it.