AI Agents
Each agent operates over a defined source set with explicit weights. MCP-tagged agents call external tools; A2A-tagged agents negotiate with peer agents (Capacity ↔ Salesforce, Finance ↔ SAP).
Risk Sentinel
37 actions · 94% confidence
MCP
Monitors project health scores, blocker age, schedule slip, and milestone proximity. Generates escalation recommendations when composite risk crosses configurable thresholds.
Source systems
JiraMS ProjectPortal CSAT
Decision weights
Schedule slip25%
Blocker age25%
Milestone proximity20%
Sentiment15%
Trend velocity15%
Most recent rationaleComposite risk computed as weighted sum across 5 signals. Solara DGP triggered escalation when blocker-age signal exceeded 6-day threshold and trend velocity dropped 14pts in 48h.
Capacity Agent
22 actions · 88% confidence
MCPA2A
Analyses workforce availability across bandwidth, skills, and complexity. Matches resources to demand and forecasts gaps using CRM signals.
Source systems
Workday HCMJiraSalesforce CRM
Decision weights
Skills match40%
Bandwidth30%
Complexity fit30%
Most recent rationaleComposite fit = 0.30·BW + 0.40·Skills + 0.30·Complexity. R. Mehta scored 81 against Senior Cloud Architect demand: BW 65% (partial avail), Skills 91% (5/5 keywords), Complexity 88% (3 comparable engagements).
Finance Agent
18 actions · 91% confidence
MCP
Tracks budget actuals from SAP, calculates forecast at completion, and flags projects where burn deviates from plan. Triggers CR recommendations on budget impact.
Source systems
SAP S/4HANACR RegisterMS Project
Decision weights
Burn rate vs plan50%
Forecast variance30%
Open CR exposure20%
Most recent rationaleMeridian flagged: actuals $2.86M / budget $2.4M = 119% burn. Forecast $2.9M (+21% vs budget). Triangulated against MS Project EVM and SAP cost-center 4-week trend.
Scope Guard
11 actions · 86% confidence
Applies NLP classification to Teams messages, emails, and meeting notes to detect scope-expansion language. Auto-drafts CRs when signals reach confidence threshold.
Source systems
MS TeamsOutlookServiceNow
Decision weights
Lexical match40%
Speaker authority25%
Repetition20%
Cross-channel15%
Most recent rationaleCR-041 raised on 4 signals across Teams + email over 72h. Highest-confidence signal (91%): client architect requested 'additional data migration module' in Teams 9 May, repeated 11 May.