New features
Nexie assistant
The built-in conversational AI assistant, available as a chat surface throughout NEXOS CORE. Ask plain-English questions about your tickets, projects, invoices, clients, purchase orders, or security posture and Nexie answers from your live database. Queries are read-only, scoped to your tenant, and row-limited; follow-up questions carry recent turns as conversation history, and each answer reports its AI cost. See Nexie assistant.NexusPulse (SMART Numbers)
A TruMethods-aligned MSP KPI engine. Each month NexusPulse auto-collects raw inputs — seats, agreements, tickets, revenue, expenses, sales activity — and computes dozens of derived KPIs including net profit %, MRR margin, Rule of 40, AISP, RHEM, reactive tickets per seat, close rates, and Enterprise Value. Each KPI is graded green / yellow / red against TruMethods benchmarks, and the analyzer emits prioritized findings with a plain-English diagnosis, dollar impact, and a recommended action that deep-links into the relevant tool. Inputs pull from contracts, helpdesk, CRM, and QuickBooks P&L, with manual overrides respected on the next run. See NexusPulse.Employee productivity analytics
Per-employee focus and app-usage analytics driven by NexusRMM agent telemetry. The agent reports 15-minute windows of foreground-app usage; EPA classifies each app (productive, communication, meeting, distraction, neutral), computes a 0–100 focus score per window, and rolls the data up into per-client dashboards, per-person drilldowns, rolling-average alerts, and AI coaching briefs for manager 1:1s. Collection is gated by a tenant-admin consent acknowledgement plus a per-person opt-in, and access is role-gated (Admin, Manager, C-Level; HR opt-in only). See Employee productivity analytics.Known limitations
- Nexie — read-only by design; write/DDL statements are rejected. Answers are generated from live queries, so spot-check numbers for high-stakes decisions. Requires a tenant Claude API key.
- NexusPulse — margin, profit, and expense KPIs require QuickBooks connected and accounts mapped; without it, expense categories fall back to 0 and those KPIs are unreliable. The monthly executive summary is rule-based, not AI-generated. Year-over-year metrics need a prior-year snapshot. Grading bands are fixed to TruMethods defaults, not tenant-tunable.
- EPA — app classification uses a fixed built-in list (substring matching); unknown apps are treated as neutral. Consent is a tenant-admin acknowledgement plus MSP-toggled per-contact opt-in, not a self-service employee consent — confirm your monitoring has the appropriate legal basis and employee notice for your jurisdiction. AI insights require a tenant Claude API key.