Source: Network Right CTO Meeting Distillation
This page distills a 2026-04-30 meeting with Network Right CTO Anthony Garcia plus public website research. It intentionally does not store the exact transcript.
Source Context
Network Right is a fractional IT and compliance provider for startups and growth companies. Public pages position it as a dedicated, human-led extension of the customer’s team, with Slack/Teams workflow integration, Jira-powered ticketing, onboarding/offboarding, asset management, SaaS management, network support, cybersecurity, vCISO/compliance support, and AI readiness.
Public website signals include 100K+ tickets handled by humans, 4.95/5 satisfaction/NPS messaging, 99% SLA adherence, pricing from 170/user/month for scale-up support, and customer stories around Clearbit remote onboarding/SaaS management and Scribe office/network buildout.
Meeting-Derived Operating Model
- Network Right frames itself as a premium alternative to a traditional MSP: it uses MSP discipline such as recurring service delivery, SOPs, runbooks, tooling efficiency, and help desk operations, but rejects the rigid standardized MSP model that forces every customer into the provider’s tool stack.
- The company works primarily with startups and aims to act as the customer’s internal IT and compliance team until the customer is ready to hire internal IT leadership. The provider is designed as a stepping stone, not necessarily a forever vendor.
- The internal team is about 40 people and serves roughly 80-90 customers. The meeting source described customer growth tracking by headcount and strategy, with IT roadmaps shaped by whether a customer is growing aggressively, preparing for enterprise maturity, pursuing acquisition, or trying to conserve capital.
- Network Right prefers customer-owned, best-of-breed startup stacks. Examples discussed included Google Workspace, Slack, Notion, Zoom, Kandji, CrowdStrike, SentinelOne, 1Password, and similar SaaS/endpoint/security tools.
- As customers become more complex, Network Right may recommend hiring an internal IT director while continuing to support execution, compliance, or transition work. This is materially different from MSPs that try to retain the whole account indefinitely. ^[inferred]
Bottlenecks And Pain Points
- The clearest pain is the absence of a modern multi-tenant MSP OS for flexible fractional IT providers that do not want to be acquired by Treeline or locked into legacy PSA/RMM tooling.
- Jira is usable as a backend ticketing system but is not purpose-built for Network Right’s multi-customer, startup-stack, service-provider workflow.
- Legacy MSP platforms such as Kaseya and ConnectWise are perceived as rigid, Windows/Microsoft-heavy, and optimized for hyper-standardized MSP delivery rather than startup-stack flexibility.
- Modern AI ITSM products such as Serval, Console, and Risotto were viewed as closer to the right direction, but not designed for the consultancy/MSP multi-tenancy pattern.
- Several AI service-desk products rely heavily on Okta or a single-customer tenant assumption. That breaks down when a provider needs to connect many customers, each with its own HRIS, IdP, SaaS apps, MDM, EDR, documentation, and access policies.
- The hardest AI failure mode is cross-tenant context bleed. A model or retrieval system that answers a customer’s question using another customer’s Notion documentation, Wi-Fi password, or runbook is unacceptable.
- In their tests, AI support flows escalated too often and asked too many clarifying questions when customer Slack requests were vague. End users often expect the support partner to infer context and “just get it,” which is exactly where current AI agents can feel brittle. ^[inferred]
- Maintaining role-based access matrices, customer-specific runbooks, and documentation freshness remains laborious. Tools like AccessOwl can help with API-driven SaaS access but do not remove the operational setup burden.
Product Gap
The strongest product gap from the meeting is not another single-tenant AI service desk. It is a multi-tenant AI operating layer for modern fractional IT providers:
- Tenant-contained AI context per customer, with hard data boundaries.
- Customer-specific documentation, runbooks, access matrices, and tool graphs.
- Slack/Teams intake that can infer likely intent from customer context without annoying users with excessive probing.
- Jira/Notion/HRIS/IdP/MDM/EDR/SaaS connectors that support many independent customer tenants.
- Human escalation designed around provider workflows: technician assignment, SLA, account context, customer history, and client-specific policies.
- A single source of truth for MSP partnerships and customer-owned tools, without forcing every customer onto the same stack.
This is an underrepresented wedge because the best-funded AI ITSM products mostly target internal IT teams, while Treeline offers a portfolio/acquisition-style service-led model rather than selling an open OS to independent modern MSPs. ^[inferred]
Strategic Implications For Init Intelligence
- Competing directly with Serval, Console, and Risotto on single-tenant AI ITSM is an uphill battle.
- Building for modern fractional IT providers may be a sharper wedge because the buyer has immediate ticket volume, repeated cross-customer patterns, and acute pain with multi-tenant context isolation.
- A Network Right-like customer could be both design partner and channel: Init Intelligence could learn from a real operator, ship leverage into their workflows, then decide whether to sell MSP enablement software, partner on managed delivery, or pursue a deeper strategic relationship. ^[inferred]
- The core product requirement is not generic “AI answers tickets”; it is safe, tenant-scoped, action-capable context and workflow management across many customer environments.
- The product must respect the human-led service posture. The AI should compress routine work and context gathering, not pretend that complex IT judgment, hardware logistics, and customer relationship work disappear.
Follow-Up Questions
- What are Network Right’s top 10 recurring ticket categories by volume and human minutes?
- Which ticket types are safe for autonomous action today, which require human approval, and which should remain human-only?
- What customer context is needed for a technician to resolve a vague Slack request quickly?
- Where does Jira fail: intake, routing, SLA, customer segmentation, reporting, automations, or integration context?
- What is the minimum secure tenant isolation model they would trust for AI over customer documentation and credentials?
- Which systems should be first-class connectors for a modern startup-stack MSP OS?
- What would let Network Right double customers without doubling delivery headcount?