Nestack Agent Care
Transparent pricing for production AI agent care

Per agent. Per month.
No surprises.

Pay only for the agents actively running in production. Choose the care level based on the agent’s complexity, usage, and business impact.

Subscriptions

Agent Care tiers

Pick your agent’s complexity class to compare pricing, monitoring coverage, and operational support across each tier.

Answers, summarises or recommends without changing external systems.

$199/agent/mo

Care Essential

Low-risk internal agents

Onboarding $1,500 one-time

  • Continuous automated monitoring and alerting
  • LLM usage, latency and cost tracking
  • Monthly final-output evaluation run
  • Standard quality and safety scorecard
  • Basic trace review for failed sessions
  • Next-business-day incident acknowledgement
  • Quarterly operational review
★ Most popular

$299/agent/mo

Care Plus

Production agents handling customer or business workflows

Onboarding $2,500 one-time

  • Everything in Essential
  • Continuous quality, safety and drift checks
  • Output, step and selected trajectory evaluations
  • Industry-specific evaluation coverage
  • Human screening of up to 200 sessions per month
  • Detailed review of flagged high-risk sessions
  • Four-business-hour incident acknowledgement
  • Monthly evidence-linked improvement review
  • Targeted re-evaluation after material fixes
  • Cost-efficiency review across models, retries and tool usage

$499/agent/mo

Care Enterprise

Regulated, high-volume, or mission-critical agents

Onboarding $4,999+ one-time

  • Dedicated AgentOps lead
  • 24/7 critical-incident acknowledgement within 30 minutes
  • Full trace, tool-action and guardrail investigation
  • Custom output, step and trajectory evaluation suite
  • High-risk human-review workflows
  • Compliance-readiness operations support
  • Change and version-impact reviews
  • Evidence-linked verification before issue closure
  • Agent improvement and retraining recommendations
  • Leadership and governance reporting
How it's delivered

We run your observability — we don't sell you a platform

Agent Care works above your existing observability stack. We connect traces, evaluations, guardrail events, human feedback and incidents into one managed operating process. Platform subscription and usage charges remain separate and are billed by the selected vendor.

Observability setup and connector configurationTrace and span mappingEvaluation managementGuardrail and review workflowsIncident-response coordinationCost and drift analysisMonthly AgentOps reportingOngoing managed supportImprovement action trackingTargeted re-evaluation and verification
Onboarding

What your onboarding fee covers

A one-time setup that gets your agents instrumented, evaluated and alert-wired before the first monthly cycle begins. Each tier includes everything in the tier before it.

Essential onboarding

$1,500 one-time
  • Kickoff, agent inventory and complexity classification (Class A/B/C)
  • Agent role, owner, version, model, tools, knowledge sources and production environment documented
  • Observability connection and trace validation to confirm usable agent, LLM, tool, error, latency and cost data
  • Lifecycle visibility assessment across goal, retrieval, workflow, task, tool, LLM, evaluation, guardrail and business outcome layers
  • Evaluation profile defined for the agent, including critical workflows, success criteria, key metrics, thresholds and human-review triggers
  • Baseline end-to-end evaluation established, with selected step-level checks for critical workflows
  • Initial failure-mode mapping and first-response runbooks for the agent’s highest-priority risks
  • Alert routing, severity levels and escalation contacts configured
  • AgentOps scorecard configured for reliability, quality, safety, cost and performance tracking

Plus onboarding

$2,500 one-time

Everything in Essential, plus:

  • Up to 25 custom regression and risk evaluation cases based on historical incidents, high-volume workflows, policies and failure modes
  • Evaluation coverage expanded beyond final output to selected step-level and trajectory evaluations
  • Evaluation slices defined for critical task types, workflows, tool paths, risk levels or user cohorts
  • Regression cases created from historical incidents and known failure modes for future re-evaluation
  • Tool-action review covering permissions, write actions, approval requirements and failure handling
  • Guardrail and safe-mode review, including pause, restriction and recovery recommendations
  • Review-queue and escalation workflow configured for high-risk or low-confidence sessions
  • Alert integrations configured with supported communication and ticketing systems
  • Cost, latency, retry and reliability baseline established for key workflows
  • Historical incidents and top workflows mapped to retrieval, workflow, task, tool, LLM, evaluation or guardrail failure categories
  • Two live operational training sessions covering trace investigation, alerts, reviews and escalation

Enterprise onboarding

$4,999+ one-time

Everything in Plus, plus:

  • Full lifecycle observability coverage assessment across critical agents and workflows
  • Compliance-readiness operations review for relevant AI, data, privacy and human-oversight requirements
  • Shared-accountability mapping across client owners, model providers, tool providers and Nestack operations
  • SEV-1 incident-response tabletop exercise before go-live
  • Tool restriction, safe-mode, pause and recovery procedures defined and tested
  • Approval authority, escalation ownership and emergency decision rights documented
  • Agent, model, prompt, tool, retrieval, guardrail and evaluator versions baselined for future change-impact and regression analysis
  • Evidence-linked verification policy defined for material changes, including what must be re-evaluated before an issue is closed or autonomy is restored
  • Human and AI evaluator calibration approach established for high-impact or ambiguous cases
  • Critical workflow trajectory and guardrail coverage reviewed before production handover
  • Named AgentOps lead assigned for onboarding and 90-day success planning
  • Enterprise escalation matrix, response targets and communication procedures agreed
Before you commit

Free assessment for qualified production AI teams

One agent, a fixed sample of real sessions, five business days. No commitment — even if you never become a client, you keep the report.

1
Scope
One production agent
Fixed sample of real sessions, delivered within five business days
20
Session review
Up to 20 sessions screened
Automated checks with deeper human review of flagged outputs, steps and trajectories
5
Failure-mode mapping
Top 5 operational risks
Severity-rated and mapped to the affected AgentOps lifecycle layer
✓
Quality probes
Basic quality and prompt-risk checks
Run against a known-answer set
$
Cost & latency snapshot
Token spend and performance review
Cost, latency, and improvement opportunities
30
Report & readout
Short findings report + 30-minute call
Risks, gaps, and suggested monitoring approach
Commercial terms

Fair terms, published upfront

Discounts, guarantees, and billing — the same for every client, no negotiation required.

%
Volume discounts
10+ −10% · 25+ −15% · 50+ −20%
50+ includes a named TAM
12
Annual prepay
2 months free
Stacks with volume discounts
30
Cancellation
30 days' notice
On Essential and Plus
5%
Price protection
Increases capped at 5%/yr
Locked for your contract term
$
Currency
USD default
Regional currency invoicing available for eligible clients
★
Nestack-built agents
Onboarding fee waived
Includes 3 months of Care Plus at no additional cost

Care Essential

About you
Your agents
Drag the slider to update your agent count
Class Aread-only$199 /agent/mo
Class Bacts or writes$299 /agent/mo
Class Cregulated or mission-critical$398 /agent/mo
Care details
Compliance & logistics
Total:$597/month$199 /agent/mo × 3 agents

Thanks — your email is ready to send

We've opened a pre-filled email in your mail app. Hit send and we'll reply within one business day.

Add-ons

Bolt on exactly what you need

Compliance packs

Compliance readiness support for HIPAA & EU AI Act with leak monitoring, incident records & monthly assurance reports.

$500 /agent/mo each

Focused AI Red-Team Assessment

Security testing for prompt injection, safety bypasses, and data leakage, with a detailed findings report.

From $3,500 per assessment

Custom eval suite

50–150 tailored test cases with scoring criteria for a new workflow or major agent update.

From $1,500 one-time

Extra human review

Detailed human review of 100 additional agent sessions beyond your plan allowance.

$250 per 100 standard sessions

Fine-tuning management

Manage one fine-tuning cycle, from data prep and training to evaluation and rollout. Bulk prep and labeling extra.

$1,500 per model cycle

War-room support

Dedicated engineer for launch weeks: live dashboard, daily briefs. After-hours coverage priced separately.

$1,500 /week

AgentOps Observability Coverage Review

Deep review of trace coverage across goal, retrieval, workflow, task, tool, LLM, evaluation, guardrail and business outcome layers, with blind spots and instrumentation recommendations.

$2,500 one-time
Accountability

Support commitments you can measure.

MetricEssentialPlusEnterprise
Incident acknowledgement / triage startNext business day4 business hours30 minutes for critical incidents
Initial containment target3 business days1 business daywithin 8 hours
Monthly scorecard deliveryBy the 7th business dayBy the 5th business dayBy the 2nd business day
Post-incident reviewIncluded in monthly reviewWithin 5 business daysWithin 3 business days
Root-cause and corrective-action recordBasicDetailedDetailed + ownership mapping
Corrective-action verificationMonthly reviewRe-evaluate in 5 daysVerify before closure
Build offers

Fixed-price builds

AI Agent Readiness Assessment

2–3 weeks: agent opportunity audit, workflow mapping, ROI estimate, data readiness review, risk classification, observability requirements, and top 3 ranked AI agent use cases.

$2,500 fixed

First AI Agent in 6 Weeks

Production-ready AI agent for one defined workflow. Includes discovery, design, development, testing, deployment, baseline tracing, evaluation setup, guardrail logging and Agent Care handover.

$8,000 fixed

Dedicated AI Engineer

A dedicated AI engineer working with your team on agent development, integrations, evaluations, observability, guardrails and workflow improvements.

$2,500 USD per month

Payment on builds: 40% start · 40% midpoint milestone · 20% acceptance. Acceptance criteria written into the SOW before start.

Not sure which package fits?

Talk to us and we’ll help you choose the right plan for your agent.

Schedule Assessment Call →

Schedule your assessment call

Tell us a little about your setup so we can make the call count — then pick a time that suits you.

About you
Your AI agents