Pricing

Start free. Scale by footprint.

AIGIS separates evaluation setup, proof, team runtime, and enterprise runtime. Configure the workspace at no cost, validate a narrow workflow with a fixed-price sprint, then choose a production band that matches company size, systems, usage, and runtime boundary.

Available now

Free

Evaluation Setup

$0

workspace configuration before production runtime

Create a workspace, choose a setup path, configure connector placeholders, prepare API keys, and document rollout scope before payment.

  • Workspace and organization setup
  • Setup profile and deployment-mode recommendation
  • Connection Center for runtime, system, database, chat, file, and AI provider placeholders
  • API key creation and rotation for SDK preparation
  • Console audit for setup events
  • No production runtime execution, live secret-backed connector testing, SLA, or customer-hosted deployment until activation
Lowest paid step

Fixed scope

Proof Sprint

$10K-$25K

2-4 week validation before production activation

Prove the first governed workflow without buying the full deployment. Best for one system, sandbox or read-only production access, capped usage, and a clear go/no-go plan.

  • One primary system and one target workflow
  • Salesforce-first where possible
  • Sandbox or read-only production validation preferred
  • Capped guided runtime validation; no production SLA
  • Customer-owned model keys for model-backed tests
  • Security and architecture readout
  • Credit toward Production Pilot when the SOW is signed within 60 days
Scoped after setup

Activation

Production Pilot

$50K-$150K

typical activation SOW for first governed workflow

Move from configuration into a live governed workflow with approved secrets, runtime boundary, connector testing, model routing, and handoff. Salesforce-first pilots usually land in this range.

  • Salesforce-first production reference path
  • AIGIS-hosted, hybrid, or customer-hosted boundary selected from evaluation setup
  • Approved secret handling and connector test flow
  • Bring your own production model keys
  • Governed read/write workflow activation
  • Runbook handoff and support plan scoped before go-live
  • ServiceNow, SAP, database, or custom connector hardening priced by scope
Runtime ladder

Team to enterprise

Runtime Ladder

From $36K/yr

team, business, enterprise, and private bands

Annual runtime pricing scales by company size, systems, governed work-unit volume, isolation, support, and compliance boundary.

  • Team: 50-250 employees, $36K-$120K/yr
  • Business: 250-1,000 employees, $120K-$300K/yr
  • Enterprise: 1,000+ employees, $300K-$750K+/yr
  • Private or regulated: $1M+/yr scoped
  • All bands exclude customer systems, cloud, and model-provider costs
  • Upgrade path as footprint, isolation, or support needs grow
  • Per-system pricing band scoped to active-user capacity, not per-conversation or per-seat tolls

Bring your own systems

Salesforce, ServiceNow, SAP, Dynamics, chat tools, identity providers, APIs, databases, and warehouses stay customer-owned. AIGIS governs access across them.

Bring your own LLMs

Use customer-owned OpenAI, Anthropic, Gemini, Bedrock, Azure, or private model keys. Model-provider spend stays outside AIGIS fees.

Bring your own databases, or Governed-hosted

Customer databases and warehouses remain customer systems. The AIGIS runtime database, cache, and audit plane can run in your cloud or in a Governed-hosted managed boundary.

Free setup does not include production runtime execution, live customer secrets, SLAs, or customer-hosted deployment. Customers bring their existing enterprise systems, cloud/runtime accounts, identity provider, and production model keys for every activated path. AIGIS governs and orchestrates that stack; it does not bundle Salesforce, ServiceNow, SAP, databases, warehouses, chat tools, or LLM provider licenses. Scope depends on company size, systems, permission complexity, integration depth, usage, support requirements, and compliance boundary.

Budget case
Planning range
Timing
What it usually covers
Proof sprint
$10K-$25K
2-4 weeks
One workflow, one system, capped usage, sandbox or read-only preferred. Creditable toward production.
Salesforce-first pilot
$50K-$150K
6-10 weeks
One governed read/write workflow, one primary system, BYO model keys, hosted or hybrid runtime.
Multi-system pilot
$150K-$350K
8-16 weeks
Salesforce plus ServiceNow, database, SAP hardening, or custom internal API work.
Team runtime (50-250 employees)
$36K-$120K/yr
Annual
1-2 systems, capped usage, standard support, and shared hosted, hybrid, or light customer-hosted boundary. A 250-person org usually starts here.
Business runtime (250-1,000 employees)
$120K-$300K/yr
Annual
2-4 systems, standard production support, stronger audit reporting, and hosted or hybrid runtime sizing.
Enterprise runtime (1,000+ employees)
$300K-$750K+/yr
Annual
Multi-system production platform, heavier usage, support boundary, operational runbooks, audit posture, and stronger isolation.
Private/customer-hosted enterprise
$1M+/yr
Scoped
Dedicated control plane, customer cloud or VPC boundary, deployment kit, data residency, and stricter SLAs.

These are budgetary planning ranges, not a quote. A 250-person company with one or two governed systems usually starts in Team Runtime, not the $300K+ enterprise band. We lock the actual number after evaluation setup confirms systems, permission complexity, runtime boundary, support expectations, and compliance requirements.

LLM spend

Model bills stay separate. Routing changes the curve.

AIGIS pricing excludes production LLM usage. Customers bring their own model-provider keys and pay OpenAI, Anthropic, Google, Bedrock, Azure, or a private model provider directly. These examples are normalized per 100K governed work units, not per 100K prompts or LLM calls.

A governed work unit is one permission-checked execution step: cache lookup, API query, database query, workflow action, write proposal, or model-backed synthesis. One user prompt can fan out to many work units, and AIGIS can batch, cache, or route those units so they do not each become an LLM call.

Direct-to-frontier baseline
Per 100K governed work units: every unit sends 20K input tokens and receives 1K output tokens from a frontier model.
~$13K / 100K work units
Illustrative GPT-5.5 list-rate math at $5/M input and $30/M output. This is the prompt-stuffing pattern AIGIS is designed to avoid.
AIGIS routed + cached
Per 100K governed work units: 35% API/cache with no model call; 35% small-model synthesis; 30% frontier calls with permission-stripped context.
~$2K / 100K work units
About 80-85% lower model spend in this example because repeated answers, direct system lookups, and simple synthesis do not all hit the frontier model.
AIGIS high-cache workflow
Per 100K governed work units: 60% API/cache with no model call; 30% small model; 10% frontier fallback for ambiguous or high-value work.
~$600 / 100K work units
This is not a whole-enterprise ceiling. Repeatable workflows get cheaper per request, while the absolute bill still scales with request volume.
Simple lookup
1-2 work units per prompt
50K-100K prompts
0-1 LLM calls per prompt
Cross-system analysis
3-8 work units per prompt
12K-35K prompts
1-2 LLM calls per prompt
Multi-step operational prompt
8-20 work units per prompt
5K-12K prompts
1-3 LLM calls per prompt

Execution routing

Not every request needs an LLM. AIGIS can answer through direct APIs, SQL, workflow execution, or cache before paying a model.

Permission-stripped context

When a model is needed, AIGIS sends the smallest governed context it can, after object, field, and record checks.

Result cache

Provider prompt caching discounts repeated input. AIGIS result caching can avoid the entire model call when a governed answer is still valid.

Model routing

Simple synthesis can route to cheaper models. Frontier models stay reserved for ambiguous, sensitive, or high-value work.

Formula: monthly model bill = input tokens times input rate plus output tokens times output rate, plus any provider-specific search, grounding, long-context, priority, or cache-write charges. The table normalizes model spend per 100K governed work units. A large enterprise with 1M similar high-cache work units would model roughly 10x the row, not the same $600 total. Every customer should rerun this with their own provider contract, work-unit volume, and usage logs.

Deployment model

Your cloud. Your keys. Your data stays put.

AIGIS is governance infrastructure for the stack you already own. Evaluation setup lets you choose AIGIS-hosted, hybrid, or customer-hosted boundaries before production activation, and every production deployment uses customer-owned system access and model keys.

Production default
Customer-hosted

You bring the cloud and the keys.

Enterprise production deployments can run inside your environment: AWS, Azure, GCP, private cloud, or on-prem. During evaluation setup, we map whether the first activation should be AIGIS-hosted, hybrid, or customer-hosted based on systems, secrets, compliance, support, and deployment readiness.

  • Customer already owns the systems of record AIGIS governs
  • Private production architecture for enterprise deployments
  • Customer pays cloud and model provider costs directly
  • AIGIS deployment kit and runbook completed during enterprise hardening
  • Compliance posture aligned with the approved customer runtime boundary
  • Customer-managed keys for both data and LLM access
Optional
AIGIS-hosted (managed)

We host. Sized to the workload.

After production activation, Governed can operate an AIGIS-hosted managed instance for you. Hosted is sized by company footprint, workload, and isolation requirements: shared hosted cells for smaller teams, dedicated single-tenant runtime/database/cache for larger workloads, and private cloud when required. Your production model usage still runs through customer-provided provider keys, so model spend stays on the customer side.

  • AIGIS-hosted means Governed hosts AIGIS, not your Salesforce, SAP, ServiceNow, database, or model account
  • Customer-owned model keys for production
  • Shared, isolated, and dedicated hosted options
  • Dedicated database/cache/audit boundary available
  • Higher operational SLA (we own uptime)
  • Quarterly access reviews and tenant audit reports
  • Runtime provisioned after activation approval
  • Railway-style hosting only for early/smaller cells

Why this matters: The whole point of AIGIS is that your data stays in your systems. Customers come with the systems, data stores, identity provider, cloud accounts, and model providers they want governed. Customer-hosted is the production architecture that makes that promise strongest, but evaluation setup chooses the safest first operating boundary instead of pretending every enterprise can self-serve a private runtime on day one.

Add-ons

Where the work that moves the needle gets priced.

Most enterprise deployments include some custom work. We don't bury it in the platform fee. We price it transparently so you can budget.

Add-on
Range
Notes
Dedicated AIGIS-hosted runtime
$15K-$30K/mo
Managed production capacity beyond the capped shared Team Runtime cell. Customers still bring and pay for production model keys.
Dedicated hosted data plane
$35K-$100K+/mo
Single-tenant MCP runtime, database/cache/audit boundary, regional placement, and higher capacity.
Custom connector development
$50K-$150K/system
Per system. Includes spec, build, hardening, and 90 days of stability support.
Professional services
$300-$500/hr
Topic engineering, change management, advanced governance design, executive workshops.
Premium SLA tier (15-min response)
$10K-$30K/mo
Tighter SLAs, dedicated escalation path, named on-call engineers.
24x7 white-glove support
$25K+/mo
Mission-critical workloads with multi-region failover playbook.
Training and certification
$10K-$50K/cohort
Per cohort, on-site or remote. Includes admin certification, advanced topic engineering, and runbook handoff.
Executive briefing program
$15K-$50K/qtr
Quarterly briefings for the executive team on AI governance, regulatory shifts, and roadmap influence.

Annual prepay carries a 10% discount. Multi-year commitments (2 or 3 years) are negotiable for additional discount in exchange for the term.

Reference economics

Vendor AI pricing needs explicit assumptions.

The old headline stack math was too easy to misread as a universal quote. Keep the economic argument, but make the buyer-visible version source-aware: public list prices where available, and scoped assumptions where products are quote-based or consumption-based.

Reference
Current public posture
How AIGIS should use it
Salesforce Agentforce
Public flat-fee access is listed at $125/user/month; other buying models include credits and conversations.
Use as a directional list-price benchmark, not a universal replacement claim.
Microsoft 365 Copilot Enterprise
Public enterprise list price is $30/user/month, with Copilot Chat included for eligible users.
Useful productivity AI; AIGIS is for governed operational-system execution.
SAP, ServiceNow, MuleSoft
SAP uses AI Units and quotes; ServiceNow and MuleSoft are quote/package/consumption based.
Model these only with explicit buyer assumptions.

Sources should be cited in sales collateral: Salesforce publishes Agentforce buying models; Microsoft publishes Copilot Enterprise pricing; SAP Business AI uses AI Units and quotes; ServiceNow and MuleSoft are buyer-specific quote or package motions.

Evaluation boundary

Free does not mean production runtime.

The evaluation workspace is the buyer's configuration and planning surface. Live connector secrets, governed execution, production model routing, customer-hosted deployment, and SLAs remain activation-scoped.

Assumptions

Free setup includes

Workspace, setup profile, connector placeholders, API keys, console audit, and rollout scoping

Proof sprint includes

A narrow validation path: one system, one workflow, capped usage, no production SLA, and no customer-hosted deployment

Production activation starts when

Runtime execution, live secrets, live connector tests, SLAs, customer-hosted deployment, or governed writes need to go live

Company size guidance

Headcount is a planning proxy, not the billable unit. Pricing follows systems, governed usage, runtime isolation, support tier, and compliance boundary

Budgetary ranges

Ranges are planning guidance, not a quote. Final pricing depends on system count, runtime boundary, connector depth, support level, and compliance requirements

Margin guardrail

Production model usage and customer-hosted infrastructure are customer-paid. Governed-hosted runtime is priced separately because MCP, database, cache, audit, observability, and support costs scale with usage and isolation

Customer stack

AIGIS does not include Salesforce, ServiceNow, SAP, databases, warehouses, cloud accounts, identity providers, chat tools, or system-of-record licenses. Customers bring the stack AIGIS governs

Database boundary

Customer databases and warehouses remain customer systems. Governed-hosted applies only to the AIGIS runtime database, cache, and audit plane unless a scoped SOW says otherwise

Model spend

AIGIS fees exclude LLM provider usage. Customers bring provider keys, set budgets directly with the model provider, and can use AIGIS routing/cache controls to reduce token exposure

Public list references

Salesforce Agentforce flat-fee access and Microsoft 365 Copilot Enterprise publish list prices

Quote-based references

ServiceNow Otto / Now Assist, SAP Business AI / Joule, and MuleSoft are quote-based, consumption-based, or package-based

Excluded from AIGIS scope

Third-party system licenses, production model usage, customer cloud/runtime costs, internal admins, legal procurement work, and custom work outside the scoped SOW

Production keys

Customers bring and pay for production model provider accounts on every deployment path

Principles

How AIGIS pricing actually works. No per-conversation tolls.

Principle
What it means
Customer brings the model
You hold the Anthropic / OpenAI / Bedrock API key. We never mark up tokens.
Per-system pricing band
Pricing scales with the number of governed systems and the active-user capacity band — not with per-conversation counts or per-seat licensing.
Quote-based above $50K ACV
Pilots and Runtime tiers above the Team band are quote-based — we'll send a written proposal after a 30-minute scoping call.
Pricing — AIGIS Enterprise AI Governance | Governed