How We Price Production AI Agents: Foundation + Per Agent + Connector

Quick verdict

Most “AI agent development cost” pages sell a blob: $8k–$400k tiers with no roadmap when you add agent #3. Cipher quotes a production agent platform as three line items:

Platform foundation (one-time) — identity, BFF, frontend rails, AgentCore/runtime plumbing, shared memory, observability, secure AWS posture. Paid once; every agent reuses it.
Per production agent — marginal cost to harden one proven or clearly scoped agent into production.
Per connector / integration — each external system (Microsoft Graph, HubSpot, and similar) as its own line so integration risk stays visible.

Best for: Australian and Singapore teams with Claude or ChatGPT workflows that already work — or clearly scoped agents — who want a predictable build path on AWS. Honest limit: we do not sell SOC 2 / ISO 27001 certification, and we do not fixed-price unbounded prompt-churn. Indicative floors live on /pricing/; every engagement is scoped.

Last updated: August 2026. Cipher Projects is an Australian AI engineering studio that builds production agents and the platforms they run on.

Why blob quotes fail buyers and delivery teams

A single fixed number feels simple until the second connector and the third agent show up. Then someone eats the overrun: either the client pays surprise change-orders, or the studio under-delivers and delivery quality collapses.

Transparent line items are a trust mechanism, not a marketing slogan. You can check the commercial architecture before you sign. We can staff and schedule what we actually promised. Delivery stays sustainable because scope is named.

The three-line model (indicative AUD)

Line	What it is	Indicative floor	Why it exists
Platform foundation	Shared rails every agent reuses	From AUD $32,000 (self-use)	Paid once so scale does not rebuild identity and runtime
Per production agent	Harden one agent (design, tools, evals, UAT, runbook)	From AUD $12,000 (typical ~$12–15k)	Scales predictably with agent count
Per connector	One external system wrapper + auth + tests	From AUD $3,500	Stops burying integration risk inside a blob

USD and SGD conversions appear on the pricing page. Product companies that stamp instances for downstream clients usually need a higher foundation (roughly from AUD $40,000 for stampable single-tenant, from about AUD $50,000 for multi-tenant). Per-agent stays in the same band either way.

What foundation actually includes

Foundation is the difference between a Claude chat that “works for the team” and a system operators can audit:

AWS landing posture (VPC, KMS, IAM, CloudTrail) that does not block SOC 2 / ISO 27001 controls
Identity with MFA — Cognito or federated Entra SSO when the company already lives in Microsoft 365
NestJS BFF bridge to agent runtime (SSE streaming) with app-layer RBAC
Next.js authenticated UI shell
Bedrock AgentCore / Strands plumbing: runtime, memory, identity, gateway, observability, model-routing skeleton
First connector wrapper via the gateway (priced in foundation or as the first connector line, depending on intake)

Stack defaults are AWS + Strands + Bedrock AgentCore. We mention them for architecture honesty, not as a shopping list. Runtime token and AgentCore meters are cloud spend — see AgentCore pricing and Bedrock pricing — separate from the build quote.

What moves price up or down

Direction	Drivers
Up	New connectors the agent depends on; human-in-the-loop approval chains; external customers on the login path; heavier compliance review; agents not yet proven in Claude/ChatGPT
Down	Agents already proven with real usage; shared tools/memory across a family; read-only agents (no writes to external systems)

Unique insight from quoting work: the fastest way to blow a “cheap agent” quote is burying two SaaS connectors and an undefined third agent inside one fixed blob. Separating connectors is how both sides stay honest when HubSpot OAuth or Graph consent takes longer than the prompt work.

Claude workflow vs production agent (pricing lens)

	Claude / ChatGPT workflow	Production agent on your platform
Identity	Vendor account / project share	Your SSO/MFA, per-user tool scoping
Isolation	Chat session	Per-session runtime isolation + audit trail
Connectors	Manual paste or thin MCP	Gateway wrappers, OAuth, least privilege
Cost control	Seat + usage surprise	Model routing, timeouts, eval sampling budgets
Human approval	Person in the chat	Explicit escalation paths in the workflow
Commercial path	Subscription	Foundation once → agents N → connectors later

Full cutover narrative: Claude workflow vs production agent.

Worked shape (anonymised)

Two agents already proven as Claude workflows, Microsoft 365 as first connector, one additional SaaS connector, internal staff only, SOC 2 / ISO 27001 aspirations (infra that does not block — not the cert itself):

Foundation in the mid self-use / product band
Two per-agent lines in the ~$12–15k area
One extra connector line
Phase total commonly mid five-figures to ~AUD $75k depending on tenancy and connector weight · roughly 8–12 calendar weeks with two senior AI engineers

That is a shape, not a promise. Intake changes the point estimates; the line-item structure does not.

What is not included

Certification: we prepare controls and evidence paths; assessors certify your organisation.
Unbounded discovery: “we will invent the agents as we go” belongs in advisory or a managed seat, not a fixed platform quote.
Cloud runtime bills: Bedrock tokens, AgentCore meters, vector/storage — operational spend.
Managed monthly seats: ongoing capacity is a different product on /pricing/.

FAQ

Is this cheaper than a managed AI engineer? Different job. Seats are recurring capacity. The agent-platform package is a fixed-scope build of shared rails plus N agents. Many clients do foundation + first agents, then keep a seat for iteration.

Can we start with one agent and add more later? Yes — that is the point of foundation once. Design the gateway so later connectors slot in without rewriting identity and runtime.

Do you only build on AWS AgentCore? That is our default for AU/SG teams that need AWS-native isolation and audit language. TypeScript-portable stacks (Mastra and similar) are a different gravity — see our Mastra vs Strands cutover.

Who is this for? Founders and operators who already feel the Claude workflow working and need production identity, connectors, and a quote they can defend to a board — not a vendor blob.

Conclusion

Price the platform the way you will operate it: foundation once, agents as you prove them, connectors as named systems. That is checkable integrity for the buyer and a deliverable roadmap for Cipher. Start from the bands on /pricing/, then scope the phase.