How we engage and price.
Every workflow has a different shape, so we scope each engagement to your actual systems, data, and goals. Here's how a Quantilus engagement is structured, what drives the investment, and how to get a firm number for your project.
How to read this page
A Quantilus engagement runs in three phases: a Strategy & Discovery Sprint to figure out what to build, a Build phase to ship the agent, and ongoing Operations to keep it improving. Our engineers embed with your team throughout. Pricing depends on scope, which is why the Strategy Sprint exists: it produces a fixed, procurement-ready quote for everything that follows. Bring us a workflow and we'll give you a firm number quickly.
Engagement Tiers
Three shapes. One lifecycle.
Start at Strategy. Move into Build. Stay for Operations. Most clients follow that path in that order.
Tier 01 · Strategy & Discovery Sprint
Figure out what to build, before building it.
Duration: 4–8 weeks · Engagement: Fixed price, fixed scope
We map your workflow end-to-end, inventory the systems an agent would need to touch, score opportunities by value/risk/readiness, and hand you a build plan you can act on, with us or anyone else.
- Workflow audit across customer, employee, and back-office touchpoints
- System inventory: what the agent could read from and act on, and what's missing
- Ranked agent backlog by value, risk, and readiness
- Compliance footprint: HIPAA / FERPA / GDPR / SOC 2 as applicable
- Reference architecture tailored to your stack
- 12-month deployment roadmap from first agent to fleet
What you walk out with
A complete agent backlog, a working architecture diagram, a fixed scope for your highest-value first agent, and a go/no-go framework for the next ones.
Best for: Leaders with budget approved for AI in the next 12 months, who want to spend it on the right workflows.
Quoted bespoke if: You need workflow audits across 3+ business units, a multi-language regional rollout plan, or formal vendor-selection support.
Tier 02 · Agent Build & Launch
Ship an agent that actually does the work.
Duration: 3–6 months to production · Engagement: Time & materials, not-to-exceed ceiling
The core engagement. We build the agent end-to-end: knowledge layer, tool actions, memory, guardrails, evaluation harness, admin console. Each capability lands one at a time. Weekly demos against a real quality bar. No big-bang launch.
- Model selection (frontier, open-weight, or hybrid) sized to cost and data sensitivity
- Knowledge ingestion: docs, policies, contracts, customer history, pricing rules
- Tool actions built one at a time: CRM lookup, draft quote, send spec, page colleague, open ticket
- Memory at the right scope (conversation, customer, org-wide)
- Guardrails: approval gates, data-isolation rules, audit logging
- Continuous evaluation harness so changes never regress quality
- Admin console your team can run on its own
What you walk out with
A production-grade agent reaching Stage 2–4 of the capability ladder, deployed inside your environment, with quality scores, full audit logs, and the runbook your team needs to operate it.
Best for: Teams with a committed workflow, the data behind it, and an executive sponsor.
Quoted bespoke if: Multi-workflow / fleet builds, custom model fine-tuning, on-prem deployment in restricted environments, multi-tenant SaaS-style productization.
Tier 03 · Operations & Evolution
Keep the agent earning its keep, every quarter.
Duration: Ongoing · Engagement: Monthly retainer
Launch is the start. Most agents quietly stall a year in because nobody owns operations. We don't. We run on-call for your agent, watch real performance, retire what isn't earning its keep, and move it up the capability ladder on a steady cadence.
- 24/7 monitoring of uptime, latency, cost, output quality, drift
- Quality regression tests on every change
- Monthly ops report: usage, cost, accuracy, ticket-deflection metrics
- Policy refinement based on actual conversations, not guesses
- Quarterly capability reviews + roadmap update
- Model upgrades when better ones ship (no rebuild)
- Bug triage, incident response, customer-side support
What you walk out with
An agent that's still earning its keep at the 12-month mark, and has climbed from Stage 2 to Stage 4 of the capability ladder. Plus a quarterly plan for what's next.
Best for: Anyone who wants their agent to still be valuable in 2027.
Quoted bespoke if: Fleet operations (5+ agents), 24/7 follow-the-sun coverage, strict-regulatory ops with named compliance officer, multi-region residency.
Add-on · Enterprise Integration
Per-system connectors, priced per connector.
Duration: 2–8 weeks per integration · Engagement: Fixed-fee per connector
Pre-built connectors are quick. Custom adapters to legacy or homegrown systems take more. We quote each connector before we build it, so you always know what an integration costs before committing.
- Fast: Salesforce, HubSpot, Slack, Zendesk, ServiceNow, Google Workspace
- Moderate: NetSuite, Workday, SAP, Canvas, Blackboard, Klopotek, Arc XP, WordPress VIP
- Involved: Epic / Cerner via HL7-FHIR, custom internal APIs, legacy mainframe gateways, hardened on-prem ERPs
What you walk out with
A working connector with permissions, rate limits, audit logging, and graceful fallback when the downstream system is unavailable.
Listed as add-on because integrations are almost always part of a Build or Operations engagement, not standalone.
Price Drivers
Six things that move the investment.
Scope is everything. Here's what we look at to size an engagement, and what pushes a project into bespoke territory.
01 · Number of integrations
An agent touching 2 systems is half the cost of one touching 6. Each integration adds discovery time, auth flow design, error handling, and ongoing operational surface area.
02 · Data residency & regulation
A workflow that requires HIPAA-BAA, FedRAMP-aligned deployment, or air-gapped operation adds the most. Compliance evidence, audit packages, and security review add real engineering time.
03 · Custom model work
Most agents work fine on off-the-shelf frontier models. If your domain needs fine-tuning, distillation to a smaller model for cost, or a custom evaluation harness, that's additional scope.
04 · Multilingual scope
English-only agents are the baseline. Each additional production language adds prompt work, eval data, redaction tuning, and reviewer staffing. Bilingual support is often included; 6-language regional rollouts are bespoke.
05 · SLA & uptime
99.5% business-hours SLA is the baseline. 99.9% follow-the-sun with named on-call, hot-standby model gateways, and multi-region failover adds to the operations scope.
06 · Fleet / multi-tenant
Building one agent for one team is straightforward. Building an agent platform that 12 business units can each configure for their own workflows is a different engagement, quoted bespoke.
AI as a Service · Custom engagements
Complex or tailored? A bespoke quote, not a band.
Quantilus also offers AI as a Service (AIaaS), fully-managed, tailored AI solutions where we run the model layer, agent runtime, and continuous improvement on your behalf. Pricing for AIaaS and other complex engagements is bespoke, scoped to your actual workflow, data volume, and SLA needs.
Quote-only engagements typically include:
- AI as a Service (AIaaS). Fully-managed, tailored AI solutions, we run the platform, you consume the capability
- Multi-workflow fleet builds. A platform where multiple business units each get a configurable agent
- Custom model training or distillation. Fine-tuning a foundation model on your domain corpus
- Air-gapped / classified deployments. No-internet operation in defense, intelligence, or critical-infrastructure environments
- Highly regulated multi-jurisdiction rollouts. Healthcare across multiple countries, financial services with cross-border restrictions
- White-labeled / OEM agent platforms. You ship our agent to your customers under your brand
How we scope it: bespoke engagements always start with a paid discovery. We won't put a number on a first call, because a real one depends on the workflow, the data, and the SLA. Discovery gives you a firm, fixed quote for everything that follows.
Pass-through costs
Model inference is billed separately.
Anthropic, OpenAI, Google, and open-weight hosting costs are pass-through at provider rates, scaling with your call volume. We size capacity, build the caching layer, and route easier requests to cheaper models so you never burn budget on token costs.
We forecast inference and infrastructure costs in your Strategy Sprint so there are no surprises later.
What we forecast for you
Model inference: sized to your call volume and model mix
Infrastructure: hosting, monitoring, data storage on your cloud
Third-party SaaS: any CRM seats, messaging, or paid APIs the agent calls
Self-hosted open-weight models flatten to GPU rental cost: lower at high volume, higher at low.
Pricing FAQ
What buyers usually want to know next.
How does Quantilus compare to Big-4 consultancies on cost?
Meaningfully less for equivalent agent scope, with faster delivery (months vs. quarters) and engineers doing the work instead of partners selling and analysts delivering. Where the Big-4 win: change management at scale across 50,000-person organizations. Where we win: the actual agent that does the actual work.
How does Quantilus compare to a custom build with internal engineers?
An in-house team can absolutely build a production agent. The trade-off is time-to-first-value (typically 6–12 months for a team learning the stack vs. our 3–6 months) and the operational overhead of staying current with rapidly evolving AI tooling. Many of our clients run an in-house team and use us for the first few agents to accelerate learning.
Are there costs beyond the engagement fee?
Three line items: (1) model inference (pass-through at provider rates), (2) infrastructure (your AWS/Azure/GCP bill for hosting, monitoring, data storage), (3) any third-party SaaS the agent calls (CRM seats, messaging, paid API endpoints). We forecast all three during the Strategy Sprint, so you see the full picture before committing.
Can we run a paid POC before committing to a full Build?
Yes. A focused proof-of-concept on a single workflow lets us build a working agent on a real integration, ship a quality report, and let you decide whether to extend into a full Build. Most clients do. We'll scope and quote the POC for you up front.
Do you offer revenue-share or outcome-based pricing?
Occasionally, for workflows where outcomes are cleanly measurable and attributable (e.g., recovered revenue from churn flagging, deflected tier-1 ticket cost). We'll propose it where it fits. Most engagements are flat-fee or T&M because outcomes get muddied by everything else the business is doing.
Will you publish a real SOW we can take to procurement?
Yes, at the end of a Strategy Sprint, you walk out with a procurement-ready SOW for the highest-value first agent. Deliverables, timeline, milestones, dependencies, acceptance criteria, and a not-to-exceed ceiling. No surprises at signing.