AI built for your business.
Software you own.

No. CogniSphere Agent Platform is MIT-licensed open-source software. You own your custom agents, applications and data, and can keep running and modifying them after our engagement ends. Monthly plans pay for engineering, consulting and ongoing support, not a software subscription. Your cloud, model usage and other third-party service costs remain separate.

For the Build, Evolve and Sustain plans, we do not track or bill actual hours. We agree a fixed estimate in hours before work begins, and that is the amount deducted from your plan balance or billed as additional hours for the approved scope. We finalise requirements during your included consulting calls, at no extra charge. For complex requirements, we may propose a separately approved exploration phase to produce a reliable implementation estimate. Your monthly report shows approved work, estimated hours charged and your remaining balance.

We explain the impact on scope, timeline and cost, then ask you to approve a revised estimate before starting the changed work. The revised total accounts for hours already charged, so you do not pay twice. If the original scope simply takes us longer, we absorb the overrun. You can keep the original scope or defer the change.

Prices are in USD and exclude applicable taxes. You pay your cloud, hosting, model/API usage, storage and other third-party services directly. Nova and Eva come free with the Agent Platform, with no separate agent fee; their infrastructure and model usage still incur provider charges. Additional human engineering or customisation uses an approved fixed estimate.

Yes. Every plan includes 1 hour strategy/consulting calls with our AI expert at the stated frequency and one monthly upgrade of the underlying Agent Platform, including installation and basic verification. Neither reduces your plan hours. Changes to custom agents, applications or integrations, including compatibility work, need a separate approved estimate. We agree a safe rollout or defer an upgrade if there is a blocker.

Our delivery plans start on Build for the first three service months while active. From month four, you can continue Build or choose Evolve or Sustain. Three complete service months must pass between effective plan changes. Either party can cancel at any time with one calendar month’s written notice, including during onboarding. Fees and new-hour allocations for a final partial month are prorated.

Unused plan hours carry over up to 120 hours on Build, 60 on Evolve and 30 on Sustain, separate from the new month’s allocation. This is a balance cap, not a three-month expiry. We use the oldest credits first. Excess carryover expires at month-end, while existing banked hours are preserved on a downgrade. Banked retainer hours have no cash value and expire when the engagement ends.

We agree the fixed estimate first, apply available unallocated plan and banked credits, then invoice the shortfall at your plan’s additional-hour rate. You approve and prepay those additional hours before the work starts. Extra hours do not authorise us to exceed the approved estimate for an unchanged scope.

Ready-to-start features begin within 1 week on Build, 2 weeks on Evolve or 4 weeks on Sustain, once scope, estimate, priority, access, credits and payment are in place. These are start commitments; delivery dates depend on the agreed work. Production support runs daily, 8 AM–10 PM IST. Incidents reported in that window are acknowledged within one elapsed hour, even if the deadline falls after 10 PM. Outside that window, acknowledgement is due within one hour of the next 8 AM opening. Resolution timing depends on the incident; investigation and remediation need an approved scope and estimate. We are expanding our team to provide 24×7 on-call support and will notify you when it becomes available.

You own your business-specific deliverables and data, and retain administrative control of your cloud and model accounts. We can deploy on your preferred cloud, including AWS, GCP or Azure, with region and access agreed upfront. The underlying CogniSphere Agent Platform remains open source. You can continue running and maintaining your custom deliverables after the engagement ends.

We agree permissions, spending limits, actions requiring human approval and escalation rules before production deployment. Nova handles minor app changes and quick debugging within those controls. Eva prepares requirements and tracks progress, while your team approves priorities and business decisions. You decide who can monitor and disable the agents.

You review deliverables against the acceptance criteria agreed before work begins. Material defects attributable to CogniSphere and reported within 30 calendar days of delivery are corrected without an additional charge or deduction of plan hours. New requirements follow the scope-change process.

Our engineers allocate 50% of their bandwidth to client billable work and 50% to research, training and Agent Platform improvements. This internal investment keeps our engineering knowledge and systems current and is not deducted from your plan. Client-specific exploration is different: it has its own deliverables and fixed estimate, which you approve before it begins.

Yes. We can staff an AI engineer who works directly with your team.

Part-time: $60 USD/hour for 20 hours per week. Full-time: $50 USD/hour for 40 hours per week.

Your engineer gets access to CogniSphere’s resources and training materials, plus guidance from our AI experts.

Staffing gives you engineering capacity. It does not include a fixed commitment to deliver specific features or outcomes.