Pilot
Test one business use with a defined set of documents and questions.
- One priority knowledge area
- Hosted deployment
- Search testing and adjustments
- Findings for your decision-makers
Use the findings to decide whether to roll out.
Discuss a pilotiKB pricing and deployment optionsStart small.
Scale when it works.
Begin with a pilot on one team's documents. Roll out when the results are clear. Choose a pilot, a business rollout or private deployment. iKB pricing is based on your sources, users, infrastructure and support needs. Request a quote after a requirements review.
Three ways to start
We provide a quote after reviewing your requirements. No fixed subscription prices are published.
Business
Use company knowledge to support employees and customer-service teams.
- Multiple topics and channels
- Studio and connector requirements
- Decision Records and model policies
- Support and rollout planning
Agree the rollout, support and responsibilities.
Plan a rolloutPrivate
Choose where the platform runs and where data is stored and processed.
- Private, on-premise, or air-gapped
- Local model integration
- Security and audit requirements
- Your IT and support requirements
Plan around your infrastructure and security needs.
Discuss private deployment
Pricing factors
What affects the cost
What affects the cost
of an AI knowledge base?
Review setup costs and ongoing costs before choosing a deployment.
- Knowledge sources
- Number of topics, files and recordings. Data sources, approved connections, update frequency and search requirements also matter.
- Infrastructure
- Hosted or private deployment, data location, offline use and local models. Include model usage, computing resources and storage. Confirm what the quote includes and which services cost extra.
- Users & channels
- Native chat, website chat, messaging connections and APIs. Include administrator roles and how many users need access at once.
- Governance & support
- Decision Records, evidence exports, retention and privacy checks for tables. Include security monitoring integration (SIEM), security reviews and rollout support.
AI models & usage costs
Choose models.
Choose models.
Review ongoing costs.
Use supported hosted providers or local models through an OpenAI-compatible interface. Set defaults, apply topic rules and review usage. Confirm where each model processes your data.
Review the architecture
Before you decideQuestions from procurement,
Questions from procurement,
IT, and risk teams.
How much does iKB cost?
iKB provides a quote after reviewing your requirements; it does not publish a fixed subscription price. Costs depend on sources, users, workload, connections, infrastructure and support. You can try the live demo before discussing a pilot. Request a quote.
Can we run iKB on our own infrastructure?
Yes. iKB supports private cloud, on-premise and air-gapped options, as well as hosted deployment. Agree the setup against your security and IT requirements. Compare deployment options.
Can we use our preferred models?
iKB supports hosted model providers and local OpenAI-compatible endpoints, with rules for each topic and usage controls. Confirm model compatibility and costs during the requirements review. See model controls.
How is answer quality evaluated?
Test your documents with real questions, including questions designed to expose errors or unsafe answers. Check citations and search results. Adjust the setup and review failures before production rollout.
What does the pilot produce?
A recommendation based on the test findings. It covers what to deploy, access and governance settings, channels, and who will maintain the system. Agree what the pilot will include in your proposal.
Are all features available in every deployment?
No. Features depend on topic settings, platform configuration, compatible models and providers. Connections need credentials, reviewed actions and topic approval. Interactive answer cards are opt-in for native chat; the embedded widget keeps plain answers. Notifications need a supported browser and permission.
What changes in an air-gapped environment?
Air-gapped means isolated from external networks. Core file processing, search and answering can run locally with local models. Features that need live external sources or hosted services are unavailable. See the air-gapped planning checklist.
Get a quote that fits your setup.
Tell us which sources, users and deployment options you need. We will help define a suitable starting point.