Magnetforce is a private AI layer for Salesforce Sales Cloud, Service Cloud, and Experience Cloud. It adds practical AI use cases to the Salesforce environment your team already uses: account research, pre-meeting briefings, call transcription, conversation summaries, sales reports, and CRM record enrichment.

It is built by Ideaport Riga, a Salesforce implementation partner.

No. Salesforce stays the system of record and the daily work surface. Magnetforce connects to Salesforce through standard configuration artefacts and APIs, so AI-assisted outputs appear inside Sales Cloud, Service Cloud, and Experience Cloud.

About 50% of agentic AI projects stay stuck in pilot, with security and compliance cited as the primary barriers (Architecture and Governance, 2026). Magnetforce is designed to clear those barriers by running in customer-controlled infrastructure and using customer-approved AI models, so AI projects can move from pilot into production inside Salesforce.

CIOs, IT Directors, Salesforce platform owners, IT architects, and security and compliance leaders at midsize and enterprise Salesforce customers who need AI inside Salesforce workflows but want control over data residency, model choice, cost, and operations.

AI orchestration sits in the open source Magnet AI layer, not in a vendor controlled black box. Prompts, RAG flows, tools, and agents are configuration artifacts you own. You can swap language, speech, and embedding model providers without rebuilding use cases, and there is no platform markup on AI model usage.

Magnet AI runs in customer controlled public cloud, private cloud, or private network environments. Sensitive sales, service, customer, case, and knowledge data does not have to be processed in a vendor controlled AI environment. Model providers are also your choice, including locally deployed models for maximum data residency control.

Sensitive data stays inside customer controlled infrastructure. Deployment components, integration patterns, and monitoring hooks are documented for architecture review. Prompts, RAG flows, and tools are versioned configuration artifacts that can be governed and audited. Configuration export and import APIs make environment promotion repeatable across dev, test, and production.

Magnet AI provides export and import APIs for prompt templates, RAG flows, retrieval tools, API tools, MCP tools, and agents. Teams move AI assets across development, test, staging, and production with a repeatable operating model and a clear audit trail, instead of ad hoc reconfiguration.

Magnet AI itself is free and open source. Customers largely own the solution: the orchestration layer, the configuration artifacts, the operating logs, and the model choice. Premium support is available separately from Ideaport Riga and is highly recommended for production use.

Magnetforce uses standard Salesforce configuration artefacts that connect Sales, Service, and Experience Cloud to the Magnet AI platform via API. A user triggers an action in Salesforce, Salesforce sends record context to Magnet AI, Magnet AI orchestrates the right prompt template, RAG flow, retrieval tool, or agent, an approved AI model runs, and structured output returns to the Salesforce record. Usage, cost, and quality are logged.

Three layers:

Salesforce layer - Sales Cloud, Service Cloud, Experience Cloud. Where users see, trigger, and consume AI-assisted workflows.

Magnet AI layer - A low-code AI orchestration platform with prompt templates, RAG tools, retrieval tools, API and MCP tools, agents, knowledge sources, channels, evaluation, usage analysis, and OpenTelemetry tracing.

AI Models layer - Customer-approved language, speech, and embedding models, deployed locally or in the customer's chosen cloud.

Magnet AI deploys in customer-controlled public cloud, private cloud, or private network environments. For Azure, the documented resources include virtual network, Azure Database for PostgreSQL, Azure Container Apps, Microsoft Foundry / Azure OpenAI as an optional model provider, and Azure Container Registry as an optional container source.

Customer-approved language models, speech models, and embedding models. Magnet AI connects to many LLM providers by storing API keys, capabilities, and costs, and can use either locally deployed LLMs or public cloud models under the provider's security and privacy policies. There is no platform markup on AI model costs.

Magnet AI supports document ingestion, parsing, chunking, vector embedding creation, semantic search, and metadata filtering across sources such as Salesforce Knowledge, SharePoint, HubSpot CMS, Oracle Fusion Knowledge, public file stores, and websites. Custom connectors are supported via OpenAPI v3 or MCP. No Data Cloud is required.

Every LLM call, RAG question and answer, and agent conversation is logged and analyzed by response quality, cost, and latency. Dashboards aggregate performance over time. Technical traces are available in OpenTelemetry format and can be sent to external monitoring tools such as Azure Monitor.

Two main commercial components.

Onboarding: 6,000 EUR one time. All Salesforce configuration artefacts are free; a 2-week onboarding engagement is mandatory.

Infra and AI models: provided by the customer (container runtime, PostgreSQL, and access to AI models). No platform tax on AI model costs.

Magnet AI itself is free and open source. Premium support is available separately and is highly recommended for production use.

The 6,000 EUR is a one-time fee covering the mandatory 2-week onboarding engagement. The Salesforce configuration artefacts themselves are free. The engagement covers Sales Cloud, Service Cloud, and Experience Cloud setup and configuration of the customer's first use cases.

Customers provide a container runtime, PostgreSQL, and access to AI models. Infrastructure runs in the customer's chosen public cloud, private cloud, or private network environment. AI model usage is paid directly to the customer's chosen providers, with no extra platform tax.

Magnet AI is used in production at organizations including Konecranes, GMC, Tet, and LIAA, supporting use cases such as technician Q&A, customer summarization, spare parts search, public information research, customer and employee chatbots, and structured evaluation of applications.

Yes. A pre-recorded demo is available so you can see Magnetforce in action before any sales conversation. Use the See demo button anywhere on this page. No scheduling and no sales call required.