Trajectory is making its post-training software available to businesses looking to adapt open-source AI models to their own workflows. Designed for specialist teams in fields such as law, medicine and finance, the platform connects to company applications, collects feedback on AI-generated outputs and uses it to improve models according to each organisation’s standards.
The system uses continual learning, allowing models to update through ongoing human input rather than relying on separate, one-off retraining projects. For example, legal teams can refine how a model reviews contracts, while finance teams can train it to follow a particular valuation approach. Companies including Harvey and Rogo already use the technology.
Trajectory’s public launch aims to make this type of model customisation more accessible to businesses without dedicated AI research teams. The company has also raised a $45 million Series A at a $300 million valuation.
Image Credit: Trajectory
Why This Trend Is Growing
- Continual Model Customization
- Ongoing human feedback loops are reshaping enterprise AI by allowing open-source models to evolve with company standards, policies and domain-specific workflows.
- Post-training Platforms
- Business-ready post-training tools create opportunities for organizations without AI research teams to refine models for specialized tasks in law, medicine, finance and other expert fields.
- Feedback-led AI Governance
- Embedded review systems can turn employee corrections and approvals into structured model improvements, linking performance enhancement with compliance, quality control and institutional knowledge.
Industries Being Reshaped
- Legal Technology
- Contract review, legal research and document analysis platforms can benefit from AI systems that learn firm-specific language, risk thresholds and preferred interpretation methods.
- Financial Services
- Valuation, reporting and advisory workflows gain new potential from models trained continuously on analyst feedback, proprietary methodologies and organization-specific decision standards.
- Healthcare Technology
- Clinical documentation, medical coding and specialist support tools are positioned for more adaptive AI as practitioner feedback becomes part of the model improvement process.
