Scale | Donovan
Scale | Donovan lets public‑sector organisations build custom AI agents, evaluate model outputs, analyse mission data, monitor live feeds for threats, and generate knowledge bases from raw inputs.
About this tool
Scale | Donovan is designed for public‑sector users who need to assemble AI agents that act on specific data streams. By feeding mission‑related data into the platform, agencies can generate insights and alerts that support decision‑making in dynamic environments.
The system also offers tools to compare model outputs, create structured knowledge bases, and monitor live data sources such as GDELT. These capabilities are accessible via API, cloud, or on‑premise installations, allowing integration with existing workflows.
What you can do
Agent Factory
Transforms raw data sources into customized AI agents that can be programmed for particular tasks, enabling agencies to automate routine analyses.
Model Evaluation
Runs side‑by‑side comparisons of model responses, helping users identify the most suitable model for a given mission requirement.
Threat Monitoring
Continuously watches real‑time data feeds and issues alerts when patterns indicative of emerging threats are detected.
Who is it for?
- A city emergency management office builds an agent that ingests weather and traffic data to forecast flood impacts and issue evacuation notices.
- A national statistics bureau creates a knowledge base from census records, allowing analysts to query demographic trends without manual data cleaning.
- A public health agency monitors global news streams via GDELT to receive early warnings of disease outbreaks and coordinate response measures.
Before you choose
How does the platform handle data sovereignty when deployed on local infrastructure versus cloud services?
What criteria are used in the model evaluation process, and can users adjust weighting of accuracy versus speed?
Frequently asked questions
Can Scale | Donovan be integrated with existing on‑premise data warehouses?
Yes, the platform supports local deployment and can connect to internal data stores through its API, allowing agencies to keep sensitive information within their own network while still using the agent creation and analysis tools.
What types of real‑time data can be monitored for threat alerts?
The system can ingest any streaming source that provides structured data, including public feeds like GDELT, sensor networks, or custom event streams, and will generate alerts based on patterns defined in the monitoring configuration.
Is it possible to customize the criteria used in model comparison?
Users can define evaluation metrics such as precision, recall, latency, or cost, and the Model Evaluation feature will rank models according to the selected criteria, enabling tailored selection for each mission.
Check the official website for current features, pricing, and terms.
Similar tools
Chatbase
AutomationChatbase is a conversational AI platform that turns business data into a custom knowledge base, enabling automated Q&A, task handling, and escalation to human agents while delivering analytics on interaction patterns.
Fulcra
AutomationFulcra enables users, agents, and applications to exchange and manage data across multiple sources, providing transparent monitoring and controlled access to shared knowledge.
Amazon Bedrock
CodingAmazon Bedrock is AWS’s managed service for building generative AI apps and agents, with access to hundreds of foundation models and the AgentCore agent platform.
Dify
AutomationDify is an open-source platform for building LLM apps and agentic workflows, with a visual workflow studio, RAG pipelines, model management and observability.
Context
AutomationThe system receives a user query and produces a response through its chat capability.
Echo Win
AutomationEcho Win is an AI‑driven call handling platform that automatically answers inbound calls, provides live transcription, and builds custom call‑flow scenarios while delivering analytics to improve service quality.