Our Agents
Artificial intelligence is significantly improving our ability to understand complex situations, such as interconnected social, ecological and economic systems. The LIA fully harnesses this benefit, but recognises that AI can reach only so far before human understanding is needed.
By combining AI and human agents, the LIA delivers local intelligence of the highest quality and completeness, and portrays a place’s dynamics responsibly.
Each agent works at the point in a process where it adds the most: some at the outset, some where judgement and field work are needed, and some as circumstances change.
Agent Atlas
Agent Atlas (“Atlas”) is the flagship agent of the LIA, deployed at the outset, before any due diligence begins or costs are committed. Atlas systematically builds a complete picture of the natural systems, industries and communities shaping a region, and how they connect. It then makes this intelligence accessible through an interactive interface, so that users can test different actions, trace exposures and effects, and take their first steps already equipped with a specialist’s understanding of a place’s interconnected risks, costs and opportunities.
Where Atlas works in the process
The same stages apply whether the process is an investment, a permit, a policy or a programme.
- A
Screening
Atlas builds the full picture of the place before anything is committed.
Scoping and engagement
Human agents set the questions and open the conversations.
Appraisal
Human due diligence, with Atlas re-run on the questions that arise.
Decision, structuring and permitting
Human agents, with AI agents on the stakeholder and regulatory checks.
Delivery
Human agents in the field.
- A
Monitoring
Atlas tracks the dynamics as they change.
- Atlas
- Atlas returns
- Human agents
Atlas in numbers
- ~290engines and algorithms, every one documented and traceable
- 400–1,500sources read per region
- 20–30 hto generate a region’s full intelligence
Choose your inputs
Choose a POV to generate a preview.
Choose your inputs and generate a preview.
