When you give AI the right context, it can support better decision-making in construction. Without it, you will still get an answer, but not necessarily the right one.
Context is the information that shapes a decision, such as constraints, goals, and conditions.
To explore this, let's test a simple scenario.
Without context
Prompt: We need to replace a rail bridge that is at the end of its life. Recommend one single type of bridge to build. Do not provide multiple options. Give a short answer in under 30 words and briefly explain why.
Each model produces a confident recommendation, but they point in different directions. Without context, the outcome is shaped by the model’s training, not the needs of the project.
With context
Now let’s provide more project-specific information, or “context”, and ask the same question again. If context matters, the recommendations should begin to reflect the actual constraints of the project. Not just general engineering knowledge.
Let’s give the models the following four pieces of context:
Human value: The project should create local employment and apprenticeship opportunities during construction.
Natural value: All equipment and materials must be brought in by rail, with no road access available.
Economic value: The solution should minimise whole life cost, not just initial construction cost.
Social value: The bridge should improve pedestrian and cycling connectivity, making it easier for people, such as students, to access nearby schools and services.
Let’s ask the same question again, but this time with this context included.
Prompt: We need to replace a rail bridge that is at the end of its life. The project should create local employment and apprenticeship opportunities. All equipment and materials must be brought in by rail, with no road access. The solution should minimise whole life cost. The bridge should improve pedestrian and cycling connectivity, including access to nearby schools and services. Recommend exactly one type of bridge. Your answer must contain only one option. Maximum 30 words. Briefly explain why.
What Changed
With context, the answers become more consistent.
Earlier, the models suggested different solutions, including prestressed concrete and various truss and steel options. With context, they all move towards a modular steel truss approach. The reasoning also changes. Instead of general benefits, the answers now reflect the project constraints.
Takeaway
AI can support decision-making, but only when the problem is clear.
Define what matters, then use AI.
Notes
- All models were tested using fresh accounts with no prior history.
- Prompts were identical across all models.
- No edits were made to any model outputs.
All content reflects our personal views and is not intended as professional advice or to represent any organisation.
/
1