Describe your applications and display
List your main games or applications and the kind of projects you use. Include the monitor resolution, refresh rate and number of displays. These details help explain where graphics performance, CPU performance or memory capacity matters most.
For professional software, check the vendor's current hardware requirements and supported accelerators. Compatibility with the application can matter more than a component's position in a general benchmark.
Set a budget for the complete setup
Explain whether your budget covers only the computer or also a monitor, operating system, peripherals, installation and delivery. This keeps the proposed configuration comparable to the amount you intend to spend.
- Main applications or games and typical project sizes
- Monitor resolution, refresh rate and connections
- Budget range and what it needs to include
- Existing parts, with exact model numbers
- Desk space, noise preferences and appearance
- Likely upgrades during the system's life
Treat memory and storage as workload choices
Capacity needs depend on active applications, project sizes and how much work runs at once. Include current usage and expected growth rather than choosing memory from a generic label such as gaming or professional.
Separate fast working storage from archive and backup needs. A large internal drive or NAS is not automatically a complete backup; decide how important data will be copied and recovered.
Plan cooling and compatibility together
Case dimensions, airflow, graphics-card clearance, power connectors and cooling requirements need to fit the selected components. Noise under sustained load also matters, particularly when the computer sits near the user.
Hardline water cooling can serve a particular thermal and visual brief, but it adds cost and maintenance considerations. Air cooling or other liquid-cooling choices may be a better fit. Discuss maintenance expectations before choosing a showcase loop.
Give AI workloads a separate brief
For local AI, identify the models, tools and tasks you expect to run. Model size, precision, context and concurrent workloads affect memory needs. An advertised GPU model alone does not establish that a particular model will fit or run at the speed you expect.
Confirm software compatibility and licences for the selected workflow. Some local tools avoid per-request API fees, but hardware, electricity and maintenance still have costs.
Agree the configuration and handover
Review the exact part list, availability, permitted substitutions, software setup and the applicable support and warranty terms. Confirm collection or delivery arrangements for your location.
Send this brief with the models of any parts you want to reuse. Their compatibility and condition need checking before they become part of the agreed configuration.