Decide what the run must solve

Start with the decision the reader needs to make: which technical directions deserve further engineering work? Then define the contradiction that should produce that shortlist. Include the system, the result you want, what must not get worse, and any hard boundary such as size, energy, cost, or process compatibility. For example, "Reduce drying time and defects in a lithium-ion battery electrode coating without increasing oven footprint or energy cost" gives the engine both the desired outcome and the boundary that rules out the obvious fix.

Claude can help you refine the problem statement and identify relevant TRIZ parameters, but it cannot retrieve the cross-industry patent cases or inventive-principle rankings that turn a contradiction into concrete candidate solutions. Patsnap Solution Engine fills that gap by running a multi-step TRIZ workflow: it analyzes your problem, retrieves relevant patent cases, applies inventive principles, and returns candidate concepts ranked by feasibility and benefit.

Write an input Claude can execute

Prepare a concise English problem description under 500 characters that includes your performance goal, the constraint you must respect, and enough context for the engine to understand your product or process. For the battery-electrode example, a complete input reads: "Battery electrode drying suffers from long cycle time and uneven moisture distribution, leading to defects. The oven footprint and energy budget cannot increase. Improve drying speed and uniformity without adding heaters or lengthening the conveyor."

Use Patsnap Solution Engine to explore this TRIZ problem.

System: lithium-ion battery electrode drying line
Current problem: long cycle time and uneven moisture distribution cause defects
Improve: drying speed and moisture uniformity
Do not worsen: oven footprint or energy consumption
Hard constraints: no additional heaters and no longer conveyor

Generate candidate solution directions grounded in cross-domain patent cases. For each direction, show the TRIZ principle, analogous case, transfer logic, expected benefit, main feasibility risk, and the first validation test.

You do not need to choose TRIZ matrix parameters yourself—the engine will analyze the problem and select the relevant contradictions automatically. Focus on clarity: name the system, the unwanted effect, the desired improvement, and the boundaries.

1Connect Patsnap Solution Engine to Claude

Patsnap Solution Engine is available as an MCP server that connects to Claude Desktop or another compatible client. To connect:

1. Obtain a Patsnap Open Platform API key (the Starter plan is free and includes 10,000 credits for initial testing).
2. Navigate to the Solution Engine marketplace page and generate a connection link.
3. Add the server to your client configuration.

Once connected, Claude can start a new solution task, monitor progress, retrieve candidate concepts, and request detailed recommendations without leaving the conversation.

2Generate candidate solution directions

Ask Claude to generate candidate solutions for your prepared problem statement. Behind the scenes, the engine starts a new task, analyzes the contradiction, retrieves relevant patent cases, applies TRIZ inventive principles, and streams progress updates until the run completes. A complete run typically takes about 5 to 10 minutes.

In the battery-electrode drying example, the input completed in approximately seven minutes and returned nine candidate concepts. Each concept combines a TRIZ principle with a cross-domain patent analogy: staged temperature-gradient control adapted from food drying, pulsed microwave heating from polymer curing, segmented airflow from automotive paint booths, and others. The engine ranks concepts by expected benefit and provides a summary of the inventive principle and supporting patent case for each.

Example directions reported in the draft run; rankings and transfer feasibility must be verified for the actual process.

4Compare concepts with engineering constraints

After reviewing the detailed recommendations, compare the top three or four concepts against your constraints. Ask Claude follow-up questions that combine the returned TRIZ principles with your system knowledge: "How would staged temperature-gradient control affect line speed and product quality for a 60-micrometer NMC coating?" "What process changes would pulsed microwave heating require?" "Can segmented airflow be retrofitted into the existing oven?"

Claude can help you structure the trade-off analysis, while the Solution Engine can retrieve additional patent cases or solution details for any candidate you want to explore further.

Select concepts for validation

Narrow the candidate list to two or three concepts that offer the best balance of expected benefit, engineering feasibility, and alignment with your constraints. For each selected concept, document the TRIZ principle, the supporting patent case, the expected performance improvement, and the validation experiments you will run. In the drying example, you might select staged temperature-gradient control for immediate lab trials because it requires only oven-controller software changes, while flagging segmented airflow as a medium-term project.

Schedule lab trials that measure drying time, moisture uniformity, and defect rate under the new process conditions, and compare the results against your baseline. If the concept does not transfer as predicted, return to the candidate list and test the next-ranked principle, using the TRIZ reasoning and patent cases to refine your hypotheses.