Intake route
When to use Idea Forge
Use Idea Forge at the start of an engagement when you know the customer or industry, but not the first useful AI proof. It is an intake tool: it helps the team turn public research and customer signals into ranked, right-sized ideas before choosing a scenario playbook.
Five-minute idea funnel
Prepare the prompt
Bring the customer or business-unit name, industry, and any known region, segment, or pain points you are allowed to use.
Run the skill
Customer Activity-Forge researches public sources only, marks gaps instead of guessing, and proposes approximately ten achievable ideas.
Select one direction
Prefer an idea with a narrow user, an owner, a safe representative sample, and a first evidence case the customer can review.
Map the implementation work
Break the idea into parts and map each part to scenario lessons across tracks. Record adaptation needs and uncovered work, then use the mapping to plan sessions.
What to prepare
Required
- Customer name or business unit.
- Industry or sector.
Helpful if known
- Region or customer segment.
- Allowed pain points or public signals to verify.
- Any data, content, or process boundary already discussed.
Keep out
- Confidential account-plan or CRM details.
- Unapproved customer data.
- Claims the team cannot cite or validate.
What Idea Forge produces
Research summary
Three to five sentences with public-source citations and explicit coverage gaps.
Ranked ideas
Approximately ten ideas ranked by customer fit, achievable first proof, and differentiation.
Session mapping
Each idea identifies the parts of the use case, relevant lessons across tracks, and uncovered work.
Scenario handoff
The top idea gets a pre-filled handoff with assumptions clearly marked for customer confirmation.
Sample handoff into a scenario
This example shows the shape of the handoff. Replace the placeholders with the skill output and keep anything unconfirmed visible for the customer conversation.
- Customer outcome
- Reduce time spent finding trusted answers for a specific user group.
- Target users and boundary
- Named role and access boundary to confirm with the customer.
- Context and source owner
- Approved corpus, data product, or synthetic representative sample plus the accountable owner.
- Use-case parts and session coverage
- Relevant lessons for each part, with adaptation needs and uncovered work made explicit.
- Evidence starter
- Routine case, edge case, refusal/review case, or access case that proves the first outcome.
- First customer decision
- The next decision the scenario playbook should help the customer make.
Agree the first implementation scope
Move to the relevant lessons once the idea has an owner, a safe sample, and a way to test it. If no track fits, identify the missing pattern. An idea is a starting point for scoping the work, not a delivery commitment.