Goal-to-Prompt Router
Match each goal to the prompt type most likely to produce an actionable result
- Difficulty
- Moderate
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 98%
The Goal-to-Prompt Router is a knowledge-grounded assistant that converts a user's desired outcome into an appropriate prompting strategy. Its reference material is one or more trusted prompt engineering guides. The user states a clear goal and supplies practical context such as role, audience, resources, limits, and success metrics. The router analyzes the underlying task, cross-references it with the prompt types in its guides, and returns three elements: the recommended prompt type, an explanation of why it fits, and a complete prompt ready to run. This separates prompt selection from prompt execution. Instead of memorizing a taxonomy, users focus on defining outcomes while the assistant performs the pattern matching and carries constraints into the generated prompt.
Origin
Extracted from Marketing Against the Grain, where a custom GPT used Google's prompt engineering guide to route marketing goals to suitable prompt types.
Core principles
- 01Start with a specific measurable outcome
- 02Use a trusted guide as the routing taxonomy
- 03Match the task mechanism rather than surface keywords
- 04Explain why the selected prompt fits
- 05Return a complete ready-to-use prompt
- 06Carry user constraints into the final prompt
How to run it
- 1
Ground the router
Provide the assistant with one or more authoritative prompt engineering guides. Define those documents as the source for available prompt types and selection logic.
Pro tip Adding guides from multiple model providers can expand the available taxonomy.
Watch out Do not let conflicting guides blur distinctions without reviewing how their terms map.
- 2
Define the return contract
Instruct the assistant to return the selected prompt type, the reason it suits the goal, and a complete prompt the user can execute.
Pro tip Use a fixed output format so recommendations remain easy to compare.
Watch out A prompt name alone does not help the user understand or trust the selection.
- 3
State a clear outcome
Describe the result to achieve, including relevant role, subject, target audience, time horizon, and measurable success condition.
Pro tip Phrase the input as an outcome rather than merely naming a topic.
Watch out Vague goals force the router to infer critical requirements.
- 4
Add operating constraints
Supply limits such as team size, available channels, budget, saturated options, or execution capacity. Require the router to preserve these constraints in its generated prompt.
Pro tip Mention what has already been tried so the recommendation does not repeat exhausted options.
Watch out Missing constraints can yield a theoretically sound but impractical prompt.
- 5
Match the task mechanism
Have the assistant analyze whether the goal requires transformation, decision-making, contextual prompting, examples, or another source-backed technique. Ask it to justify the match in terms of the task's actual demands.
Pro tip Check whether the explanation discusses impact, feasibility, and the kind of reasoning needed.
Watch out Keyword matching can select a prompt type that sounds relevant but has the wrong mechanism.
- 6
Run and evaluate the prompt
Execute the generated prompt with the required source material and assess whether its output advances the original success metric. Refine the goal or constraints if the result is too generic.
Pro tip Test the router on several different goal types before relying on it routinely.
Watch out A persuasive routing explanation does not guarantee that the generated prompt will perform well on every model.
In the wild
A sole product marketer asks the router to turn a product webpage into a 30-day launch campaign designed to win five customers. The router selects a transformation prompt and incorporates the source page, staffing constraint, timeline, and target outcome into a ready-to-run instruction.
→ The executed prompt produces a four-week launch playbook with guardrails, checkpoints, content ideas, list-building activities, social proof, and referral tactics.
A paid marketer explains that Google Ads and Meta are saturated, identifies the product as a B2B AI SDR tool, and targets 20% more signups from paid experiments. The router selects a decision-making prompt that weighs alternative channels by impact and feasibility.
→ The marketer receives a structured prompt for prioritizing the first new channel and designing an experiment.
Common mistakes
Supplying a topic instead of a goal
The router needs a desired result to distinguish between transformation, decision-making, contextualization, and other prompt mechanisms.
Omitting practical constraints
Ignoring staffing, budget, timeline, or existing channel saturation can produce recommendations the user cannot execute.
Treating routing as execution
Selecting a prompt type is only an intermediate step; the generated prompt still needs to be run and evaluated against the original outcome.
Is it for you?
Best for
People who repeatedly use AI for varied goals and need reliable help selecting an appropriate prompting technique.
Not ideal for
Vague exploratory requests that lack a clear outcome, relevant context, or meaningful constraints.
From the transcript
“This prompt engineer is going to say, well, this is the goal you have, and because of that, here's the type of prompt you should…”
“You have to get a clear goal.”
“it actually will pattern match the goal that you have against the right prompt type.”
From the episode
Become a Google-Level Prompt Engineer in 20 Minutes