How TripleDart Runs Keyword Research With Claude and Ahrefs

How TripleDart Runs Keyword Research With Claude and Ahrefs

Team Pick
AI & Agents
TripleDart
Workflow guide
The split we run in production: Claude takes the mechanical 80% of clustering and intent, an analyst keeps the three judgment calls.
Organisation
TripleDart
Written By
TripleDart SEO team
PUBLISHED
August 5, 2026
Read
3
min

What it is.

Our own workflow, written up with the division of labor spelled out. Ahrefs owns the volume data; Claude owns clustering, intent reclassification and calendar sequencing.

The number that decides whether this is worth doing is intent accuracy. A junior analyst calls a query informational 80% of the time; Claude against SERP context calls it commercial correctly about 90% of the time in our testing.

That 10-point swing moves roughly 15% of a typical content calendar out of the wrong bucket, which is where the return comes from.

What you get.

  • The split between what Ahrefs owns and what Claude owns
  • The export shape Claude reads most reliably
  • Three prompts covering semantic clustering, intent reclassification and calendar prioritization
  • Intent accuracy of about 90% against a junior analyst's 80%, and where it drops
  • The three calls we keep human, including which cluster to cut despite high volume
  • What the finished calendar looks like coming out of the pipeline

Read it at the source.

Opens on TripleDart. Free, and the figures on this page all come from it.
Read the guide

Frequently Asked Questions

How accurate is the model half?
What stays with the analyst?
Can AI do keyword research?
Does intent classification change results?
How accurate is AI at classifying search intent?
What should stay human in keyword research?
Strengths
  • States the accuracy figure both ways, so the claim can be argued with.
  • Publishes the prompts, so the workflow can be rebuilt without a call.
  • Names what stays human, which most AI workflow posts skip.
Limitations
  • Our own workflow, and TripleDart sells the service it supports.
  • Accuracy figures come from internal testing, with no published sample behind them.
  • Assumes an Ahrefs seat and someone comfortable in Claude Code.
Skip this if
  • Skip it if you already have a keyword process you trust end to end.
Ideal for
  • SEO leads
  • Content strategists
  • GTM engineers

The plays behind the reading.

We build inbound GTM engines for B2B software teams, and these are the frameworks we build from. Bring us a pipeline target and we'll map the route.