Skip to main content
The Positioning tab answers a question citations and mentions cannot: when an AI answer calls a product “the easiest to integrate” or “the best for regulated industries”, is it talking about you or a competitor? You write the claims you want to own. After every probe run, DevTune reads each relevant answer and records which product it singles out for each claim: your brand, one of your positioning competitors, another product, several products, or nobody.
Positioning is available on Pro, Agency and Enterprise plans. If you don’t see the Positioning tab, contact us to turn it on for your account.

​
How answers are scored

An answer is scored against a claim when:
  • the probe completed successfully
  • the answer mentions your brand or one of your tracked competitors
  • the prompt’s intent is one of the intents the claim applies to
Each scored answer credits the claim to exactly one of: DevTune records an outcome only when it is confident in it. Answers it can’t read confidently are left out rather than guessed. Share is the answers crediting a product divided by the answers crediting any single product, including Other product. Several and Nobody are shown beside the shares but aren’t part of them. A share needs at least 10 credited answers. Below that, you see the count without a percentage.

​
Dashboard tile

The Positioning tile on the AI Search dashboard shows how many of your claims you lead. You lead a claim when answers credit you more often than any single competitor. The tile also names the claim you are closest to leading, with your share and answer count. The tile appears when the dashboard shows the last 30 or 90 days, with no topic filter and no more than one platform selected, because positioning is measured over those fixed windows.

​
Share over time

The chart shows your share, and your competitors’, for one claim at a time. Pick the claim from the menu at the top right.
  • Each point on a line is the share over the seven days ending that day. Single days rarely have enough answers to mean anything on their own.
  • Dots show single days for the selected product, sized by how many answers they rest on. A small dot is a day with only a few answers.
  • Earlier competitors appear as muted lines when your competitor set changed during the window. A marker shows the day it changed.
The first six days of a window have no line, because their seven days reach back before the window starts. Days with no scored answers have no point.

​
Claim comparison

Beside the chart, Claim comparison lists every product in your competitor set for the selected claim, with its share and answer count over the window. Other product is listed last.
  • Select a row to highlight that product’s line in the chart, and select it again to clear it. Selecting a product in the chart’s legend does the same.
  • Select an answer count to open the answers behind it.

​
Change since the previous period

Each share in Claim comparison, and your brand’s column in the claims matrix, shows how much it moved against the previous period of the same length: the 30 or 90 days before the window. Changes are in percentage points. Your own change is green when it rises and red when it falls. Competitors’ changes are grey, because a rival gaining isn’t good news for you. A change is shown only when the two periods compare:
  • Your competitor set didn’t change during either period. After a change, comparisons return once a full previous period has been measured against the new set.
  • The previous period was fully measured. Every answer that should have been scored was. A claim added recently, or answers still being scored, means no change yet.
When a change isn’t shown, Claim comparison says why. Each share still needs 10 credited answers on both sides.

​
Claims matrix

Below the chart, the matrix shows every active claim against every product in your competitor set, plus Other product, Several and Nobody. Your brand’s column comes first. Select any count in the matrix to open the answers behind it.

​
The answers behind a number

Opening a count lists the answers it includes, newest first, with the AI platform and date of each. Select an answer to open that exact response on its prompt’s page, even when the platform ran the prompt more than once that day. The list uses the same window and filters as the page, so the number of answers matches the count you selected.

​
Filters

  • Window: 30 or 90 days
  • AI platform: one platform or all
  • Prompt intent: one intent or all
Every number on the tab, including the chart, table, matrix and answer lists, follows the same filters.

​
Competitor set

Shares compare your product with up to five competitors. Choose them in Project settings → Brands, under Positioning competitors. DevTune suggests the five competitors AI cited most over the last 30 days. If you don’t choose, it uses the first five competitors you track. Changing the set starts a new comparison. Shares from the earlier set stay in history but aren’t compared with new ones, and DevTune warns you how many days of comparable history a change cuts off before you save it. After a change, the page tells you how many days of the window the current shares cover.

​
Managing claims

Select Manage claims to add, edit, approve, dismiss, archive or restore claims.
  • Add a claim to start measuring it. DevTune also scores the answers it still holds from recent runs, up to about 30 days, so a new claim usually has some history on day one.
  • Suggest claims drafts up to ten proposed claims from your saved company profile. Proposed claims aren’t scored until you approve them. Dismissed claims are kept, so they aren’t suggested again, and you can restore them.
  • Edit a claim that already has history, and DevTune starts a new claim and archives the old one with its history. A trend never mixes two different questions.
  • Archive a claim to stop scoring it. Restore it to start again. The archived period stays a gap and isn’t filled in.
Pro plans can have 10 active claims per project, and Agency and Enterprise plans 20. Proposed, dismissed and archived claims don’t count toward the limit. See Writing positioning claims for how to write claims that measure well.

​
API and MCP

Your own agents and scripts can read the same data:
  • GET /projects/{projectId}/attributes lists your claims
  • GET /projects/{projectId}/attributes/match returns counts and shares over a fixed window, and the previous period’s totals when the two periods compare
  • GET /projects/{projectId}/attributes/{attributeId}/answers lists the answers behind one count
The MCP tools are devtune_list_attributes, devtune_get_attribute_match and devtune_get_attribute_match_answers. All of them require the attributes.read scope. See the API overview and MCP server.

​
Next Steps