Start a free auditFree audit
Product

From the answers assistants give to the report your client approves.

Six parts, one per client: measure the answers, find the sources cited around them, rank the work, check what followed it, report it in your brand, and pull any of it into your own tools.

Measure: every question, on every assistant you switch on

Write the questions your client’s buyers ask, import a list, or generate a draft from templates and edit it. Each question is asked several times on every assistant you switch on, and each cell is an aggregate of those answers. Every two weeks by default; weekly or daily if you switch it on.

Prompt × assistant · Fernpost · last 28 days · example dataestimated · confidence: medium
Prompt · client: FernpostChatGPTPerplexityClaudeGrokswitched on
best project tool for a small design studio44%33%22%33%not measured
Fernpost vs Quillstack89%100%78%89%not measured
project software with a client portal22%11%33%11%not measured
cheapest way to share timelines with clients0%competitor named0%competitor named11%–not measured
how to move a studio off spreadsheets11%0%competitor named0%22%not measured
project tool with time tracking and invoicing33%22%44%22%not measured
share of answers naming Fernposta competitor named, Fernpost notfewer than 3 answers, so no numbernot measured
Share of sampled answers mentioning the brand. Every cell is an aggregate over the window, never one answer. “Not measured” means we do not ask that assistant, so there is nothing to report either way.
Scroll the table sideways to see every assistant →
3 + 2Three assistants on Starter; all five from Growth, switched on per client.
≥ 3answers per question per assistant before a cell shows a number.
2 weeksbetween measurements by default. Weekly or daily if you switch it on.
Keptevery raw answer, with its model version and cost.

Microsoft Copilot and Google AI Overviews / AI Mode offer no public API, so they are not measured, and no report estimates them.

Diagnose: which sources are cited where competitors are named

Every answer comes with the pages it cited. Grouped by kind of source, they show where competitors are named and your client is not, and whether the gap sits on the client’s own site or out in the category.

Cited sources · Fernpost · example data
share of answers citing each source · last 28 days
  • [1]reviewhub.example· gap18%
  • [2]forum.example· gap12%
  • [3]toolreview.example· names both9%
  • [4]listings.example· gap7%
  • [5]fernpost.example· names client6%
“Gap” = cited in answers that name a competitor but not the client.
Diagnosis · Fernpost

The gap is driven by third-party sources rather than by a shortage of owned content.

[1]reviewhub.example· gap[2]forum.example· gap[3]toolreview.example· names both
Based on the sources models actually cited in the measured period. Presence is inferred from answers where the source was cited, not from checking the page itself.
Evidence · answers citing reviewhub.exampleexamples from the sample, shown after the aggregate
Perplexity · answer 1 of 323 Jun
Most reviewers recommend Quillstack, with Loambox as a cheaper option …
[1]reviewhub.example· gap
ChatGPT · answer 2 of 324 Jun
Reviewers rate Quillstack highest for larger teams; Tidepin suits freelancers …
[2]reviewhub.example· gap[3]tidepin.example
Grok · answer 3 of 325 Jun
On review sites, Loambox and Quillstack collect the most studio reviews …
[1]reviewhub.example· gap

Opportunities: ranked work, each with a reason

Each opportunity is derived from measured answers: where the client is losing, on which questions, and what the cited sources suggest doing about it. Ranked by an internal score, not by certainty of outcome. No opportunity and no action exists without a reason.

Opportunities · Fernpost · example data5 of 11
  1. 1

    Get Fernpost covered on reviewhub.example

    Reason: reviewhub.example is cited in 18% of answers for this category (14 citations). Quillstack and Loambox appear in those answers; Fernpost does not.

    estimated impact: higheffort: mediumtype: review platform
    Size of the gaplarge
    Share of tracked questions38%
    Commercial intenthigh
    How much of it you controlpartly
    How well it is measuredmedium
    Share of the questions this client tracks, not search demand — the product does not measure how often people ask them.
  2. 2

    Get Fernpost covered on forum.example

    Reason: forum.example is cited in 12% of answers here (9 citations), and threads comparing the category name competitors without mentioning Fernpost.

    estimated impact: mediumeffort: medium
  3. 3

    Publish a page that answers the comparison questions directly

    Reason: no page from fernpost.example appears among the 9 sources cited for the comparison cluster, so there is nothing of the client's own to cite.

    estimated impact: mediumeffort: medium
  4. 4

    Refresh the fernpost.example pricing page

    Reason: the page is cited 6 times here, but the brand is not mentioned in those answers. It is being read without carrying the name.

    estimated impact: mediumeffort: low
  5. 5

    Get Fernpost covered on listings.example

    Reason: listings.example is cited in 7% of answers here (5 citations), with two competitors listed and the client absent.

    estimated impact: mediumeffort: low
Actions board · this sprint · example

Backlog 2

Review listingReason: reviewhub.example cited in 18% of answers
AKowner
Comparison pageReason: no own page among 9 cited sources
MRowner

In progress 1

Pricing page refreshReason: cited 6 times, name not carried
JSowner

Done 3

Migration guideReason: 3 answers send migrations to a competitor
AKdone
+ 2 more
Untouched topics · invoicing, time tracking. This sprint’s work does not touch them, so they are what later movement gets compared with.
  • Every action opens as a brief: the objective, why it matters, the numbers behind it, steps and acceptance criteria for its type.
  • Dismissing needs a reason too, so the next person does not re-open it. A dismissed opportunity comes back only if it grows noticeably larger than it was when you dismissed it.
  • “Not this month”: Kept out of the list until the date passes, then shown again with whatever the measurements say by then.
  • Nothing is published for you. The board records work your team does; any change to a client’s site stays a human decision.

Experiments: what followed the work, as an estimate

When a piece of work is marked done, the product compares the topics it touched with the client's other topics that it left alone, over the same weeks, and shows the difference as an estimate with a confidence level. It is a record of what followed, not a claim about why.

Migration guide refresh · Fernpost · next sprint · example data
Topics the guide coversUntouched topics: invoicing, time tracking
010%20%30%40%W22W23W24W25W26W27W28W29W30W31W32W33Guide marked done · W27Covered topics 25.5Untouched 22.2
010%20%30%40%W22W25W28W31W33Marked done · W27
Before: the 28 days before the work was marked done (W23–W26) · after: every week since (W27–W33) · 3 assistants × 3 samples per question
Estimated incremental effect
+2–5 pp

An action's estimated contribution is shown as a band around the estimate. The band has a fixed width: it marks the figure as rough, it is not a statistical interval.

estimated · confidence: medium
TopicsBeforeAfterChange
Covered by the guide19.0%23.7%+4.7
Untouched19.9%21.5%+1.6
Difference+3.1

This is a record of what was done and what followed, compared with topics the work did not touch. It is evidence, not attribution of cause.

Without untouched topics there is nothing to compare against, and the report says so instead of showing a difference.

White-label reports, in your agency’s name

A report is a page your client opens from a link, without an account, in your logo and colour. You can download the same page as a PDF to forward.

What the client can do

  • Open the report from a link, with no account and no login
  • Approve the report and the next sprint in it; the name and date are recorded
  • Read the caveats in the “How to read this” section at the end

What the report page never shows

  • Our name or logo
  • “Powered by” or any link back to us
  • Pricing, plan names or upsell messages
  • Worth knowing. The link opens without an account. It can be served from a domain you own — ask us and we set it up — and on that domain nothing but the report exists. Sent from the product, the email goes out under your agency's name from our sending address.
  • What always stays. Figures are called estimates, the measurement basis is stated, and the caveats close the report. In an audit report every ranked item carries its reason; a quarterly report’s next sprint is the plan you agree with the client.
Fernpost
2026-04-01 — 2026-06-30
AI answer visibility

Across the tracked buyer questions, Fernpost was named in an estimated 28.6% of answers this period, up from 19.4%. The gap to the strongest tracked competitor stands at -13.4 pp. Every figure is an estimate from repeated samples of assistant answers, not a count of real buyer conversations.

Named in answers19.4% → 28.6%+9.2 pp over the period
Competitor gap-21.8 pp → -13.4 ppAgainst the best-performing tracked competitor
Work completed
  • Pages refreshed4
  • Source outreach3
  • New pages published2
  • + 2 more lines
Results
  • Newly cited sources7
  • Brand mentions in AI answers34
Next sprint
  1. Get covered on the two remaining review platforms
  2. Publish the head-to-head comparison the answers keep asking for
  3. Refresh the integrations page with current partners
How to read this
  • Visibility is the share of AI answers mentioning the brand, measured across repeated samples on several platforms over weekly windows.
  • There were no untouched topics to compare against in this period, so movement cannot be separated from platform-wide changes.
Abridged preview · example data
Open the full example report →

API: the same numbers, in your own tools

A read-only REST API for your agency’s data. Responses carry the same intervals, sample counts and reasons as the screens, so a figure does not lose its basis when it lands in another dashboard.

Endpoints · GET · Bearer key
GET/api/v1/clientsClients of your agency
GET/api/v1/clients/{id}/visibilityPrompt × assistant matrix, intervals, movement
GET/api/v1/clients/{id}/sourcesCited sources and presence
GET/api/v1/clients/{id}/opportunitiesRanked opportunities with reasons
GET/api/v1/clients/{id}/actionsWork queue with reasons
GET/api/v1/reportsReports and their status
A key is shown once, right after it is created; only a hash of it is stored. Another agency’s client returns “not found”.
# window: 7–90 days, default 28
curl https://app.answertally.example/api/v1/clients/cl_8f2a/visibility?windowDays=28 \
  -H "Authorization: Bearer $ANSWERTALLY_KEY"

// abridged, illustrative response
{
  "data": {
    "client": "Fernpost",
    "windowDays": 28,
    "cells": [
      {
        "prompt": "best project tool for a small design studio",
        "platform": "chatgpt",
        "sharePct": 44.4,
        "interval": [27.1, 62.9],
        "samples": 36,
        "sufficient": true
      }
    ]
  }
}

See it on one of your own clients first

The free audit runs the first steps once, from questions to ranked work, and ends on an opportunity report in your brand. Nothing is charged to run it.

Start a free auditSee pricing →