Skip to content
← All posts
Planning for 20276 min

Tracking AI answers for agency clients in 2027: a planning guide

By the MentionedOn team

Planning guides written in November 2026 for the year ahead, revised when their facts are re-checked.

To track AI answers for several clients in 2027, give each client a fixed set of questions, run them on the same assistants on the same weekday, keep the dated answers and cited links, and report only changes that hold across several runs. Plan it in January: agree the questions and record a baseline with each client before any work starts, so the December report compares like with like.

Written November 2026. This is a planning guide for 2027, written before the year starts. It says what to do, check and budget for, and what is true today; it does not predict what the assistants will do. We will revise it when we re-check the facts it relies on.

Disclosure. MentionedOn publishes this guide and sells one of the tools it names. Other tools are described only from their own pricing and product pages, each with the date we last read them (between September 2026 and November 16, 2026). Those are 2026 readings, not 2027 prices. Confirm with the vendor before you buy.

January: set up each client

  • Agree ten to thirty questions per client: the service and city, the comparison a customer makes, price and urgency. Get the client to confirm the list in writing; it is the scope of what you will report.
  • Agree the assistants that matter for the client's customers, and keep the list fixed for the year.
  • List competitors twice: the ones the client names, and the ones the answers name. They are often different.
  • Record a baseline for four weeks before any new work goes live.
  • Record the client's facts (name, address, hours, service area) and note where the answers get them wrong.
  • Set up a work log with the date every page, listing or article went live. Without it, no later comparison holds.

The 2027 quarter by quarter plan for a single business is the same plan, one client at a time.

The weekly routine

  1. Run every client's set on the same weekday.
  2. Log the date, question, assistant, businesses named in order and links cited. The weekly agency log template has the columns.
  3. Flag swing questions, where the client is named in some answers but not most (MentionedOn's own term).
  4. Note any answer that cites a page you published for the client.
  5. Add new cited sources to the client's outreach list.

Weekly is a choice, not a rule. A weekly series on a fixed day is easier to read for monthly reporting; daily runs help while a client is in the middle of a change.

The monthly client report

Keep it to what the client can check:

  • Questions where the client was named, by assistant, as counts such as 3 of 4 runs.
  • Questions a competitor won, and the cited pages behind those wins.
  • Work shipped, with dates.
  • What changed since the baseline, and whether it held across several runs.
  • What you do not know yet, said plainly.

Share of voice or mention rate across the same questions and weeks gives a client one figure to follow. Report a change only when it holds, and do not present a change after your work as proof your work caused it.

Why the caution: in our destination report review, every one of 30 list pairs changed between the two most recent stored reports. That was a look back, not a controlled test, and it covered places to stay, eat and visit rather than local services, but it is a reason to wait for a change to hold. Do not carry one client's numbers over to another: in our eight city plumbing snapshot the most mentioned plumber appeared in anywhere from 5 to 10 of 10 sampled answers, depending on the city. Clients who check on their own phones may see answers move too. Why the app and a report can differ is a useful page to send them.

Quarterly reviews

  • Q1: present the baseline and agree two or three target questions per client.
  • Q2: ship the fixes and pages, with dates.
  • Q3: compare equal windows before and after each publication date.
  • Q4: compare December with January and agree next year's questions.

Pricing the work for clients

Price from your own numbers, not someone else's benchmark:

  1. Time one client's weekly round, by hand or in the tool you use.
  2. Add the reporting time per client per month.
  3. Add the tool cost per client. For example, MentionedOn's Growth plan is $399 a month for up to 5 businesses, so the tool cost per client depends on how many slots you fill.
  4. Price the fixes separately, since they vary most from client to client.

The 2027 budgeting guide has a worksheet you can use per client.

Choosing a tool for several clients

The facts below are each tool's own statements as we read them, with the date. They are 2026 readings.

  • BrightLocal: "White-labelled tools and reports". Starts at $31/mo billed annually ($369/yr) or $41/mo billed monthly (Track, 1 location) (checked November 12, 2026).
  • Local Falcon: "White-label reports listed among its plan features". Starts at $24.99/mo billed monthly (Starter) (checked November 12, 2026).
  • Peec AI: "Agency pricing for multiple brands". Starts at $95/mo billed monthly (Starter) (checked November 12, 2026).
  • Otterly.AI: "Agency pricing" and "API and MCP requests from Standard". Starts at €29/month (Lite) (checked November 12, 2026).
  • Scrunch AI: "Separate plans for agencies". Starts at $250/mo (Core) (checked November 12, 2026).
  • Trakkr: "White-label client portals on Scale". Starts at $100/mo billed monthly or $83/mo billed yearly (Growth) (checked November 12, 2026).
  • Rankscale: "REST API from Pro+". Starts at $17/mo, yearly billing only (Essentials); Pro $99/mo billed monthly (checked November 12, 2026).

MentionedOn, which publishes this guide: Growth covers up to 5 client businesses with 120 tracked questions each, 25 tracked competitors and 10 seats, a plain-English weekly report for each business and CSV exports, for $399 a month. It has no white-label reports on any plan, so if reports under your own brand are a requirement, one of the tools above that lists white-label is the better buy. More on the agency setup is on the agencies page, and the 2027 buyer's checklist covers what to test in a trial.

Client conversations to plan for

  • "ChatGPT said something different on my phone." That can happen; send this explanation and show the series rather than one answer.
  • "Can you guarantee we will be recommended?" No one can. This page covers what can and cannot be bought.
  • "How long will it take?" There is no reliable timeline; this page explains why, and the work log shows when to look.
  • "Does our SEO cover this?" Some of it does; the measurement is separate.

FAQ

How many clients can one person track by hand?

Time one client's full round of checks and divide the hours you have by it. The answer depends on the number of questions and assistants per client, so measure it rather than borrow someone else's figure.

Can I put my agency's brand on MentionedOn reports?

No. MentionedOn has no white-label reports on any plan, Concierge included. Its CSV exports can go into your own decks, and the tools listed above that offer white-label reports are the better buy if you need them.

Should I report a score or the answers?

Both. A score is easy to chart, and the dated answers and cited links behind it are what let a client check it.

What should I promise a client for 2027?

A fixed set of questions, checked on a fixed schedule, with dated answers and an honest read of what changed. Not a recommendation, because no agency or tool controls what an assistant says.

What would change this advice

  • A tool listed here changes its agency features or prices. Each line updates with its new check date when we re-read the vendor's pages.
  • MentionedOn adds white-label reports or changes the Growth plan. We would update this guide when it ships.
  • An assistant your clients' customers use changes how it cites sources, which would change what goes in the monthly report.
  • Your clients' customers move to an assistant you do not track. Add it as a new series from the date you start.

Questions or corrections: hello@mentioned-on.com.

See where you actually stand

The trial covers ChatGPT, Claude, Gemini, Perplexity, Grok and Copilot: all six when it starts and in the Monday sweep, with daily checks on ChatGPT and Gemini.

$0 today, then $99/month. Card required. Cancel anytime.