How to Build an AI Visibility Dashboard for Free

You do not need a five-figure platform to know whether AI engines cite you. A free dashboard built from tools you already have will answer the only questions that matter: where you show up, where you do not, and whether that is changing. This guide walks through building one in an afternoon.

What the Dashboard Should Answer

Before you open a spreadsheet, write down the three questions your dashboard must answer. A useful starting set is: which of my target queries return my brand in an AI answer, how does that compare to last month, and where do competitors appear instead of me. Everything you build serves those questions.

A dashboard that tracks everything tracks nothing. Pick a small, decision-driving set of metrics and let the rest wait until you have a habit of reading the first version.

The Free Stack

You can assemble the whole system without paid software:

• A spreadsheet (Google Sheets or any equivalent) as the data home.
• Looker Studio, a free visualization layer that connects to that spreadsheet.
• Google Search Console, free, for crawl and impression context.
• A manual prompt-testing log where you record exactly what each AI engine answered.

The spreadsheet is the source of truth. The other pieces either feed it or display it. Keep that separation so you can swap any single tool later without rebuilding the whole thing.

Step 1: Define Your Query Set

List the twenty to fifty questions a potential customer might ask an AI engine that your product could answer. Write them as natural questions, not keywords. "What is the best tool for managing AI crawlers" beats "AI crawler tool." Save this list in the first tab of your sheet.

Step 2: Build the Tracking Tab

Create a tab with one row per query and columns for date, engine, whether your brand appeared, the position or phrasing of the mention, and any competitor named. Run each query through each engine once, paste the results, and you have week one of data.

Repeat this on a fixed cadence, weekly or biweekly. Consistency matters more than frequency; a steady rhythm beats an occasional deep dive.

Step 3: Add a Perception Log

In a second tab, log how engines describe your brand, not just whether they name it. Capture the exact phrasing. A mention that calls you "a budget option" when you sell enterprise software is a problem the appearance column would hide.

Step 4: Connect Looker Studio

Point Looker Studio at your spreadsheet. Build one line chart for mention rate over time and one table ranking queries by your presence. You do not need fancy visuals; a clear trend line is worth more than a pretty gauge.

Step 5: Set a Review Ritual

Block fifteen minutes after each data refresh to read the dashboard and write one sentence on what changed. That sentence is what keeps the project alive when the novelty fades.

Metrics Worth Tracking

Track mention rate, the share of queries where you appear, and competitor overlap, the share where a rival appears and you do not. Those two numbers tell you most of the story. Add a freshness flag for queries you have not tested in thirty days so stale data does not masquerade as stable.

Reading the Dashboard Honestly

A flat line is not failure. It may mean your content is holding while a competitor slipped. Compare your line to the competitor overlap line before drawing conclusions. The point is to notice movement early, not to celebrate a single good week.

Example Dashboard Layout

A useful first version needs only two visuals. The first is a line chart of mention rate over the last eight to twelve refresh cycles, with one line per engine if you test more than one. The second is a table ranking your queries by whether you appear, sorted so the missing ones rise to the top and demand attention.

Keep competitor overlap as a third element once you have the habit. A simple stacked view showing "you," "competitor," and "neither" per query group makes prioritization obvious in a glance during a monthly review.

Picking Your Refresh Cadence

Weekly refresh catches drift early but costs more effort; biweekly is a fair compromise for most small teams. The wrong answer is "whenever I remember," because gaps in testing hide exactly the moments you would want to see. Tie the refresh to a calendar slot, not a mood.

If you test several engines, stagger them so no single week becomes a wall of work. A steady, lighter rhythm beats a quarterly panic where you re-test everything at once and cannot compare like for like.

Why Free Beats Expensive Early

Paid platforms shine when query volume, refresh frequency, and team size all climb. Early on, none of those justify the cost. A free stack forces you to define what matters, which is the part paid tools cannot do for you. The discipline of picking twenty queries and reading the results weekly is worth more than a dashboard that tracks ten thousand you will never open.

When you do outgrow free, you will know exactly which capability to buy, because you will have felt the gap. That is a stronger buying case than any sales demo.

Common Dashboard Mistakes

The first mistake is tracking too many queries too soon, which turns the weekly refresh into a chore and stops the habit. The second is logging only appearances and missing phrasing, so a damaging description hides behind a green "mentioned" cell. The third is never reviewing the data, which makes the whole effort a graveyard of stale numbers.

Avoid all three by keeping the query set small, adding the perception log from the start, and blocking fifteen minutes after every refresh to write one sentence about what changed.

Sharing the Dashboard With Stakeholders

A dashboard nobody sees changes nothing. Export a snapshot to a slide or doc each month and walk stakeholders through three numbers: mention rate, competitor overlap, and the one query that moved. Resist the urge to show every tab; the summary is what earns continued support.

When a number moves, lead with the action you took, not the metric itself. Stakeholders fund responses, not reports, and a dashboard that drives a published fix is easier to keep funded than one that merely informs.

Frequently Asked Questions

How many queries should I track to start?

Start with twenty to fifty of your highest-intent questions. That is enough to show trends without turning the weekly refresh into a part-time job. Expand only once the habit is established.

Is manual testing too slow to be useful?

It is slower than automation, but it is accurate and free, and it teaches you exactly how engines phrase answers. Many teams keep a manual core even after adding paid tools, because the evidence is verifiable.

When should I move from free to paid?

Move to paid when your query set passes a few hundred, when you need daily refresh, or when multiple teammates need shared access and alerts. Until then, the free stack scales further than most teams expect.