The Digital Footprint Audit

What this is for

Someone is going to look you up before they hire you, hire your firm, or take a meeting with you, and increasingly they're not using Google to do it. This is for finding out what that lookup actually surfaces, including the old accounts and outdated bios still contradicting how you position yourself now.

Why this has to run in AI tools, not Google

A recruiter, client, or due-diligence analyst doing a background check today is as likely to open Perplexity or ask ChatGPT as they are to type your name into Google. Those tools don't just list links; they read across sources and summarize a conclusion about you ("appears to be a marketing executive with a background in..."). That summary is what forms someone's first impression, and it can be wrong, outdated, or built from a source you forgot existed. You need to see that summary yourself before someone else does.

What these tools can and can't actually see

Be clear-eyed about this before you start:

• They can see what's public and indexed: news mentions, published articles, public LinkedIn and company bio pages, conference listings, public social posts, old company "about us" pages still live on the web.
• They generally cannot see: private social media accounts, content behind logins, most people-search and data-broker sites in full detail (though they may reference that such listings exist), and anything that isn't crawled or indexed at all.
• They can be wrong. AI search tools summarize and sometimes misattribute; two people with your name can get merged into one profile, or an old role can be reported as current. Treat every result as a lead to verify, not a fact.
• Results differ by tool and change over time. What Perplexity surfaces today and what it surfaces in three months won't be identical. This is a spot-check, not a permanent record.

The prompts

Prompt 1: What a recruiter finds

Use Perplexity, or ChatGPT (its web search engages automatically when a prompt needs current information; confirm your plan includes it).

ROLE: You are a recruiter doing a routine background look-up on a
candidate before an interview.

CONTEXT: The candidate's name is [YOUR FULL NAME]. They currently work
in [INDUSTRY/FUNCTION, e.g. B2B marketing, Northeast US] and have
previously worked at [1-2 PAST EMPLOYERS, to disambiguate from other
people with the same name].

CONSTRAINTS:
- Search for and summarize what you find about this person in the
  first 2 minutes of a normal background look-up.
- Note anything that looks outdated (old job titles, old employers
  listed as current) or contradictory across sources.
- If you find information about someone else with the same name, say
  so explicitly rather than blending it in.

OUTPUT FORMAT: A short paragraph, as if reporting back to a hiring
manager, followed by a bullet list of every source URL you used.

Prompt 2: What a client finds

Same tool as above. Run this separately from Prompt 1; a client is evaluating credibility, not employability, and the AI's framing changes with the role you give it.

ROLE: You are a prospective client doing due diligence on a consultant
or vendor before a first meeting.

CONTEXT: The person's name is [YOUR FULL NAME], and they present
themselves professionally as [YOUR CURRENT POSITIONING, e.g. an
independent operations consultant for mid-size manufacturers].

CONSTRAINTS:
- Search for and report what you find that would inform whether to
  trust this person with a paid engagement: credibility signals, past
  client or employer names, any public reviews or testimonials,
  anything that looks inconsistent with the stated positioning above.
- Explicitly flag anything you find that contradicts or undercuts the
  stated positioning.

OUTPUT FORMAT: A short credibility summary (as if writing a note to a
colleague deciding whether to proceed), plus source URLs.

Prompt 3: The red-team pass on yourself

The least comfortable prompt and the most useful one. Run it last. Note the framing: this is written as a self-audit, because asking an AI to dig up leverage "on a person" trips most tools' safety rules and gets you a refusal. Auditing your own public exposure is a legitimate request, and saying so up front is what makes the prompt work.

ROLE: You are a security researcher helping me red-team my own public
footprint. I am [YOUR FULL NAME], located in [CITY/REGION], working
in [INDUSTRY]. This is a self-audit of my own publicly available
information, with my consent.

CONSTRAINTS:
- Search for anything publicly available that could be used against
  me professionally: old controversial posts, litigation records,
  negative press, publicly indexed financial or property records, or
  contradictions between my public claims and public records.
- Do not fabricate anything. If nothing turns up, say so plainly
  instead of stretching a weak result.

OUTPUT FORMAT: A blunt list of findings, each with the source URL, or
a single line stating nothing notable was found.

If you still get a refusal, that's the tool's policy against targeting individuals doing its job; restate that you are the person being searched and that this is a self-audit, and most tools will proceed.

Prompt 4: Old accounts and bios still live and contradicting you

Use Perplexity for this one; it tends to be better at listing distinct sources rather than folding them into one narrative summary.

ROLE: You are a researcher cataloging every public profile and bio
associated with a specific name across the web.

CONTEXT: Name: [YOUR FULL NAME]. Current professional positioning:
[1 SENTENCE, e.g. Fractional CFO for early-stage SaaS companies].
Past employers/roles for disambiguation: [LIST 2-3].

CONSTRAINTS:
- Find every distinct public profile, bio, or "about" page you can
  locate: old company team pages, conference speaker bios, alumni
  pages, old personal or company websites, directory listings, cached
  social profiles.
- For each, note whether it's still live, and whether the title,
  employer, or description on it conflicts with the current
  positioning I gave you.
- Sort the list with the most contradictory or most visible ones
  first.

OUTPUT FORMAT: A table: URL, what it says about me, live/dead if you
can tell, conflict level (high/medium/none).

Step 5: The image check (not a prompt)

Chat tools are unreliable at image search, so do this one directly; it takes five minutes:

1. Google Images: search your name in quotes, plus your name with your city and your industry. Note every photo of you that appears and where it's hosted.
2. Bing Visual Search / Google reverse image search: upload your current headshot and see where it appears; then do the same with any old headshot you know exists. This catches your photo on pages that don't mention your name in text.
3. While you're in Google: set up "Results about you" (in your Google account settings, or myactivity.google.com/results-about-you). It's free, first-party, monitors search results containing your contact info on an ongoing basis, and lets you request removal of results exposing your phone, address, or email. For the ongoing-monitoring half of this audit, it does the job automatically.

What to do with what you find

Triage everything you surfaced into three buckets:

1. Fix now (you control it): stale bios on your own sites and profiles, old titles on pages you can edit or request edits to. This is usually most of the list and costs an afternoon.
2. Request removal (someone else controls it): data-broker and people-search listings. That's a process of its own, and the Delete Kit is the step-by-step for it, including California's one-request DROP system.
3. Live with it (public record): property records, court filings, old press. You can't remove these; you can make sure the first page of results tells your current story so they're context, not headline.

Where this goes wrong

• You run one prompt and call it done. Each one asks the AI to adopt a different vantage point, and each surfaces different results even for the same name. Recruiter framing misses what the red-team framing catches, and vice versa. Run all of them.
• You treat every result as confirmed fact. AI search tools blend and sometimes misattribute information, especially for common names. Click through to the actual source URL before you act on anything, especially anything negative.
• You skip this because "nothing bad exists." The point isn't just finding bad information. It's finding the outdated bio calling you a "Manager" three titles ago, or the old company page still listing you as an active employee two jobs later. That's the more common problem, and it's the easiest one to fix.

The 2-minute version

Open Perplexity. Run Prompt 1 (recruiter view) with your name and current role. Skim for anything outdated or contradictory. If you have more time, run Prompt 4 to catch every stale bio still live (that's usually where the real cleanup work is), then turn on Google's "Results about you" so the monitoring runs without you.

---

Want to know where you actually stand with AI? The free AI Readiness Assessment takes 10 minutes and gives you a personalized roadmap.

Weekly, on Tuesdays. The Gen Xcelerator: practical AI intelligence for people who already know how to do the job. Subscribe free.

Experience Is the API. GenXcelerate