AI in investigations, responsibly.

What automation is good at

Most of the time in a genetic genealogy case goes to work that's large, repetitive and rule-bound:

• Sorting hundreds of DNA matches and estimating how each one is related
• Grouping matches into the family lines they share
• Building trees from public records, generation by generation
• Checking the people in those trees against other records
• Watching for new matches after the first search

Software does this faster than a person and doesn't get tired at the thousandth record. That's where Indago spends its effort.

What stays with people

Judgment stays with your team. That covers whether a case qualifies, which relationship fits when the DNA allows several, whether two records describe the same person, and above all whether a candidate becomes a lead. Those calls carry legal and human consequences. They belong to investigators, genealogists and prosecutors who can be asked to explain them.

Five principles we build on

1. Leads, not verdicts

Everything Indago produces is an investigative lead. A lead is a reason to look, not a finding. It's corroborated through traditional investigation and confirmed by a direct DNA comparison before anyone acts on it.

2. People approve

Automation proposes and investigators approve. Nothing becomes a lead in Indago until a person signs off on it.

3. Show the reasons

A suggestion you can't inspect is a suggestion you can't defend. Every relationship estimate, tree connection and candidate in Indago comes with the evidence behind it and how confident the estimate is. If the support is thin, you should be able to see that at a glance.

4. Keep the record

Cases get reviewed by supervisors, by prosecutors and in court. Indago records what was found, what was proposed, who reviewed it and when. Your team can always reconstruct how a lead came about.

5. Authorized data, kept inside

Indago works only with data your agency is permitted to use, under each source's terms. Your case data never trains our AI and never goes to an outside AI service. It's returned or deleted when your contract ends.

Know the limits

Good tools are honest about where they can be wrong. In genetic genealogy, the common traps are:

• Overlapping relationships. The same amount of shared DNA fits several relationships. An estimate is a probability, not a fact.
• Endogamy and pedigree collapse. In communities where relatives often marry, people share more DNA than their relationship suggests. That makes matches look closer than they are.
• Unexpected parentage. Adoptions, donor conception and undisclosed parentage break the link between the paper trail and the DNA.
• Record errors. Public records can be wrong, stale, duplicated or attached to the wrong person.
• Thin data. A case with few or distant matches gives any method, human or automated, less to work with.

None of these is a reason to avoid automation. Each is a reason to keep a person reviewing its output.

Questions to ask any vendor

• Can my team see why the tool made each suggestion?
• Does anything become a lead without a person approving it?
• Is every action recorded, with who did it and when?
• Where is our case data stored and processed? Is it used to train models, or sent to an outside AI service?
• What happens to our data when the contract ends?
• Can we run the tool on a case we've already solved and check its answer?

We'll answer every one of these for Indago, and we'd encourage you to ask them of anyone else. The last one is the best test there is. Bring us a solved case and check what Indago finds against the answer you already know.

For the method itself, read What is investigative genetic genealogy? For how we protect case data, see the Trust center.