Skip to main content

Getting the Most Out of Reporting Intelligence

Best practice tips for asking Tali the right questions and getting accurate, useful reports back.

Written by Sadie Malkin

Tali is Talos360's AI-powered reporting assistant, part of Reporting Intelligence, and it lets you ask questions about your recruitment data in plain English and get instant, ready-made reports back. There's no report builder to learn and no fields to configure. You just ask, the same way you'd ask a colleague. This guide covers how to phrase your questions so Tali understands exactly what you're after, first time.

Note: you will need access to the 'Can use Reporting' and 'Can use Reporting Intelligence' permissions to use Tali, as this is part of the Reporting+ package. If you don't currently have this feature and want to find out more, reach out to your Account Manager.

Early Access Notice: Reporting Intelligence is currently in early access. Always review and verify any reports produced by Tali before relying on them for decision-making.

Structuring a good question

The clearer your question, the better the report Tali builds. A strong question usually gives Tali three things: what you want to measure, how you want it broken down, and where (which filters) it should apply to.

Compare a weak question with a strong one:

Weak: "How are my vacancies doing?"

Strong: "What is the total number of applications for each vacancy in the Sales team for Q2?"

The strong version tells Tali the metric (applications), the breakdown (by vacancy), the filter (Sales team) and the timeframe (Q2), so there's nothing left for Tali to guess.

Be specific rather than relying on Tali to read between the lines

Tali can handle open-ended questions, but it uses its own judgement to decide how to break the data down, so it's worth being explicit whenever you have something specific in mind.

Avoid ambiguity in rankings. Instead of "our best performing sources", ask "top 5 sources by number of hires for last year" or "sources with the highest number of applications last month for the Sales team". "Best" and "most" can mean different things depending on what you're measuring, so being specific removes the guesswork.

Spell out how a group should be defined, rather than assuming Tali will infer it. Instead of "the London team", if there's any ambiguity about which teams that includes, say "the London team, excluding the London Support team" or "the London team, including London Support".

Always specify a timeframe

Instead of "total applications for the Sales team", say "total applications for the Sales team in Q3". Without a timeframe, Tali has to guess whether you mean this month, this year, or all time, and a report built on the wrong assumption is easy to misread as fact.

Explain the logic behind a calculation, not just the result you want

For anything more complex than a simple count, spell out how you want it worked out. Instead of "percentage of applications from each source for the last three months", say "each source's applications as a percentage of total applications for the last three months, broken down monthly". The extra detail costs you one sentence, and it saves Tali from having to guess at your intent.

Remember: these are guides, not rigid rules. Tali adapts to natural, conversational questions too. Since AI can occasionally interpret things differently than intended, always review the results to make sure they meet your needs.

Start broad, then refine

You don't need to get your question perfect on the first try. A good approach is to start broad, look at what Tali gives you back, then refine from there by adding grouping, filters or a calculation in your next message. It's the same way a conversation with a colleague would naturally narrow down.

For example, you might start with "Show me applications by source this year", see which sources stand out, then refine with "Now just the Sales team" or "Break that down monthly instead", without needing to repeat the original question.

Tali treats a conversation as a continuing thread. Once Tali has answered, you can ask as many follow-up questions as you want in the same conversation, and each one is understood in the context of what came before, so you don't need to repeat detail you've already given it.

Checking Tali's working

Treat a report from Tali the way you would a colleague's: useful, but worth a sanity check before you rely on it. Here are a few ways to verify what you're looking at.

  • Ask how it got there. "How is this calculated?" or "How did you calculate that percentage?" gets you a plain-English explanation of the data and logic behind the figure.

  • Ask why something was excluded. If a number looks lower than expected, try "Why did you exclude those records?" to check whether a filter is doing something you didn't intend.

  • Watch for vague, numberless answers. If Tali gives you a general answer with no specific figures attached, that's usually a sign the question was too broad for it to find a good match in your data, rather than a genuine data-driven result. Try rephrasing with more specifics, along the lines above, rather than taking a vague answer at face value.

  • Review before you rely on it. Since AI can occasionally interpret a question differently than intended, always check a report against what you'd expect before using it for a decision, particularly while Reporting Intelligence remains in early access.

If a question doesn't come out right the first time, that's expected. See Troubleshooting Tali for how to fix the most common issues.

Getting more from Tali's answers: Ask Follow-up and Explain

Whenever Tali writes something back to you, whether that's commentary underneath a chart or report, or a written answer in the Tali chat, you're not limited to just reading it. Highlight any part of that text and a small menu appears with two options, 'Ask follow-up' and 'Explain', so you can carry on the conversation right there without opening a fresh question from scratch every time something catches your eye.

Highlighting text to bring up the menu

To use either option, highlight the specific word, figure or sentence you want to act on. A small menu appears above your selection showing 'Ask follow-up' and 'Explain'.

The more precisely you highlight, the more precisely Tali can respond to either option. Highlighting a single figure or phrase gets you a focused response about that detail, while highlighting a whole sentence gets you a broader one.

The same menu appears the same way inside an ordinary Tali chat answer, not just underneath a dashboard chart. Here, Tali had already suggested a next step in its written answer, "If you'd like to break this down by source or team, just let me know", and highlighting that sentence brings up the same 'Ask follow-up' and 'Explain' options.

Highlighting a sentence in a Tali chat answer showing the 'Ask follow-up' and 'Explain' menu

This is a useful habit to build. Whenever Tali's own answer proposes something ("just let me know" or "let me know if you'd like this broken down further"), you don't need to type that request out yourself, just highlight the suggestion and choose 'Ask follow-up'.

Explain

Choose 'Explain' when you understand the answer overall but want more detail on the specific part you've highlighted, for example a score, a percentage, or a term like "Manual Reject" that you want unpacked before you act on it. Tali responds with more context on just that highlighted piece, without changing or adding to the original text.

Tip: use 'Explain' as your first move whenever an answer references a calculation or outcome you don't recognise. It's the quickest way to understand what a figure means before deciding whether you need to dig further.

Ask follow-up

Choose 'Ask follow-up' when the highlighted part raises a genuinely new question, one that builds on what Tali has told you rather than just asking it to clarify. This keeps your question connected to the exact context Tali was already looking at, so you don't have to repeat details like the date range, team, or vacancy involved.

Tip: treat your follow-up the same way you'd treat any question to Tali, be specific. Instead of "why", try "why did this candidate score 45% on Work Ethic", so Tali has the context it needs to give you a useful answer rather than a generic one.

Explain or Ask follow-up, which should you use?

As a rule of thumb, use Explain when you want more detail on exactly what's highlighted, and use Ask follow-up when you have a new question that goes beyond what's highlighted, such as wanting to see the same figure broken down differently or compared against something else.

Getting a PDF of your conversation

If you want to share what Tali has told you with someone who doesn't use Talos, or keep a copy for your own records, you can ask Tali directly for a PDF rather than screenshotting or recreating the answer elsewhere. Just ask in plain English, for example "Can you give me this as a PDF?", and Tali will generate a downloadable PDF of the conversation for you.

Tip: ask for the PDF once you're happy with the answer. If you go on to refine the question further afterwards, ask again, so your PDF reflects the final version rather than an earlier draft.

Analysing your own data by uploading a CSV

Tali isn't limited to your existing Talos reporting models. If you have a report or spreadsheet that was built outside of Talos, either in a spreadsheet from another system or one you've been keeping manually, you can upload it as a CSV file and ask Tali questions about it directly.

To upload a file, select the '+' icon in the Tali chat input box and choose 'Upload file'.

This is particularly useful if you're trying to work out whether something you currently maintain outside Talos could be recreated inside Reporting+. Upload the data as it stands today, ask Tali the same questions you'd normally answer manually, and see how close Tali gets before deciding whether to build it properly as a report or dashboard.

Did this answer your question?