What is the Cogent prompt library?
The Cogent prompt library is a free AI prompt library for travel and expense: 50 prompts organised into ten categories from air sourcing to duty of care. If you have ever wondered what to ask AI about your programme, this is the answer, and these are AI prompts for business use rather than novelty. Every prompt is a complete brief drawn from live enterprise deployments or written against confirmed Cogent capabilities, so it can be copied verbatim and run on your own live T&E data. Answers come back validated, cited and exportable.
What should I ask AI about travel and expense? 50 prompt examples
Ask for outcomes, not data: the drivers behind a spend increase, the modelled impact of a policy change, the off-channel bookings your system never saw, the negotiation position before a renewal. The 50 prompts on this page are real briefs from live enterprise deployments, organised by job: sourcing, hotel programme, savings, policy, compliance, sustainability, forecasting and executive reporting.
How do I write a good AI prompt for T&E reporting?
Brief it like an analyst: what you want, the timeframe, and the output you need. "Show all travellers who booked above the hotel rate cap in Q2, with total overspend by department" will return exactly that, cited. It's natural language: no query syntax, no keywords, no special phrasing. If the brief is ambiguous, Cogent asks a clarifying question before it runs, and every answer is explainable: the reasoning comes back with the figure.
Are these AI prompts free?
Yes. All 50 prompts are free to view and copy, with no email, sign-up or gate. Running them against your own live T&E data requires a Cogent connection, which a demo can set up.
Do these work as ChatGPT prompts too?
As ChatGPT prompts they are portable, and the wording works in any AI chat tool. But a general-purpose chatbot has no access to your booking, card, expense, HR or finance data, so it can only answer in the abstract. Cogent runs the same brief against your live, governed, consolidated T&E data and returns a cited, validated answer rather than a plausible-sounding one.
Can I schedule a prompt to re-run automatically?
Not in Cogent. Every prompt here runs on demand, so the cited answer is always current when you ask. When one brief becomes a standing deliverable, hand it to Orchestra: it runs the project on a schedule and delivers the finished pack.
Will these prompts work on our data?
Yes, once your feeds are connected. Cogent runs on the TMC, card, expense, HR and finance systems you already have, through 200+ maintained connectors, and every answer respects your permissions down to row and field level. A live demo runs them on data shaped like yours.
What is the difference between a prompt and a project?
Scope. A prompt is a single brief with a direct, cited answer in seconds. A project is a multi-step engagement: a plan you approve, specialists executing in parallel, a validation pass and a boardroom-ready deliverable. A prompt that surfaces something significant can be promoted into an Orchestra project without losing context.
Do I need prompt engineering skills to use Cogent?
No. The prompt library ships pre-built, so the prompt writing is already done: every brief on this page is copy-and-run, and plain business language works because intent interpretation determines what is being asked. If you have searched how to prompt AI or how to structure LLM prompts, the short answer is that you do not need to here. Nobody on your team has to become a prompt expert to get validated answers.
.png)
