Overture · Executive reporting for corporate travel
Executive reporting is how a travel programme reaches the people who fund it. Overture writes that executive briefing every month from your own data: what changed, what it cost, and what needs a decision. Nobody on the travel team assembles it.
The month is explained before anyone asks you to explain it.
Executives stop reading travel dashboards because a dashboard hands them homework. It shows the figures and leaves the comparison, the interpretation and the conclusion to a reader for whom travel is one line in a much longer brief.
One landing page shows every role the same screen. Each of them needs a different one.
"It's kind of like if the New York Times had the same stories every day, you'd kind of stop reading it, right?"
.png)
Every figure leadership needs is already in the platform, unread. The problem is the reading.
Executive reporting presents performance to senior leaders in a format built for decisions, not analysis. For a corporate travel programme: one monthly read of spend against plan, policy compliance, savings, market conditions and actions.
It replaces the landing page your travel platform opens on. Everyone lands on the Briefing, composed from their own role and data scope, and the Workspace sits beneath it for the people who have to act. Both read from one detection run, so the number the CFO quotes is the number the travel manager is working from.
It carries you from an insight to an action: "one employee booked three out-of-policy trips in 30 days, go and coach them" arrives as a card with a name on it, not a chart to interpret. Open-ended analysis lives in the wider platform.
Overture builds each month's briefing in five steps, and every one of them finishes before you log in.
Bounded detection runs over the data warehouse on a scheduled refresh. Insights arrive as computed facts with the evidence attached.
Cogent, the travel and expense assistant, writes the story for each insight from templated prompts that require that evidence. This is narrative reporting built on natural language generation, anchored to the figures it describes.
Role and data scope decide which surface you land on, which insights you see, and how each one is framed.
The page paints from cached content; the model never runs on page load.
Cogent is on the page, scoped to the same monthly snapshot everything above was composed from.
Overture reads a monthly snapshot. Re-run the pipeline mid-month and it still shows the month that closed, so the words and the numbers stay locked together. Consistency beats currency here, by design.
The monthly executive briefing is a written account of the travel programme, composed from configurable modules and arranged for whoever opens it. It reaches the business unit owner as a short email first: spend against prior year, position against plan, identified savings and the low-effort subset of them, with a link into the full read.
Avoidable spend is travel spend that runs outside the terms the programme intends, and that would stop once someone with the authority to act on it sees it.
The month in a few sentences, what changed and why it matters. Every figure inside the prose carries its own drill affordance.
Year-to-date spend, budget pacing, compliance, average cost per trip, CO2 year to date and avoidable spend, each showing its movement against prior year or plan.
What the programme banked this period, sitting beside what is still on the table.
The one thing worth doing this quarter, written for an executive reading at a glance, with follow-up questions seeded as one-click chips.
Monthly spend against budget and prior year, forecast on a toggle, and a year-end outlook that gives both numbers: where the year closes on current trend, and where it closes with the recommended moves applied.
Category and segment splits, each carrying its own compliance position, plus the external drivers: jet fuel index, carrier load factor and currency basket.
Every drill-down does the same work, explaining the driver behind the number: open the compliance tile and the panel names the gap to target, the components dragging it down, and what would move it. Segment comparisons carry explain popovers, so the surface holds up when somebody interrogates it live in the room.
The briefing arrives every month, already written. The quarterly business review still happens; this is what it is built on. Three months of writing done and evidenced, waiting for the meeting.
The order never changes, so the reader learns the travel programme briefing once and then reads it in minutes every month. What sits inside that order does change, because layout is templated per client and the data contract can rearrange modules.
Overture renders one insight system at the altitude each reader works at. One detection run sits behind all of it, so two people reading two layers quote the same figure. Everyone lands on the Briefing, and the Workspace is one opportunity away rather than one login away.
The Briefing has already done the "what changed" job at your scope, and the Workspace is where you act on it. Opportunities are curated at global scope or computed per transaction, so narrowing the filter re-cuts what applies rather than re-running the month.
The same detection, scoped to the deterministic set: out-of-policy travellers, off-channel bookings, spend variation, your own top travellers. Sourcing language is stripped out, because those levers sit above their authority.
Business travel reporting belongs to the programme, whatever tool renders it: spend against last year and against plan, travel policy compliance, savings identified and the subset that is easy to take, supplier and category breakdowns, and the market conditions moving the cost base.
Compliance arrives as five named KPIs, each scored against the target you set and each showing how many points short it sits, or that it is on target
Emissions report on the same cadence as spend: CO2 year to date sits beside average cost per trip, and intensity per trip moves as carrier mix changes.
One further module does work most reporting leaves out.
Puts the market next to your own spend. When air spend rises because carrier load factors on your routes rose, the briefing says so, and what that means for fares. That decides whether the next conversation is about how your travellers booked or about what the market charged, and those are two very different meetings.
Reporting for a UK business unit sits inside the same briefing. Currency is configured per cut, so that unit reads its own figures with no conversion step.
The Workspace is Overture's second layer, the one where the work happens. The Briefing narrates what changed and surfaces the few opportunities worth attention; opening one drops you into the Workspace, where that opportunity becomes a piece of work with a target, a value and a way to close it.
There is no toggle between two products. Every Overture user lands on the Briefing, and the Workspace sits beneath it, one opportunity away.
Programme-wide opportunities, such as missing cities in the hotel programme, are curated and exist only at global scope. Deterministic ones are computed per transaction and re-cut as the scope moves: off-channel bookings, out-of-policy travellers, spend variation, top travellers.
Narrow the scope and anything that no longer applies is not rendered. A programme-wide opportunity does not shrink to fit a segment. It stops applying. Dimming was tried and reversed: a greyed row reads as broken rather than as out of scope.
Travellers are named in the Workspace and nowhere else. Not on the Briefing, not in the monthly email digest, not in anything distributed. That is a design rule, not a setting.
Every opportunity opens on the same template, whatever kind it is. The Workspace does not have a different screen per opportunity type, so the second one you open works exactly like the first.
One thing the template deliberately does not do. It does not execute: every action is drafted through Cogent rather than fired from the page, so nothing leaves Overture that a person has not read.
Because an opportunity that names a lever the reader cannot pull is not information, it is an accusation with a number attached. Tell a line manager their unit would be cheaper if Sourcing renegotiated the hotel rate and you have told them their team is a cost problem and handed them no remedy.
Overture's two opportunity categories divide along exactly that line.
Curated, global scope, and sourcing-shaped: the hotel programme, contracted coverage, where the programme itself has a gap. These belong to the people who can open a negotiation.
Computed per transaction and scoped to whatever the reader holds: out-of-policy travellers, off-channel bookings, spend variation, their own top travellers. These belong to the person who can change the behaviour.
That is why the team-lead view is a coaching list and not a scorecard. The money is the same either way. Only the framing changes.
Cogent is reached for, used and closed, not kept open. The travel and expense assistant is included for every Overture user whether or not they hold a separate Cogent subscription, and there is no upgrade to buy. What follows is the shape of one use of it.
One task, one draft, and a clear edge where it stops.
See the Workspace on your own spendExecutive reporting summarises a period for the people who fund the programme. Operational reporting carries the day-to-day detail the programme runs on. Overture is executive reporting with a layer beneath it: a Briefing that narrates what changed, and a Workspace that turns each opportunity into an action with an owner and a value. Both come from one detection run.
What separates the altitudes is what the reader is expected to do next. An executive is being asked to decide, so the month arrives interpreted. A travel manager is being asked to act, so the month arrives as a worklist. One detection run, two jobs.
No. Overture points at what is worth opening and hands off for the rest. The transaction-level work stays in the analytics dashboards you already have, and "Explore more" carries your filters into them rather than rebuilding them here.
That boundary is a design decision, and it shows in three of the things Overture deliberately does not have.
The test we hold it to: Overture should shorten the list of things worth opening. It should never become another thing to open.
Because detection and writing are separate steps, and only the writing is done by AI. Detection is deterministic and runs at the data warehouse. The model narrates what detection found, from templated prompts that require the evidence behind each claim. Take the AI away and every number on the page is unchanged. You lose the writing, not the figures.
Insights are computed facts. AI narrates them.
Written at refresh time from templated prompts, before anyone opens the page.
Every reader sees which slice of data they are on, and the header carries a dated stamp for the period the briefing covers.
A department user sees their department, and the metrics change when their filter applies. Which cut a reader gets is enforced by their role rather than chosen by them.
Cogent reads the same monthly snapshot the Briefing was written from, so whatever it drafts agrees with what you just read. When a question needs transaction-level detail, Overture hands off to the wider travel and expense platform.
Overture records what the travel team achieved in the period and attributes it, so the contribution is documented by the reporting itself.
Each win names the team that delivered it, carries the value achieved, and says how the value lands, written up beside the spend figures by the system that detected them.
Stakeholder management is why this surface exists. When leadership understands what the travel function does, it treats that function as a strategic partner rather than a transaction desk, and the team gets more of what it asks for.
Management reporting is the operating detail a function runs on: volumes, categories, exceptions, month-to-month movement. Executive reporting is the layer above it, the same programme summarised for a decision with the analysis already done. Most organisations have the first and improvise the second.
Against a traditional reporting platform, the real choice is what the surface does with the numbers. One shows them. The other tells you what they did, with the detection, the narration and the travel data model already built. The alternative is an 18-month in-house BI programme.
Overture works alongside your analytics dashboards. It tells you where to look, then hands off for the transaction-level work. And it is newly announced, so an early conversation shapes the configuration.
Good management reporting for a travel programme follows five rules, and none of them are about adding charts.
If the Briefing and the Workspace can disagree, both lose their authority.
A compliance figure that moved is only useful beside the components that moved it.
Fuel, load factors and currency move spend without any traveller behaving differently.
AI insights someone cannot act on read as criticism.
A report that arrives when somebody has time to build it arrives late.
The last rule is the one that fails manually, which is why programmes drift to reporting quarterly and then find a pattern that has been running for two months. Automatically generated is the only version of monthly that holds.
We build the briefing against your programme: your categories, your cuts, your compliance definitions. You see what your executives would receive, on your own figures. What you ask for now is what gets configured.
