AI for insights and analytics
The same idea everywhere: you ask in your own words, the model chooses the measure, a tool computes the figure from your files, and every line links to its rows. What changes from one industry to the next is the questions, the measures and the files. One page for each.
AI sales analytics for B2B teams: ask where revenue comes from, who is uncovered and what went quiet, and get figures a tool computed from ledger and CRM.
AI for FP&A: ask what is driving the cost increase, how much is timing and where headcount is above plan, and get a bridge that sums to the variance.
Each page has the questions people in that industry ask, the measures that answer them with their formulas and sources, a worked conversation with figures that sum, and the files it reads.
Net flows by client and strategy, redemption watch, share of intermediary flows and coverage at cadence.
Lending-only relationships, deposit flight, products against sector norm and return per relationship.
Strict control coverage, testing on schedule, issue ageing by owner, training against role and evidence.
Category gaps by trade, dormant trade accounts, credit headroom, quote conversion and price against terms.
Utilisation against realisation, proposal win rate, sold work against the bench and project margin against plan.
Contact volume against each account's baseline, SLA by contract, surge and silence lists, effort against revenue.
Dormant accounts, contribution after cost to serve and price leakage, answered in plain words from the ledger.
Licence utilisation by school and trust, renewal watch, uplift realised and budget cycle coverage.
What is driving the cost increase, actual against budget and forecast, headcount against plan, timing.
Single customer view, products against segment norm, opened and unused, attrition and contact recency.
What is driving revenue, distribution voids, promotions that paid back, and share from the market data you upload.
Kitchens dropping a delivery day, order guide slippage, shorts and substitutions, contribution per drop.
Margin per load after claims and detention, primary tender share, tender rejection and dormant lanes.
Contract compliance per facility, invoiced against contract price, tier earned against priced, facility mapping.
Corporate production against commitment, pace by segment, property penetration per account, group wash and channel cost.
Quotes that never convert, categories bought elsewhere, counter sales with no account, price against contract.
Aftermarket attach against fleet norms, installed base completeness, contract renewal and catalogue fit.
Renewal watch, income retention against count retention, lines against sector norm and remarketing activity.
Share of client fee wallet, pitch to mandate conversion, wallet-weighted recency and senior contact breadth.
Dormant clients by prior fees, practice areas against norm, realisation and lock-up by client, cross-referral.
Operator wallet share by basin, activity-weighted coverage, service lines per operator and rental utilisation.
Access-aware reach and frequency, calls on blocked accounts, pull-through after an access win.
Savings realised against claimed, maverick spend by requester, contracted share, terms compliance and supplier performance.
Net interest income split into balance, rate and mix, margin by product, fees by segment, accounts never funded.
Net revenue retention by cohort, seat utilisation before renewal, whitespace reconciled to ARR.
Value coverage at cadence, dormant accounts by prior value, share of wallet, win rate from a stated stage, rep load.
Gross profit per shipment and TEU, trade lane share, detention and demurrage, transit reliability.
Delivery ratio per partner, renewal against delivered value, inventory utilisation, hospitality yield.
Supplier OTIF against baseline, forecast bias by SKU and site, inventory ageing and single-source exposure.
Fee against size norm, service lines per client, season watch, advisory triggers and realisation.
Estate currency, orphan and ghost services, site penetration, product attach per site, renewal coverage.
Hit ratio by count and by size, inquiry share, the quiet client list, axe hit rate and traded unsent.
Net new assets after markets and fees, held-away assets, outflow watch and contact recency by value.
The AI model chooses the measure and writes the explanation. Deterministic tools compute every figure, and each one carries an identity that must hold before it is shown.
A number in an answer is a link to the rows that make it, so the meeting is about what to do, not whether the figure is right.
Start with the exports you already produce. Every source, from a file to a connection to your systems, is listed on the upload page.