Sign in
Set up for Commercial banking

AI for insights and analytics

AI analytics for commercial banking

A relationship manager's book has clients who borrow and bank elsewhere, deposits that are quietly leaving, and facilities drawn to the limit. AI analytics reads the product, balance and facility files, compares each client with its sector, and answers the team head's question in figures that tie to the ledger.

Ask it in your own words

These are the questions commercial banks ask. Each one maps to a measure our tools compute from your files. The AI model chooses the measure and explains the result; the arithmetic is done by our code, and every total is checked.

The questionThe measure behind itWhat comes back
Which clients borrow from us and bank elsewhere?Lending-only relationshipsClients holding only credit products, with the products similar clients hold and the estimated fee value of the gap.
Where are deposits leaving?Deposit flightClients whose balances fell against their own baseline, by amount and pace, with the last contact.
Which relationships hold fewer products than their sector norm?Products held against sector normProducts per client against the median for its sector and size, ranked by the gap.
Whose facilities are nearly fully drawn?Facility utilisationDrawn against limit by client, with those over ninety percent and their trend.
Which relationships do not earn their capital?Return per relationshipIncome less cost of capital and servicing per client, with the loss-making ones and why.
Are relationship managers seeing their largest clients?Coverage at cadence, by revenueRevenue of clients met within cadence over revenue assigned, by manager and team.

A conversation, with the figures

Every figure below was computed by a tool from the rows and checked before it was shown. The lines sum; the percentages match; each line opens to the rows that make it.

Where are deposits leaving the book?

Deposit balances are down $167.0m against each client's own baseline across 3 teams, on clients whose lending has not changed. Mid-market team accounts for $84.0m, 50% of the outflow.

Mid-market team$84.0m50% of the total
Real estate team$52.0m31% of the total
Small business team$31.0m19% of the total
The 3 lines sum to$167.0m0 unexplained

Which mid-market clients?

3 clients hold $76.0m of the $84.0m outflow in the Mid-market team, 90% of it. Two are lending-only relationships whose operating accounts were never with us, and the third moved its balances after a rate review elsewhere.

Northfield Manufacturing$31.0mlending-only; operating account elsewhere
Grange Logistics$27.0mbalances down 40% in two months; no contact logged
Ashcombe Foods$18.0mmoved after a competitor rate offer
These 3 are$76.0m90% of Mid-market team

What does the team do with that?

Each client gets a named conversation about the whole relationship, not the loan. The manager goes with the products the client holds elsewhere and what similar clients hold with us, and logs the outcome.

Mid-market relationship managersMeet the three clients with the product gap analysisThis month
Team headReview deposit flight and lending-only lists weeklyWeekly
Head of commercialPut return per relationship in the quarterly reviewQuarterly
Which clients borrow from us and bank elsewhere?Where are deposits leaving?Which relationships hold fewer products than their sector norm?Whose facilities are nearly fully drawn?

The measures behind the answers

Each measure has one formula, one source and one meaning. They are computed per relationship manager and team and in total, and every one carries an identity that must hold before it is shown.

MeasureFormulaFromWhat it tells you
Products held against sector normProducts held ÷ median products held by clients of the same sector and size bandProduct holdings file; client masterRelationships thinner than their peers
Lending-only relationshipsClients with credit exposure and no operating account, payments or fee products; exposure and return on eachHoldings file; facility fileBalance sheet used; fees earned elsewhere
Deposit flightAverage operating balance, last 3 months ÷ same months last year, per relationship; flagged below a stated ratioBalance fileCash management moving to another bank
Facility utilisationDrawn ÷ committed, per facility, with trendFacility fileUnused lines costing capital; clients drawing to the limit
Share of wallet by productBank's revenue from the client in the product ÷ estimated client spend on the productRevenue file; estimates by source gradeWhere the client's fee spend is going
Coverage at cadence, by revenueRevenue of clients contacted within tier cadence ÷ total revenueCRM; revenue fileWhether top relationships are being seen
Referrals between business linesReferrals sent, accepted, converted, and revenue, by origin and destinationReferral log; revenue fileWhether one bank is working as one bank
Return per relationshipRevenue − cost of funds − expected loss − cost to serve, over capital usedRevenue, capital and cost filesRelationships that do not earn their capital
Portfolio load against capacityClients and required touches per relationship manager against available hoursCRM; client tiersPortfolios too large to cover
Primary bank indicatorsShare of client turnover seen through the account; payroll and tax payments presentTransaction file; client turnoverWhether the bank is the client's main bank

Each measure is worked through, with the export it comes from and what to drop, in Relationship KPIs for commercial banking.

What the AI model does, and what our tools do

The AI model chooses the measure

From your question and the measures declared for commercial banking, the model picks the one that answers it, and the period and comparison the question implies.

Our tools do the arithmetic

Deterministic code reads the rows, computes the measure, and checks the identities below. The same question on the same data gives the same answer, every time.

The AI model explains, and cites

The AI model writes the sentence around the result, naming the relationship manager or account behind it. It states no figure that is not in the result, and every figure links to its rows.

The identities that must hold

HoldingsProduct revenue per client sums to the revenue ledger
BalancesRelationship balances sum to the general ledger deposit total
ReferralsSent = accepted + declined + pending; accepted = converted + lost + open
CoverageClients = covered + uncovered; each with one owner

What it reads

The exports commercial banks already produce. Column names are mapped once and the mapping is reused. A file is the way in; scheduled delivery and connections to your systems come with the plan, and every source is listed here.

  • Product holdings file
  • Client master
  • Holdings file
  • Facility file
  • Balance file
  • Revenue file
  • Estimates by source grade
  • CRM
  • Referral log
  • Revenue, capital and cost files
  • Client tiers
  • Transaction file
  • Client turnover

Who owns each answer

An answer is a list with an owner and a cadence, or it is a chart nobody works.

MeasuresOwnerCadence
Deposit flight; coverage at cadenceTeam heads; relationship managersWeekly to monthly
Lending-only; products against norm; referralsHead of commercial bankingMonthly
Return per relationship; facility utilisationHead of commercial banking with finance and creditQuarterly
Portfolio loadHead of commercial bankingTwice a year, and on every reassignment

Questions commercial banks ask about AI analytics

How does it estimate the value of a product gap?

From your own book. If similar clients in the sector hold five products and this one holds two, the fee income those products earn from similar clients is the estimate, and it is labelled as an estimate. The tool shows the comparison set it used, so the manager can see whether the peers are fair.

Can it see deposit flight before the balance is gone?

It compares each client's balances with their own baseline month by month and flags a fall against pace, not just level. A client down forty percent in two months appears while the balance is still large, which is when the conversation is worth having.

What exports does a commercial bank need?

Products held by client, balances by month, facility limits and drawn amounts, and the CRM activity file. Income by client turns product gaps into return per relationship. Client names can be replaced with identifiers before upload, and are on Enterprise.

Does it reconcile to the ledger?

Every balance figure carries a control total against the file it came from, and the deposit flight lines sum to the change in the book. If a client appears twice under two entities, the tool says so rather than counting it twice, and the roll-up is shown.

See it on your data

Bring a few thousand rows. The data map opens next, every column mapped once, and the first question is answered in minutes. Free, in your browser, no account.