SME Lending Operating Model: From Silos to Solution Pods
This is Article 2 of 6 in the Q-Lana Customer Centricity series. An SME lending operating model built on solution pods cuts proposal time from three weeks to four days. It covers a library of client problems and a copilot in the RM's workflow.
Most SME lending operating models were built around departments, and clients pay for that design in waiting time. In a typical bank, a proposal for a mid-sized food processor passes through five desks over about three weeks. That delay matters in markets where the MSME finance gap stood at US$5.7 trillion in 2019. This article sets out an SME lending operating model built on small cross-functional solution pods. In the cases we describe, it brings the same proposal down to about four days. Nobody is reorganized and no reporting line changes. What changes is who sits together and works the file.
Where Does the current SME lending operating model lose time?
A relationship manager (RM) spots an opportunity during a site visit. The file then routes to a product specialist, a credit analyst, operations and finally compliance. Each unit works its own queue and passes the documents on after its own review.
The client experiences repeated questions, mixed messages and slow answers. The bank loses context, and nobody owns the outcome from start to finish.
Most of the damage happens at the handovers. Take a processor whose largest buyer pays at 90 days but is negotiating down to 60. The RM hears this over coffee. The CRM note says "payment terms under review". When the credit analyst models working capital, the 60-day prospect has dropped out of the file. Nobody removed it on purpose. The analyst simply was not in the conversation.
How solution pods work in an SME lending operating model
A solution pod is a small cross-functional team serving a defined set of client segments. It is the starting point of Q-Lana's customer centricity model. It has four members, each with a clear remit.
The four roles in a pod:
- The RM leads the client dialogue and the discovery work.
- The credit analyst structures the facility and prepares the credit assessment from inside the pod, not from a central unit.
- The sector specialist knows how cash moves through the segment, including seasonality, buyer concentration and input price swings.
- The product and operations liaison checks that the proposed structure can be booked, documented and disbursed. That person clears blockers before they turn into delays.
Segment focus is key
Each pod covers three to five priority segments, such as food processing, agribusiness trade, manufacturing and construction. Focus is what builds judgment. An analyst who reviews forty grain traders a year learns what a normal post-harvest stock build looks like. She can tell a seasonal overdraft peak from a real liquidity problem without asking for more statements. Over time the pod writes this down as playbooks, so a new analyst starts from the same base.
None of this needs a new structure. Reporting lines stay where they are. Four people from four departments sit together and share one pipeline.
Turnaround times in a food and agribusiness pod
Consider a pod covering about 140 borrowers in food and agribusiness.
Before: five desks in sequence
Under the old SME lending operating model, a proposal moved through five desks one after another. The RM collected documents, and credit placed the file in its queue. The product team priced it. Operations re-keyed the data into the core banking system, and compliance reviewed it last. Three weeks was normal. The client explained its business again at almost every step.
After: one team working in parallel
Inside the pod, one team builds the proposal at the same time. The sector specialist already knows the bakery cash cycle, so nobody asks why flour stocks double in October. The credit analyst starts spreading the financials while the RM is still at the client's premises. Meanwhile the liaison confirms booking and collateral requirements.
Time from discovery to proposal falls from about three weeks to about four days. The client tells its story once instead of four or five times.
Who owns the outcome?
The more important change is accountability. In the relay model, a declined or delayed deal is always partly someone else's fault. However, in a pod, the four members own the outcome together, including what happens after disbursement. Suppose the bakery breaches a covenant in month seven, the team that structured the facility also handles that conversation with the client. Faster turnaround follows from this shared ownership.
A Library of Recurring Client Problems

Client needs mostly follow a small number of patterns. A trader squeezed by late-paying buyers and a processor facing its annual pre-harvest gap look different on paper. Underneath, both need short-term liquidity tied to a predictable cash conversion cycle. Once a bank sees this, it can prepare its answers in advance. The idea draws on Clayton Christensen's jobs-to-be-done theory.
What each library entry contains
A Jobs-to-Be-Done Library records these patterns as entries in the loan management system. Each entry starts with the problem archetype, named the way a client would put it. A typical example is "my buyers are paying later". Next come the data signals that flag the problem early and can be monitored continuously. The entry then sets out a solution toolkit of products, covenants and monitoring rules. That toolkit is pre-cleared within the risk appetite framework. Finally, the entry defines the success metrics the bank will track once the facility is live.
How an entry is triggered
A distributor's payables have grown faster than its receivables for four consecutive months. The system matches this to the late-paying-buyers archetype. It proposes a revolving working-capital line and invoice finance against the top buyers. It also suggests an early-warning covenant on debtor days. The toolkit already sits inside agreed appetite and pricing thresholds. The RM can therefore discuss it with the client that week. Credit still reviews the file, but against a known template instead of a blank page.
What the library does for less experienced RMs
Every bank has one or two RMs who spot a seasonal cash-flow problem months before it shows up in arrears. Today that instinct lives in their heads and leaves when they do. Written into a library entry, it becomes a reference every colleague can use. Judgment still decides the case. The library makes sure the right questions are asked first.
Data Governance Behind the SME Lending Operating Model
The library and the copilot are only as reliable as the data that feeds them. Data governance therefore belongs inside the SME lending operating model and has to be designed alongside it. The Basel Committee's BCBS 239 principles make a similar point about risk data at large banks.
Who owns each data signal?
Each library entry depends on signals such as debtor days, facility utilization and repayment patterns. Someone must own the definition of each one. Suppose finance calculates debtor days on year-end receivables while credit uses monthly averages. The trigger will then fire at different times for the same client. Sound data governance assigns one owner per signal, documents the formula and fixes the source system. The pod can then act on an alert without rechecking it by hand.
Data quality checks before an alert
Missing or late data causes many false signals. A borrower that submits management accounts six weeks late looks healthy until the numbers arrive. Data governance rules should therefore treat stale inputs as an alert in their own right. A typical rule: no management accounts within 45 days of quarter end opens a monitoring task for the credit analyst. A data gap then becomes a portfolio monitoring action instead of a blind spot.
Audit trail and model feedback
Every recommendation, acceptance and override is logged together with the data behind it. Internal audit can trace why a covenant was set at a given level. The same log feeds model review. When overrides cluster in one segment, the risk team checks whether that archetype's rules need recalibrating.
Put the Intelligence (Copilot) in the Workflow

The diagnosis runs on a an AI copilot built into the screens the RM already uses every day. Stand-alone analytics tools tend to go unopened after the first month.
The copilot does the following:
- It pulls together financial statements, account behaviour and CRM notes from across the bank's systems.
- Looks for anomalies in utilization or repayment before they turn into arrears. When a pattern matches a library archetype, it suggests a next step.
- Drafts the proposal outline from policy-aligned templates. If a recommendation would breach a risk appetite threshold, it raises an alert straight away.
The override is the interesting part
Six weeks before an agri-input importer's seasonal buying cycle, the copilot flags a pattern. Unhedged euro payables above twenty per cent of cost of goods, and the local currency down four per cent in a month.
It surfaces the FX exposure archetype, prepares a one-page brief, and proposes forward cover on the next two import orders, a documentary letter of credit, and a concentration covenant on the top three buyers. The appetite check passes.
The relationship manager reads it, agrees with the hedge, and knows something the data does not: the client is about to sign a fourth large buyer, which eases the concentration. She loosens the covenant and documents why.
Three months later, the new buyer is confirmed. That logged rationale becomes evidence the model learns from.
This is how a well-designed system handles disagreement. It records the reason, compares it with what actually happened and uses the result in the next calibration.
Consistent Service Across Relationship Managers

Every client should get the same quality of diagnosis, structuring and follow-up, whichever relationship manager they happen to have. That is a service design problem, and four artifacts solve most of it.
A standard discovery script that guides the diagnostic conversation. A proposal blueprint that makes every credit request complete and policy aligned. An onboarding checklist covering documentation, covenants and client expectations. And a 30/60/90-day milestone plan that sets early outcome targets and the follow-up that goes with them.
What does standardization free up?
Think about how much of an RM's week goes into reformatting spreadsheets and chasing missing documents. Standard formats and automation take that work off the desk. The RM and the analyst then spend their time on client conversations and structuring decisions that no template can make.
Want to build this in your team?
The Q-Lana Financial Skills Campus is launching Customer-Centric SME and Corporate Finance, a class for relationship managers and credit teams. It covers the diagnosis-first conversation, the structuring decisions that follow from it, and how to hold both inside the risk appetite framework. Taught on your own files rather than on cases, by practitioners who carry an SME book.
Starting the SME Lending Operating Model Without a Transformation Program
Many institutions see customer centricity as a large transformation programme, and that perception delays the start. In practice most of the pieces already exist. RMs know their clients but lack the tools to act quickly on what they know. Credit teams understand risk but wait too long for the data to assess it. Product teams have suitable facilities but rarely see the client's actual problem.
The sequence is short. Form pods around priority segments. Codify recurring client problems into a library, with clear data governance over the signals it uses. Then build the copilot into the RM's workflow. Customer experience and risk quality improve together, and the org chart stays as it is.
Next in This Series
The next article looks at what goes into the library. It covers the five problems behind most of an SME book and the pre-approved toolkits that answer them.
See it in practice
Our argument is that intelligence belongs inside the workflow. The quickest way to judge that is to look at a working RM workspace.
See the RM Workspace
The full operating model, including the pod design, the library structure and the copilot boundary, is set out in the Q-Lana white paper Customer Centricity in SME and Corporate Lending.
All Articles in This Series
- Customer Centricity in SME Lending Is a Risk Strategy
- SME Lending Operating Model Built on Solution Pods (this article)
- Solution Toolkits Settle the Risk Before the Client Asks
- Client Discovery in Ten Minutes Changes the Credit File
- Let AI Handle Credit Policy Enforcement and Free Your RMs
- RM Incentives Decide Whether Customer Centricity Sticks
Follow Q-Lana
One short email a week on credit risk, lending and impact finance, free. Or browse more guides in the Q-Lana Knowledge Center.
Get Q-Lana Weekly
One short email a week on credit risk, lending and impact finance, free. Or browse more guides in the Q-Lana Knowledge Center.