Your insolvency CRM knows what happened, not what happens next

Traditional insolvency CRMs are very good at recording what has happened. They now need to evolve to help teams understand what needs to happen next while keeping responsibility for regulated judgement, accountability and customer care firmly in human hands.

For years, case-management technology has been designed primarily as a system of record. It stores customer details, case stages, financial information, creditor claims, documents and notes. That foundation is essential, but it still leaves case managers spending significant time rekeying information, reviewing correspondence, checking incomplete records, preparing summaries and moving work between queues. Rules-based automation has already removed some of that effort and it works well when input is structured and the next action predictable. Insolvency casework, however, often involves incomplete evidence, changing customer circumstances, varying creditor responses and information arriving through different channels. This is where carefully governed AI can assist: not by replacing the case manager or insolvency practitioner, but by reducing administrative work and presenting the right evidence at the right point in the workflow.

Why AI needs strong foundations

Useful automation begins with a reliable operational foundation. Insolvatrack brings case workflows, customer information, standard-financial-statement (SFS) data, debts, claims, review history, tasks and audit records into a controlled workspace. Solid foundations matter because AI can't compensate for unclear ownership, inconsistent data or undefined processes; it will only amplify them. The opportunity is not to place a chatbot over an existing CRM, it is to combine structured data, explicit workflow rules and AI assistance so that each is used for the work it handles best. Workflow rules control required steps and deadlines. AI helps interpret unstructured information and identify patterns. People remain responsible for decisions, advice and exceptions.

The Financial Conduct Authority (FCA) says it will continue working with the Insolvency Service and recognised professional bodies to raise standards and tackle consumer harm in the debt advice market, while using Consumer Duty to drive improved outcomes. A well-run records system is necessary for that scrutiny. It is not, on its own, the differentiator.

How AI can assist

FCA-commissioned research published as part of the July 2026 Mills Review found that demand for AI assistance among retail financial services customers appears strongest in complex areas including debt advice, pensions and investments. Separately, the same research found that 20% of this demographic would be open to using AI capable of acting autonomously for certain tasks. Neither finding tells us what a customer expects from a specific insolvency case system but they provide an incentive to identify the exact scope of AI rather than striving to move faster. The failure I see most often in casework isn't an inaccurate AI-generated suggestion. It's a case that has drifted since its last SFS review, a changed circumstance, a missed follow-up, a channel switch with nobody specifically tasked to notice. That's the gap this kind of assistance needs to close first.

1. Understanding incoming information

Customer emails, call transcripts, creditor correspondence and supporting documents can involve substantial manual review. AI can help classify incoming material, produce a concise summary and identify candidate information for the case record. Any extracted value should retain a link to its source, carry an appropriate confidence indicator and require human review before it changes the official record.

2. Keeping cases moving

A case may need attention because evidence is missing, a deadline is approaching, a creditor claim has changed, an SFS review is incomplete or the customer's circumstances have altered. AI can help spot those signals and prioritise work. Conventional automation can then complete low-risk actions such as creating a task, routing an item to the correct queue or issuing an approved notification. The distinction is important: AI identifies context; controlled workflow executes the permitted action.

3. Assisting decisions without making them

AI can compare the current SFS with an earlier review, highlight unusual movements, identify potential inconsistencies and prepare a case summary for an authorised reviewer. It can also suggest questions that may need to be answered before the case progresses. It should not approve an arrangement, determine suitability or replace the professional judgement required for regulated advice. The human reviewer must be able to inspect the evidence, accept or reject the help offered and record the rationale for the eventual decision. This is not just good practice, it is where the responsibility actually sits, and it doesn't move because AI is involved. For insolvency casework specifically, that responsibility cannot be delegated to AI, regardless of its capability.

4. Reducing the burden of case communication

Case teams repeatedly produce summaries, handover notes, update emails and standard letters. AI can create a first draft using approved language and the information held in the case file. The value is not autonomous communication; it is a faster starting point. Communications should still pass through the organisation's approval, quality and vulnerability controls before being issued.

5. Connecting activity to customer outcomes

Case activity alone does not show whether the customer received a good outcome. Where evidence held in Insolvatrack is made available alongside Better Outcomes, conversations and CRM events can be reviewed against the context of the wider customer journey. This can help quality assurance teams examine whether vulnerability was recognised, commitments were completed and the eventual outcome was consistent with the evidence available at the time.

The route to AI automation

The safest route to AI-enabled automation is progressive. Begin by observing what the AI identifies. Measure it against experienced reviewers. Introduce AI where staff can inspect and correct the output. Automate only low-risk, reversible actions once performance and controls are understood. Continue to monitor results after deployment rather than treating implementation as the end point.

The control principle

Every AI-assisted action should be evidence-linked, reviewable, reversible where appropriate and recorded in the audit history. It must be clear who is assigned responsibility at every stage. That requires clear governance around data protection, access, model and prompt changes, accuracy, exceptions and the circumstances in which operational data may be used for improvement. In a regulated environment, explainability is not a slogan. It means being able to show what information informed an output, what happened next and who was accountable.

Moving from case recording to controlled automation

The AI-assisted insolvency CRM is not a system that makes autonomous decisions about customers. It is a system that can interpret more of the administrative context, help teams find what matters and allow predictable work to move with less manual intervention. Insolvatrack offers the opportunity to build on a structured case-management foundation and introduce assistance where it produces a clear operational benefit: quicker triage, less rekeying, stronger case summaries, more consistent workflows and better visibility of work requiring human attention. The point isn't to remove professional judgement but to stop wasting experienced people's time preparing information that a controlled system can safely assemble.