From Podcast to Practice: reflections on AI and automation in Insolvency

A few years ago, we discussed a thought-provoking question on the Turnkey Exchange podcast: is the insolvency industry ready for AI and automation?

At the time, the conversation was grounded and pragmatic. Rather than focusing on hype, the discussion explored where technology could genuinely help automating repetitive work, improving compliance, supporting investigations and reporting and where caution was essential. The conclusion was clear. Insolvency would remain a profession built on responsibility and human judgement, with technology there to support, not replace.

Several years on, it feels like a useful point to return to those conclusions and test how they stand up in practice.

What happened

One of the key expectations from the podcast was that automation would lead the way. That has largely proven to be the case in practice. Across the industry, routine processes such as onboarding, compliance checks, reminders and data handling have become faster and more consistent. Not through dramatic change, but through steady, embedded improvements. At the same time, AI is increasingly woven into everyday workflow tools:  from search and support to document generation, anomaly detection and communications. As a result, AI is less about adopting standalone solutions and more about enhancing existing ways of working, meaning firms often no longer experience it as a separate capability at all. This hasn’t made insolvency ‘smarter’ in a human sense, but it has made it more reliable in day‑to‑day practice.

In many ways, it has delivered exactly what was anticipated the most immediate gains in cost, compliance and consistency. It also reinforces a point discussed at the time: Automation and AI are increasingly being viewed as necessary, rather than optional, in supporting a modern, regulated profession.

Deborah Baxter, CEO of Turnkey commented:

What has become clear is that AI is no longer a question of ambition, but of governance. Automation has raised the baseline for consistency and compliance, while generative tools are proving valuable when carefully supervised. The task for firms now is to use these technologies in a way that preserves judgement, accountability and trust at the heart of the insolvency practice.

Generative AI and its role

The podcast also raised questions around more advanced AI and how far it could go. In practice so far, generative AI has found a role, but a defined one. It is being used to support drafting, summarising and explaining complex information. But the risks highlighted in the original discussion have become equally clear. Outputs can be inaccurate, misapplied or overly confident, requiring careful oversight from the insolvency expert. As a result, its use has tended to remain largely assistive and supervised.

A core theme from the original conversation was technology should support professional judgement, not replace it. Professional judgement, however, may be changing. Instead of engaging with raw facts, practitioners may be reviewing an AI shaped version of those facts. Whether it is through copilot generated content or summarised AI notes. This raises important questions about where firms may be most exposed today: the quality of underlying data, the risk of over‑standardised outputs, and the potential for skills to erode as professionals become further removed from the detail of their work.

The gains, the pains, and the jobs to be done

Looking back, many of the benefits anticipated in the original podcast discussion have materialised. Greater consistency in process, stronger compliance, reduced pressure on cost‑sensitive work, and clearer communication with creditors and stakeholders have all become more achievable through automation and carefully applied technology. Together, these improvements have helped raise the baseline standard of insolvency work, broadly as expected.

At the same time, the concerns raised then remain very real. There is an ongoing risk of skill erosion where professionals become too removed from the detail of their work, alongside continued challenges around over‑reliance on automated outputs and data security. More subtly, there is the risk of standardisation. AI tools naturally favour averages, the safest answers and most common wording which can be a limitation in a profession where judgement, nuance and context matter.

The job to be done, therefore, has not fundamentally changed. Technology must continue to remove friction from routine, process‑heavy tasks and strengthen consistency and compliance, while supporting clearer communication. Crucially, it must do so in a way that reinforces rather than replaces professional judgement and responsibility, which remain central to effective insolvency practice.

How insolvency is evolving

One of the most interesting themes discussed in the original podcast and one that has become clearer with hindsight is how the structure of insolvency work appears to be changing.

What the discussion anticipated, and what experience since suggests, is a growing distinction between different types of work:

• Highly automated, process‑driven tasks become more efficient, standardised and involve lower levels of discretion
• More complex, judgement‑led work is where investigation, strategy and professional expertise remain critical

Technology has played a role in supporting both areas. In practice, it has helped streamline routine, process‑heavy activity, while reinforcing the importance of professional judgement in more complex cases.

Do we need external regulation and internal governance?

The podcast also explored whether regulation would eventually follow, and that now feels increasingly likely. While formal legislation may take time to catch up, expectations around disclosure, data use, accountability and supervision are already shifting. Importantly, the challenge is not limited to external regulation. Many firms are adopting AI at pace internally, sometimes faster than governance frameworks can keep up. This makes it critical for firms to think seriously about how they validate AI outputs, manage data quality, maintain professional skills, and preserve clear lines of accountability.

Final reflection

• Looking back, the original conversation holds up well.
• Automation has delivered where expected, becoming embedded and dependable.  
• Generative AI has moved rapidly forward, valuable but still requiring careful oversight.  
• Insolvency remains grounded in responsibility, trust and professional judgement.

If anything, the last few years have reinforced the original views discussed on the podcast. The insolvency industry is well placed to adopt AI and automation, provided this is done in a way that maintains control, trust and transparency.  

What has become clear is the range of approaches firms have taken, with some moving quickly to automate while others remain more cautious. Taken together, this reinforces that the real challenge for firms is not whether to use AI, but how to govern it responsibly while preserving professional judgement and clear accountability.

As the discussion has moved beyond generative tools alone, advances in document intelligence, grounded search and retrieval, multimodal analysis and agent‑led workflows are starting to influence how AI is applied in practice. The developments are beginning to shift the practitioner’s role away from direct analysis towards validating AI‑generated outputs. As seen across many professions, this points to a broader professional shift, with implications for training, supervision, career paths and how expertise increasingly develops.

You can listen to the original podcast here: Is the insolvency industry ready for AI, ChatGPT and Automation? Hear from our host, Mark Simpson, joined by special guests Darren White, Greg Clough, and Stephen Hunt, where they discuss the impact of new technology, such as AI, ChatGPT and Automation, and whether the insolvency industry is ready for this level of innovation.

‍