A lot of settlement strategy depends on how well you understand your own position, and how well you can assess the other side’s willingness to settle. It is not just about the number. It is about the pressure points sitting underneath it: merits, evidence, costs, timing, leverage, uncertainty, risk appetite, and what each side thinks happens if settlement does not happen. That is what my settlement pressure tester skill is designed to help with. It stress-tests a proposed settlement position by looking at things like: – how strong the case actually is – what assumptions the position depends on – where the evidence is uncertain – what leverage each side may have – how the other side is likely to respond – how timing and costs affect the strategy It does not tell you whether to settle - that remains a legal and commercial judgment. But it helps make the underlying pressures more visible before an offer is made.

I recently built a disclosure strategy mapper. The idea is simple: you give it a case summary, pleading, chronology or issue list, and it helps you think through the documents that may matter. It maps things like: – likely document categories – potential custodians and sources – adverse material that may exist – evidential gaps – search themes – privilege or sensitivity flags – disclosure risks I wanted something that could help with the early-stage thinking that often sits behind litigation strategy. What documents would we need to prove this? Where are they likely to be? Who might hold them? What might undermine the case? What are we assuming exists, but have not yet found? It is not a substitute for applying the relevant disclosure rules, of course. But it is a tool to help you structure your thinking.

Not every use of AI in legal work needs to be documented in the same way. But there are instances where keeping a clear record really matters. For example, where AI is being used to help make decisions that affect people’s lives, rights, access to services, employment, finances, immigration status, or legal position. In those contexts, it is not enough to say: “we used AI as part of the process.” You need to be able to explain what actually happened. What tool was used. What materials were uploaded or pasted in. What the tool was asked to do. What output it produced. How that output was checked. What was relied on, and what was not. That is what I had in mind when building the AI Audit Trail skill, available on Lawve AI. It helps create a structured record of AI-assisted legal work, including prompts, inputs, outputs, verification steps, reliance, and risk flags. The point is not to create paperwork for its own sake. It is to make sure that, where AI use impacts decisions that affect people, there is a clear record of how the tool was used and how the output was treated.

One of the key skills in legal work is making complicated things understandable. Legal advice can be technically correct and still be very hard for a client to use. Too much legalese and not enough explanation of what it actually means for the person reading it. Most clients are not looking for a lecture on the law. They want to understand: – What is happening; – Why it matters; – What the risks are; – What choices they have; and – What they should do next. That is what I had in mind when building my latest skill, available on Lawve AI. It helps turn legal analysis, case updates, research notes, or internal drafting into language a client can understand and act on. The aim is to preserve the nuance, but make the practical implications clearer.

It is great to see Serhan Handani put another of my Lawve AI skills to test in the UK Legal AI Brief, this time Judicial First Impression. Open work only gets better when people test it openly, and Serhan does. Serhan highlights the value of running Judicial First Impression alongside Opposing Counsel Review on the same document. He notes that the two skills catch overlapping but distinct issues: the adversarial skill goes further on strategy, the judicial one on drafting defects. That is exactly the pairing I had in mind. Different reading positions highlight different problems, and writers benefit from both.

We all talk about the risk of AI hallucinating cases that don’t exist. There is a quieter risk that gets less attention: the inferences AI draws that do not immediately raise red flags. A date that was never in the document. A rule reconstructed from training data rather than the statute. A paragraph reference that looks right. Plausible but wrong, and by the time you notice, it is in your draft. I’ve created a skill to address this, now available on Lawve AI. It forces a different mode. Claude answers only from materials you provide or online sources it has actually accessed. Every claim is anchored to a pinpoint reference and labelled in one of five categories: expressly stated in materials, verified online, supported but not expressly stated, not found, or flagged inference. Gaps stay as gaps. “Not found” is treated as a correct answer, not a failure. Built for document review, chronologies, citation checks, evidence summaries, and any drafting task where fidelity to the record matters more than a complete-sounding answer.

Tired of Claude outputs sounding like… Claude? Fluent. Structured. Technically correct. Still not persuasive. Persuasive legal writing isn’t just clarity. It’s how the argument is framed, how it lands on a reader who may be sceptical before they’ve finished the first paragraph, and what the writer chose to leave out. That’s the layer AI gets wrong by default. It produces prose that reads well - and then sequences the weakest argument first, buries the concrete example, hedges the strong claim into a soft one, and reaches for the same three rhetorical patterns every time. The result feels organised. It doesn’t feel written. So I put together a skill focused on that layer, available on Lawve AI. It draws on techniques from Justice Kagan, Boies & Olson, and other top advocates - opening with a frame instead of a summary, sequencing arguments by persuasive force, using concrete analogies instead of abstract principles, ending on a parting thought rather than “for the foregoing reasons.” It also includes a long section on the structural and phrase-level tells that mark prose as AI-generated, with examples of what to cut and what to write instead. Generating text is the easy part. Shaping it into something a sceptical reader will actually be moved by is the real work.

Serhan Handani built a deliberately flawed skeleton argument and ran my Opposing Counsel Review skill against it. Then wrote about what happened in UK Legal AI Brief: https://uklegalaibrief.co.uk/p/correction-stress-testing-larissa-meredith-flister-s-opposing-counsel-review-skill He drafted (with Opus 4.7) a Part 24 skeleton for a fictional commercial dispute, planted the kinds of defects practitioners actually see - inflammatory tone, clauses cited but never set out, no engagement with UCTA s.3 or Misrepresentation Act s.3, performance asserted on the word of the Claimant’s own CTO - and then handed the lot to the skill to see what it would do with it. A few things he flagged that I was particularly pleased to read: - The skill identified the structural weakness immediately: the Claimant had collapsed two arguments (contract is sound; performance is a fact) into one, and the second was contested and unsuited to summary disposal. - It picked up the statutory reasonableness controls the skeleton had quietly omitted, and cited the right binding authorities for them (First Tower Trustees, AXA Sun Life, Geldof v Simon Carves, Three Rivers). - It caught a planted internal inconsistency at paragraph 5 that Serhan was not sure it would catch. - The “if I were the judge” section produced a single quoted judicial question that, in his words, would “almost certainly sink the application in court”. The Opposing Counsel Review skill is free, open-source (Apache 2.0), and available on Lawve AI.