Where there's a will, AI still has work to do

UK law firm SE Solicitors tested an AI chatbot drafting a will for a fictional 42-year-old with a £440,000 house and £10,000 savings, and found the output looked plausible but omitted key probate questions (e.g., inheritance tax planning, partner status, vulnerability of heirs) and even left placeholders, such as an unspecified dog-care sum. The firm warned AI use could cause consumers to miss changing inheritance tax rules and to expose sensitive personal/financial data without solicitors’ confidentiality safeguards. Broader demand signals cited include a 312% rise in “legal AI” searches and 72% of UK adults (age 30–34) considering AI for will updates, but the test suggests AI is not ready to replace probate solicitors.

Analysis

This is less a model-quality critique than a monetization critique: in regulated, high-liability workflows, the economics accrue to whoever owns the review layer, audit trail, and indemnity, not the entity that can produce a passable first draft. That favors incumbents like Thomson Reuters and RELX over generic chatbot interfaces, because buyers will pay for workflow certainty rather than raw generation. For GOOGL, the near-term hit is mostly reputational in sensitive consumer use cases, not revenue-level impairment.

The time horizon matters. Over days, the market may read this as another reminder that consumer AI is brittle in high-stakes tasks, but over 1-3 months the more important variable is whether enterprise AI sales teams can prove lower error rates and stronger data controls in legal/financial verticals. The real tail risk for GOOGL is privacy/regulatory backlash if users paste confidential estate or tax data into consumer tools, which could increase procurement friction and push more workload into locked-down enterprise environments. A reversal would require evidence that AI copilots are being adopted with human review and confidential-mode guarantees, reducing the perceived liability gap.

The contrarian view is that the article likely overstates substitution risk: wills are episodic, so the realistic TAM for fully self-serve drafting is small, and most consumers will still want a human sign-off for anything consequential. If anything, this supports a hybrid model where AI generates the first pass and a professional validates it, which is constructive for legal software and services rather than a direct threat. On that framing, the signal for GOOGL is neutral-to-slightly negative at most, unless this kind of headline starts showing up in enterprise renewal conversations.

Key Decisions for Investors

• Do not short GOOGL on this headline alone; treat it as a sentiment check, not a fundamentals event. Reassess only if the next earnings call shows slower Gemini/Workspace adoption in regulated industries or any privacy-policy tightening.
• If GOOGL trades down 2-3% on the headline, consider a small tactical long for a 2-6 week mean-reversion trade, with a tight stop if follow-on commentary ties consumer AI usage to privacy incidents or enterprise churn.
• Build a watchlist long in legal-workflow incumbents such as TRI or RELX on any AI skepticism selloff. These names should capture the spend that migrates from drafting tools to review, research, and compliance workflows over 3-12 months.
• Avoid allocating to pure consumer-facing AI drafting apps until there is evidence of liability coverage, confidential-mode controls, or court-admissible audit trails. The market is likely underpricing adoption friction in regulated tasks.
• Set an alert for any UK/EU regulatory or bar-association guidance on AI-assisted legal drafting. That would be the real catalyst; absent policy, this remains a micro-signal with limited direct earnings impact.