Legal AI in 2026: The 10 Most Searched Solutions Compared

TL;DR

The most searched legal AI solutions in 2026 — Harvey, Legora, Claude for Legal, Westlaw Edge — are powerful tools that deserve their success. None, however, was built with data sovereignty as a structural requirement.

For a Swiss law firm subject to art. 321 SCC, that is precisely where the decision lies. In February 2026, the Heppner ruling reminded the profession that using a third-party AI tool without confidentiality guarantees can cause professional secrecy to fall. This article presents the landscape by functional category, then sets out the questions to put to every vendor before adoption.

---

Introduction

Legal AI has moved, in less than three years, from demonstration object to daily working tool. The Legaltech Hub ranking of most searched solutions confirms it: at the top, one now finds AI legal assistants, contract analysis platforms, AI-augmented case law databases. The market is mature, adoption is real.

Behind the enthusiasm lies a question most rankings carefully avoid: where does the data go when AI processes a file? In February 2026, United States v. Heppner placed that question at the centre of the profession. A US federal court ruled that documents generated via a consumer AI tool were not covered by professional privilege, on the ground that transmitting them to a third-party service amounted to disclosure. The effect on law firms was immediate.

This article does two things: it presents the ten most searched legal AI solutions honestly by functional category, then asks the question the ranking leaves aside — that of data sovereignty, and what it means for a Swiss law firm.

---

Ten solutions, three functional categories

Before comparing, one clarification is necessary: these tools belong to different categories. Grouping them under the "legal AI" label is convenient for search rankings, misleading for the buying decision. A legal research assistant and a case law database do not solve the same problem.

AI legal assistants (copilots)

Claude for Legal (Anthropic) leads searches. A state-of-the-art general AI assistant adapted for legal use, capable of analysing documents, reasoning on complex questions and drafting. Its strength is reasoning quality. For use in regulated sectors, hosting and data processing conditions warrant examination.

Harvey is the most established legal AI assistant among large firms. Adopted by a significant share of AmLaw 100 firms, it offers an assistant, a document vault and workflows. Harvey holds solid certifications (SOC 2 Type II, ISO 27001) and regional hosting. An important point: Harvey relies on proprietary LLMs (OpenAI, Anthropic, Google). Data associated with AI queries can therefore be transmitted to third parties.

Legora (formerly Leya), based in Sweden, is a collaborative legal AI platform specialising in large-scale contract review and structured data extraction. Where Harvey targets collaborative analysis, Legora targets the standardised and auditable processing of large document volumes.

Wexler positions itself on factual analysis and chronology construction from large document sets, particularly useful in litigation.

Orbital Copilot specialises in real estate law, with an AI that automates title review and real estate due diligence.

Case law and research platforms

Westlaw Edge (Thomson Reuters) is an AI-augmented legal database: case law research, citations, precedent analysis. The value lies in proprietary content as much as in the AI. This is a category of its own.

Management and automation software

3E by Elite is a practice management system (accounting, billing, financial management). It appears in "legaltech" searches but belongs to the legal ERP space, not document analysis AI.

Stencil by Page Vault focuses on document automation and web evidence capture.

Consilio is a leading player in e-discovery and large-scale document review management, primarily in litigation and investigations.

Fern AI completes the picture in the legal task automation segment.

---

Comparison table: function, origin, sovereignty

A cross-sectional reading focused on the decisive criteria for a Swiss organisation under confidentiality obligations. The goal is not to name a winner — each tool excels in its category — but to illuminate the shared blind spot.

Reading the table: almost all the most searched solutions are American or British in origin and rely on cloud infrastructure outside Switzerland, often backed by third-party LLMs. None was conceived with data sovereignty as a structural requirement. They were designed for the Anglo-Saxon market, where the legal and cultural framework differs from ours. This creates a blind spot for Swiss law firms, fiduciaries and banks.

---

The Heppner ruling and its implications for Swiss firms

A precedent that reframes the professional secrecy question

In February 2026, in United States v. Heppner, Federal Judge Jed S. Rakoff ruled that documents produced by a defendant via a consumer AI tool were protected by neither professional privilege nor the work product doctrine. The reason: transmitting information to a third-party AI service whose terms of use do not guarantee confidentiality amounts to disclosure to a third party. The privilege falls.

This decision concerns consumer tools. It has illuminated a principle that extends across the profession: the choice of tool matters as much as how it is used. A tool that transmits identifiable data to a third-party LLM hosted outside your jurisdiction creates a risk that neither a contractual clause nor good intentions can fully neutralise.

Why the issue is more acute in Switzerland

For a Swiss law firm, three frameworks overlap. Art. 321 SCC makes breach of professional secrecy a criminal offence. The nLPD strictly regulates the transfer of data outside Switzerland. For financial actors, FINMA requires precise documentation of where data lives and who accesses it.

Most solutions in the ranking, designed for the American market, process data via infrastructures subject to the Cloud Act — the law that can compel an American provider to hand over data to a federal authority, even if physically stored in Europe. A "Swiss region" or "European region" of an American provider does not resolve this problem: the vendor's domicile takes precedence over server location.

---

Ketl: a different category

For clarity: Ketl does not compete directly with Harvey or Legora. Harvey and Legora are copilots for legal reasoning and drafting. Ketl is a sovereign, AI-augmented DMS (document management system). Ketl will gradually introduce similar functionalities to those two providers through integrations and open-source solutions. For now, using specialised software alongside Ketl remains sensible depending on the use case.

The structural difference is this: at Ketl, sovereignty is a foundation, not an option.

The entire chain (storage, exploitation, AI models) is hosted and operated in Switzerland, by Swiss companies. No document leaves Swiss territory, including during AI processing. Where Harvey relies on American AI, Ketl uses open-source models as well as its own models, developed and trained in Switzerland, approximately 1,000 times lighter than a large generalist model — efficient enough to run on sovereign infrastructure without calling an external service.

For a firm subject to art. 321 SCC, this difference is not cosmetic: it separates a tool usable without reservation from one that creates a professional liability risk.

In production: Ketl processes over 46 million documents for more than 600 active users in the legal sector. Ketl covers 11 additional regulated activity areas (insurance, banking, fiduciary...). AI models are developed, trained and hosted exclusively on Swiss infrastructure (Exoscale and Infomaniak).

The right way to see it: the copilots in the ranking help reason and draft faster. Ketl ensures documents remain sovereign while being intelligently classified, extracted and exploited. Both logics can coexist, provided the sovereignty question is asked from the outset.

---

Questions to ask every legal AI vendor

When evaluating a legal AI tool for your firm, fiduciary or bank, here are the questions to put in writing to every vendor, regardless of their ranking position.

On AI processing. Does the tool rely on a third-party LLM? If so, which one, and where does it process data? Are identifiable data transmitted, or anonymised before processing?

On hosting. Where is data physically stored? Is the vendor subject to Swiss law, or to a foreign jurisdiction exposed to the Cloud Act?

On retention. Are your data used to train models, even anonymised? What is the zero data retention (ZDR) policy with the LLM provider?

On professional secrecy. In the light of cases like Heppner, does the tool's use preserve professional secrecy, or expose it to a qualification of third-party disclosure?

These questions do not disqualify any solution for uses where sovereignty is secondary. In regulated sectors, they shift the centre of gravity of the decision: from "which tool is most capable?" to "which tool can I use without compromising my obligations?".

---

FAQ

Harvey is ISO 27001 certified. Is that sufficient for a Swiss firm under art. 321 SCC?

ISO 27001 certification attests to serious security practices. It does not resolve the jurisdiction question. Harvey is an American company relying on third-party LLMs (OpenAI, Anthropic, Google), themselves subject to the US Cloud Act. For a firm under art. 321 SCC, the question remains open: does transmitting identifiable data to an LLM operated by an American entity constitute disclosure to a third party under Swiss professional secrecy? The Heppner ruling provides elements of an answer.

It should also be noted that the tool itself may not be suited to Switzerland. Managing the country's particularities — notably its case law written in multiple languages — is a significant obstacle to optimal performance.

Can Harvey or Legora be used for some tasks and Ketl for others?

Yes. A common configuration uses a drafting copilot for redacted or anonymised documents, and a sovereign DMS for storage, classification and archiving of originals. Both logics are complementary. The key is to define from the outset which documents may transit through a third-party service and which require sovereign processing.

Does the Heppner ruling apply in Switzerland?

The Heppner ruling is a US federal decision. It does not apply directly under Swiss law. Its lesson is nonetheless transferable: once a document covered by professional secrecy is transmitted to a third-party service with insufficient confidentiality conditions, the risk of qualification as third-party disclosure exists, including under art. 321 SCC. Several cantonal bar associations have published guidance on this topic since early 2026.

Does Ketl offer drafting or legal research functionality comparable to Harvey?

Ketl is an AI-augmented DMS designed for classification, extraction, document search and workflows. LLM-assisted drafting and case law research fall within Harvey, Legora or Westlaw Edge. The two categories address distinct needs and can coexist in the same firm.

How long does Ketl deployment take for a law firm?

Standard deployment is four weeks, including training. Configuration covers connection to Microsoft 365 or a sovereign storage solution (kDrive, Nextcloud), historical document import, naming convention setup and initial workflows. Current operations remain active throughout the deployment period.

---

Conclusion

The ranking of most searched legal AI solutions reflects a maturing market, dominated by powerful AI copilots — Claude for Legal, Harvey, Legora — and specialist high-performing tools. These solutions deserve their success on their terrain: reasoning, drafting, analysis, research.

None was built with data sovereignty as a structural requirement. For a Swiss law firm, fiduciary or bank, that is precisely where the decision lies. The Heppner ruling made it clear: the choice of tool is bound up with the protection of professional secrecy. Ketl does not position itself against these copilots, but on the dimension they leave vacant — a document AI whose data never leaves Switzerland.

Sovereignty is a regulatory requirement, and Ketl has integrated it as an architectural foundation.

Want to see it on your own files? We have been supporting Swiss law firms and regulated professions in their digital transformation since 2019. Request a demo or contact@ketl.ch

---

Written following a conversation with James McGill, co-founder of Ketl, and structured with Ketl AI, our sovereign AI hosted in Switzerland.

The ranking of most searched solutions is established by Legaltech Hub. Ketl is not affiliated with the vendors cited, presented here for informational and comparative purposes.

---

Sources:

• United States v. Heppner, S.D.N.Y., February 2026 (Judge Jed S. Rakoff)
• Legaltech Hub — legal AI solutions ranking 2026
• nLPD — Federal Data Protection and Information Commissioner (FDPIC)
• Cloud Act — official text (Congress.gov)

Related pages

• AI in regulated industries: what changes for Swiss law firms and banks
• Swiss document management system: why hosting alone is no longer enough
• SharePoint, kDrive or Nextcloud: which solution for storing your data in Switzerland