How might we enable law firms to review large volumes of documents more efficiently by identifying key clauses, red flags, and deviations at scale, while maintaining legal accuracy, explainability, and effective senior oversight?   

Background: 
Law firms regularly review large sets of similar documents such as leases, warranties, NDAs, litigation bundles, and due diligence materials. This work is time-intensive, highly repetitive, and often delegated to junior lawyers under tight supervision. We want to assess where automation genuinely adds value, where risks remain, and how reliably outputs can be supervised and trusted. 

We are interested in: 
Solutions that use AI to support first-pass review, clause extraction, and red-flag identification, producing structured outputs that can be efficiently checked and validated by lawyers. 

Why this matters to firms: 
Because document review consumes significant junior and senior capacity, and firms need evidence-based clarity on where automation reduces cost and risk rather than introducing new ones. 

How might we support faster and more reliable legal research by using AI to structure sources, surface relevant authorities, and reduce time spent on repetitive searching, without introducing hallucination or trust risks? 

Background: 
Legal research remains a core but resource-intensive activity. Lawyers often spend significant time searching, filtering, and cross-checking sources. Concerns around accuracy, traceability, and hallucinations continue to limit adoption. 

We are interested in: 
AI-enabled research approaches that prioritise source transparency, reliability, and relevance, while keeping lawyers in control of judgment and conclusions. 

Why this matters to firms: 
Because unreliable research creates professional risk, and firms need trusted ways to increase speed without compromising quality. 

LOIL 2.0 Benchmark Output: 
Source traceability scores, hallucination incidence rates, and research-time savings by task type. 

How might we enable senior lawyers to rely on AI-generated summaries of long documents and bundles for risk assessment and decision-making, rather than generic or superficial abstracts?

Background:

Senior lawyers need fast, accurate overviews of complex documents. Existing summarisation tools often fail to surface legally relevant risks or omit critical issues.

We are interested in:

Solutions that generate structured, decision-ready summaries with clear links to source material and transparent confidence signals.

Why this matters to firms:

Because senior time is scarce and expensive, and poor summaries can lead to missed risks or unnecessary rework.

LOIL 2.0 Benchmark Output:

Decision-usefulness ratings, omission rates for key risks, and senior validation time per summary.

How might we assist junior lawyers in drafting structured legal briefs and internal memos that improve quality, consistency, and learning outcomes, while reducing senior rewrite time and supervision burden?

Background:

Juniors often struggle with structure and relevance, leading to heavy senior rewrites. Firms are increasingly concerned about sustainable supervision and training quality.

We are interested in:

AI tools that support structured drafting and produce outputs that are easier to review, validate, and teach from.

Why this matters to firms:

Because drafting inefficiencies directly affect leverage, training outcomes, and supervision capacity.

LOIL 2.0 Benchmark Output:

Senior rewrite reduction rates, structural quality scores, and junior learning impact indicators.

How might we enable contract drafting that adheres to client-specific playbooks and constraints, while clearly flagging deviations and preserving lawyer control and accountability?

Background:
Institutional clients require strict adherence to agreed standards. Automation is attractive, but only where deviations, provenance, and accountability are fully transparent. 

We are interested in:
AI-supported drafting that operates within defined constraints and makes all deviations explicit and reviewable. 

Why this matters to firms:
Because client trust, regulatory compliance, and professional liability depend on knowing exactly where and why documents depart from playbooks. 

LOIL 2.0 Benchmark Output: 
Playbook compliance rates, deviation transparency scores, and senior validation effort per draft. 

How might we help lawyers draft routine legal correspondence more efficiently, ensuring appropriate tone (i.e. firm style), accuracy, and risk awareness, while keeping lawyers firmly in control of final communications?

Background:

Routine correspondence consumes significant time despite being relatively low risk when structured correctly. Tone or clarity failures can still create reputational issues.

We are interested in:

AI solutions that assist with first-draft correspondence, adapt tone to context, and surface risk areas clearly.

Why this matters to firms:

Because communication quality directly affects client trust, while inefficient drafting quietly erodes margins.

LOIL 2.0 Benchmark Output:

Time-to-first-draft metrics, tone appropriateness ratings, and lawyer edit distance scores.

How might we increase pro bono and legal clinic capacity by using AI to support form completion, triage, attendance notes, and draft advice, without increasing risk and while improving accessibility and supervision?

Background:

Pro bono clinics face growing demand but limited capacity. Supervising lawyers spend disproportionate time reviewing junior or student work.

We are interested in:

AI-enabled workflows that allow juniors or students to contribute effectively, while supervisors retain control and risk is reduced rather than increased.

Why this matters to firms:

Because pro bono delivery, regulatory expectations, ESG commitments, and talent development increasingly intersect.

LOIL 2.0 Benchmark Output:

Capacity uplift metrics, supervision time savings, accessibility scores, and error rates.