Evaluate legal research software with a controlled pilot: define the firm’s coverage, workflow, security, and budget requirements, then give every finalist the same representative research questions and score the results against an agreed rubric. A polished demo or an AI-generated answer is not proof that a product retrieves the right law, supports its conclusions, or fits the firm’s obligations.
Start with the firm’s requirements
Before contacting vendors, write down what the software must do for the firm—not just the features that look useful in a demonstration. Include the jurisdictions and practice areas where the firm works, the kinds of users who will rely on the product, recurring research tasks, current subscriptions, existing technology, and the budget available. Those details define what “adequate coverage” means and make it possible to compare complete configurations rather than marketing claims.
- Map essential sources. List the primary law and secondary materials the firm needs for each jurisdiction and practice area. Identify any content gaps in the subscriptions already available to attorneys.
- Describe representative work. Include routine questions as well as difficult, recurring matters. Use realistic facts and procedural postures, and note whether a task calls for controlling authority, contrary authority, a statutory comparison, a survey across jurisdictions, or useful secondary-source context.
- Identify the users and workflow. Specify who will search, review, administer, and support the product, and how research is currently saved, cited, shared, and incorporated into work product.
- Set constraints. Record the acceptable contract term, seats, jurisdictions, implementation effort, security requirements, and total spend. Include existing subscriptions and any free government or legal resources the firm routinely uses.
The American Bar Association’s summary of its 2024 Legal Technology Survey Report, published March 3, 2025, says 67% of attorneys relied on fee-based online services for legal research and 55% used free platforms such as government websites and legal databases. Those survey figures describe reported use; they do not establish which mix is right for a particular firm.
Test every finalist on the same matters
Use a written pilot protocol rather than relying on separate vendor demonstrations. Each finalist should receive the same questions, starting facts, and evaluation criteria. Keep a record of the search or prompt used, the materials returned, the time spent, and what a lawyer had to do to reach a dependable result.
Recommended Free Tools
- Select the questions. Draw them from actual firm work, while removing client-identifying or confidential information unless the firm has approved the product and pilot for that use. Include both common tasks and matters that expose jurisdiction, procedural-posture, or contrary-authority challenges.
- Run equivalent tasks. Give each finalist the same question and comparable time. Use the product’s ordinary search and review workflow rather than a specially prepared demo environment.
- Check the underlying law. Open the cited decisions, statutes, regulations, or other authorities. Confirm that they are the cited materials, that they support the proposition, and that the firm’s usual citator workflow shows their current treatment.
- Record failures as well as successes. Note missing controlling authority, irrelevant results, incorrect or unsupported propositions, overlooked contrary authority, unusable citations, and extra review steps. Record elapsed time, but do not treat speed as a benefit if the result requires substantial correction.
- Include intended users. Have the attorneys and staff expected to use the product perform normal tasks. Record onboarding, training needs, export and citation handling, accessibility, administrative work, and support experiences.
- Apply the same scorecard. Agree on criteria and their relative importance before reviewing results. Weight them according to the firm’s work and risk profile; a universal ranking or weighting will not fit every firm.
| Scorecard area | Evidence to record during the pilot |
|---|---|
| Coverage and retrieval | Required jurisdictions and source types available; whether controlling authority and relevant facts or procedural posture surfaced; gaps in databases or jurisdiction coverage. |
| Research quality | Relevance of results, treatment of contrary authority, usefulness of secondary sources, and any unsupported statements or omissions. |
| Citations and status checking | Whether citations open to the identified primary material, support the proposition, and fit the firm’s process for checking current treatment. |
| Workflow and adoption | Time and steps for searching, reviewing, saving or exporting, handling citations, training users, and administering access. |
| Security and contract fit | Written answers on client-data handling and security, plus whether contract commitments meet firm requirements. |
| Total cost and support | Cost for the specified configuration, implementation and training needs, support terms, usage constraints, and renewal conditions. |
Verify citations and AI-assisted analysis
Apply source-level verification to conventional search results and AI-generated summaries alike. For each material proposition, follow the citation to the authority, read the relevant text in context, check its status using the firm’s normal citator workflow, and look for contrary authority. Record whether a citation is missing, does not resolve to the claimed source, fails to support the proposition, or leaves out a material qualification. A citation shown in an interface is a lead to check—not proof that the legal analysis is correct. Lawyers remain responsible for evaluating authority and analysis.
During a pilot, score unsuccessful examples as carefully as successful ones. A small number of persuasive answers cannot establish how a product will perform across the firm’s jurisdictions, practices, or matter types.
Review confidentiality, security, and professional duties before a client-data pilot
Do not put client material into a pilot until the firm has reviewed current written vendor information and involved its ethics and security stakeholders. Ask specifically about retention and deletion, whether prompts or uploaded materials may be used for model training, access controls, encryption, hosting location, subcontractors, auditability, and breach notification. Check the answers against client requirements, internal policy, confidentiality obligations, and applicable ethics guidance.
The ABA Legal Technology Resource Center links to a state-by-state collection of cloud ethics opinions. Consult the jurisdictions relevant to the firm rather than assuming one rule applies everywhere. The ABA summary of its 2024 survey, released March 3, 2025, reports that 60% of firms had implemented formal cybersecurity policies; that survey finding says nothing about the security of a particular research product. Vendor assurances should be assessed through current documentation and contract language, not treated as a substitute for the firm’s own review.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Compare products by verified fit, not feature lists
Two examples described in the available product materials illustrate why feature descriptions should be treated as claims to verify in a pilot, not as comparative performance findings.
- Westlaw Precision with CoCounsel. The Maryland State Bar Association describes AI-assisted research, summaries that reference primary sources, KeyCite, comparisons of statutes and regulations, and multi-jurisdictional surveys. It says pricing varies by attorney count, jurisdiction, and contract length, and describes trial availability as by request. Thomson Reuters describes CoCounsel Deep Research as grounded in Westlaw and Practical Law content, with hyperlinked sources and KeyCite integration. Confirm current functionality, coverage, security terms, pricing, and contract scope directly with the vendor.
- Lexis+ AI. The Maryland State Bar Association resource describes AI legal search, case summarization, analytics, Lexis+ ecosystem integration, conversational assistance, and document drafting and review. That is a feature description, not a comparative evaluation. Current price, coverage for a particular practice, and contract terms should be confirmed directly.
These examples are not a complete market survey, and the descriptions do not establish that one product is more accurate or valuable than another. Apply the same firm-specific questions and verification process to every finalist.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Compare the complete cost and build a local value estimate
Request a written quote for the exact bundle the firm would buy. Specify seats, jurisdictions, content add-ons, AI functions, usage assumptions, term, renewal terms, onboarding, training, and support. Headline prices or features cannot show whether two proposals cover the same work. The Maryland State Bar Association notes that Westlaw Precision with CoCounsel pricing varies by attorney count, jurisdiction, and contract length; a quote for one firm should not be generalized to another.
Estimate value using the firm’s own pilot observations and workload. Compare time on representative tasks with the current process, then account for research volume, implementation, training, administration, and contract costs. Treat vendor ROI guidance as a prompt to calculate local value, not as independent evidence of savings: the available sources provide no comparable, independently established savings figure. Preserve the assumptions so the estimate can be revisited when actual use is known.
Document the decision and revisit it before renewal
Keep the completed scorecard, pilot questions, observed results, security review, quote, assumptions, and decision rationale together. A written record makes it clearer why the firm selected, retained, or rejected a product and gives the renewal review a baseline: compare expected value with actual use, workflow outcomes, support, and total cost. If usage or the firm’s work has changed, reassess the requirements rather than renewing solely because the product is already in place.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
