FAQ
Services
-
We help legal departments get better results from their AI tools by structuring legal data into formats that LLMs can effectively understand and use.
We transform unstructured legal information into machine-readable datasets, evaluate LLM performance, and create specialized AI frameworks (Legal Intelligence Assistants) that understand your specific business and legal context.
Think of us as translators between your complex institutional knowledge and AI systems.
-
Our business model and approach are unique.
Business Model:
We help convert what you already know into a format commercial LLMs can use directly, and our goal is that you won’t need us after that unless you want to expand or change something.
Our business model is based on delivering results impressive enough that you’ll want to engage (or refer) us for other use cases when the time comes.
Approach:
Unlike general AI consultants, we specialize exclusively in legal applications of AI with a focus on making LLMs truly understand your unique legal and business context (including institutional knowledge that may never have been written down).
We don't build software, provide IT infrastructure, or sell pre-built AI products.
Instead, we focus on the critical "last mile" of AI implementation by structuring your existing knowledge and creating frameworks that help any commercial LLM consistently deliver outputs that are more accurate and relevant to your team.
-
No. We've designed our solutions specifically for legal professionals without technical backgrounds.
Our approach eliminates the need for coding expertise or complex technical infrastructure.
We handle the technical aspects of structuring data and configuring AI frameworks, while you maintain control over the legal and business guidance that shapes these frameworks.
-
We address common challenges legal teams face when implementing AI: inconsistent results, lack of business-specific context, hallucinations (factual errors), and lack of meaningful evaluation tools.
Our customers typically come to us after experiencing disappointing results from generic LLMs that don't understand their unique business context or (if Vining Legal is engaged) legal standards.
We help transform "that's interesting but not quite right" results into outputs that align better with your specific approach.
Implementation & Process
-
Absent unusual circumstances, we strongly recommend that all customers start with a Foundation Data Assessment (FDA) so they can make a data-driven decision about whether AI currently makes sense for their business at all, and if so determine the use case.
The FDA evaluates how well your existing data can support the use case you want to assess, and will also identify other use case candidates that may be supported.
For most customers, the primary input for the FDA is completing our structured questionnaire about your business context and priorities, with minimal document sharing required.
If you choose to proceed after the FDA, we create structured foundation data from the questionnaire and any supplemental content you’ve approved or chosen to provide.
Note: If you choose to provide documents to us as a foundation data source, we may use established, enterprise-grade AI platforms (OpenAI, Anthropic, Google) with robust security controls to support service delivery. We remove identifying information from these documents to the best of our ability before they are input into the LLM, but we strongly encourage you to proactively remove any information not approved for model input before providing documents to us.
If purchased, we’ll then perform an LLM context evaluation to measure performance improvement and deliver a report.
The process is designed to be minimally disruptive to your existing workflows.
-
It depends, but for our Core Foundation Package implementation typically takes 2-3 weeks from kickoff to delivery.
In general, the process is significantly faster if a customer designates an internal project lead who can dedicate 3-5 hours (total) to respond to questions and review content.
More complex implementations involving multiple extensions or custom orchestration may take 4-6 weeks, or longer.
Our standard process is designed to include only a small number of focused touchpoints rather than extensive time commitments.
-
We've deliberately designed our process to minimize the burden on your team.
Most implementations require roughly 3-5 hours of direct engagement from key stakeholders, primarily for an initial questionnaire, a review session, and final implementation guidance.
For specialized extensions, subject matter experts may need to provide additional context in focused sessions.
We handle the heavy lifting of data structuring and framework configuration.
-
For flat fee services, deliverables are deemed accepted on completion rather than through a formal acceptance process.
This approach allows us to offer predictable, competitive pricing and maintain efficient delivery timelines.
If you have feedback after delivery, we'll address reasonable requests during our standard 30-day professional services warranty, which is included with all of SLI’s standard offerings.
Customers with procurement policies that require formal acceptance testing may consider an hourly engagement model, which provides greater flexibility for revision cycles.
Business & Value Questions
-
We offer both flat-fee and hourly engagement models.
The initial Foundation Data Assessment is $3,500 flat.
Our Core Foundation Package starts at $3,500, with specialized extensions and LIAs available at published rates on our Services page.
Work beyond the fixed scope of flat fee services is invoiced at our standard hourly rates.
Absent unusual circumstances, we strongly recommend that all customers start with a Foundation Data Assessment so they can make a data-driven decision about whether AI currently makes sense for their business at all, and if so determine the use case.
We structure our services and pricing to deliver immediate value while building toward more advanced deployments (e.g., specialized extensions or LIAs) over time if you decide to expand.
-
Most customers see an immediate reduction in time spent reviewing and correcting AI outputs, increase in consistency in AI-generated content, and ability to handle higher volumes of routine legal tasks.
It’s difficult to measure the higher-order value of enabling your team to spend more time on strategic work rather than correcting AI errors.
-
If you purchase an LLM evaluation, we use the data points captured in your foundation data as a baseline to measure performance in specific areas (typically reduction in hallucinations and improvement in context-awareness).
For contract redlining use cases, Vining Legal will additionally evaluate legal reasoning and alignment with your Legal Frame.
Beyond metrics, success means AI tools that feel like they understand your business and produce outputs that require minimal correction or oversight.
Relationship with Vining Legal
-
No, engaging SLI doesn't require you to also engage Vining Legal.
While we offer complementary services, some customers may choose to work with just one company based on their specific needs.
That said, SLI is not a law firm and Vining Legal is. Use cases involving translating or incorporating legal judgment or advice (like contract redlining) into AI do require customers to engage Vining Legal to provide legal oversight.
We're happy to discuss whether a joint approach or single-company engagement makes more sense for your specific needs.
-
Joint offerings are structured with clear role separation: Vining Legal provides legal services (legal frame assessment and creation, and legal oversight), while SLI provides technical services (foundation data assessment and structuring, context evaluation, LIA configuration, and orchestration).
Depending on customer needs and preferences, this can be structured either as a legal engagement with VL (where SLI provides technical consulting as a subcontractor) or as separate coordinated engagements with both firms. Budget, security requirements, and procurement policies typically guide this decision.
Data & Security
-
There are no guarantees in life, security, or the evolving AI regulatory environment. Any time you engage a third party like SLI to provide a service there’s an inherent level of risk, and moreso when the service involves your confidential information and a frontier technology like AI.
That said, we've designed our approach specifically to minimize these risks across several dimensions:
Confidentiality & Legal Ethics: Our Foundation Data methodology extracts business context while preserving confidentiality through the following controls anonymization protocols.
Most information is collected through a questionnaire with primarily multiple-choice questions that don’t call for identifying information. We supplement this with publicly available information (with your approval).
None of the “default” foundation data points are inherently identifying. You can choose to include data which identifies you, like your company name, if you want an LLM to know it.
Document uploads are rarely required for the Core Foundation Package, though extensions like Contract Redlining may benefit from sample documents.
Note: If you choose to provide documents to us as a foundation data source, we may use established, enterprise-grade AI platforms (OpenAI, Anthropic, Google) with robust security controls to support service delivery. We remove identifying information from these documents to the best of our ability before they are input into the LLM, but we encourage you to proactively remove any information not approved for model input before providing documents to us.
For legal departments concerned about privilege, depending on the project we may be able to implement additional safeguards like working within your existing knowledge management structure or using information barriers.
Data Security: Rather than requiring unlimited access to your documents, our approach focuses on extracting and structuring targeted information elements within secure environments. We use established, enterprise-grade cloud providers with robust security controls. Within reason, we can also adapt to work within your existing security environment rather than requiring you to conform to ours.
AI Regulation & Compliance: While we aren’t a law firm and can’t provide legal advice on compliance matters, our Foundation Data approach creates clear documentation of what information feeds into AI systems, which in principle can support compliance with emerging AI governance requirements around transparency, data provenance, and traceability.
Our structured foundation data sheet provides a framework you and your legal advisors could choose to adapt to serve as a pillar of your AI compliance program.
Control & Flexibility: You maintain control over what information is shared and how it's used.
Our data minimization practices mean we collect only what's needed for your specific use case, with clear boundaries between what stays solely with SLI versus what might be shared with LLM providers under appropriate safeguards.
Note: At the risk of stating the obvious, it's important to understand that if you select a Foundation Data point that is inherently identifying (like company name) in your Foundation Data, and opt to include that data as part of more advanced AI deployments (like LIAs, if you want them to know your company name), that information will be processed by your LLM providers according to their terms.
No technology or method, including SLI’s, can magically detect and remove all identifying information if you've explicitly chosen to include it. We offer many foundation data point options that don't require sharing identifying information with third-party LLMs at all.
For more information on our data handling practices, please contact us.
-
SLI requires specific business context information to create effective Foundation Data. For the Core Foundation Package, we need basic information about your business description, products/services, industry, target customers, and communication style, plus 5 additional data points you select based on your priorities (such as business terminology, resource constraints, etc.).
Note: At the risk of stating the obvious, it's important to understand that if you select a Foundation Data point that is inherently identifying (like company name) in your Foundation Data, and opt to include that data as part of more advanced AI deployments (like LIAs, if you want them to know your company name), that information will be processed by your LLM providers according to their terms.
We've designed our intake process to be minimally invasive while capturing essential context needed for your LIAs to function effectively.
Most information is collected through a structured questionnaire with primarily multiple-choice questions. We supplement this with publicly available information (with your approval) to reduce your workload.
Document uploads are rarely required for the Core Foundation Package, though extensions like Contract Redlining may benefit from anonymized sample documents.
Note: If you choose to provide documents to us as a foundation data source, we may use established, enterprise-grade AI platforms (OpenAI, Anthropic, Google, xAI) with robust security controls to support service delivery. We remove identifying information from these documents to the best of our ability before they are input into the LLM, but we encourage you to proactively remove any information not approved for model input before providing documents to us.
For more information on our data handling practices, please contact us.