Want AI? Consider Workflow Automation Technology

The hype around Generative AI has contributed to a wave of interest in how to apply many forms of artificial intelligence, not just GenAI. As companies promote the adoption of AI enterprise-wide to remain competitive, some in-house legal departments are enjoying less restraint on investments tied to the use of AI-powered technology. Our observation is that many are doing it the hard way – trying to apply emerging GenAI tech to address specific, narrow pain points – and overlooking broader legal operations needs. 

While GenAI can perform incredibly useful tasks – from answering questions about policies to redlining contracts, these tools typically lack the workflow orchestration and administration features that are critical for managing processes and gathering performance management data. We are most bullish about the application of many forms of AI through or alongside purpose-built legal technology – most notably contracting, document management, and workflow automation tools. 

In a recent whitepaper, Corporate Legal Workflow Automation; Trends and Key Implementation Strategies, we explained the key benefits of low/no code tools for corporate legal departments. As the surge in interest in using AI-driven technology continues, we now take a deeper dive into how the leading legal workflow automation (WFA) tools function, and how many of the leading providers are quickly enhancing their solutions with GenAI. 

Low/No Code Workflow Autoamtion SOlutions' Capabilities

Before getting to the AI underpinnings in workflow automation tools, their broader capabilities deserve a bit of explanation. The leading solutions marketed to corporate legal departments (Bryter, Checkbox, Josef, Kim, MyLegal, Neota, OnitX, TAP, and Tonkean) provide workflow applications and “citizen developer” building blocks, along with strong integration, administration, document automation, compliance and security capabilities.  Geek alert - what follows is an explanation overview of the technology concepts we recommend that buyers consider when assessing fit for their anticipated use cases and resourcing models.  

Workflow Management – Controlling the flow of tasks, processes or actions is generally handled under one of two approaches: 

• Decision Tree-oriented solutions use a flowchart structure with predetermined rules and somewhat limited flexibility. Their logic-based approach is intuitive to lawyers, with easy-to-follow visual representations but interdependencies are more difficult to manage.
• Orchestration-focused solutions provide a centralized way to design, and control workflows involving multiple tasks, systems, and actors. The tools ensure that tasks are executed in correct order by gathering data, monitoring status, and triggering the next events/tasks. They are well-suited for complex workflows but managing the complexity can be challenging.

Integration – Interacting with adjacent systems, whether in the legal department or across functions, tends to involve up to three approaches that can be combined to facilitate making connections without engaging IT:

• Out-of-the-box integrations – Solutions most focused on taking a no-code approach have invested in providing pre-built integrations to commonly used collaboration tools such as Slack, Teams and Outlook to “meet clients where they are.” Such integrations are typically as simple as providing credentials which then enables a link between the two systems.
• Included Intergration Platform as a Service (iPaaS) – Pre-built connectors enable integration with a wide variety of applications and databases. Data transformation is part of the process. iPaaS often goes hand-in-hand with Orchestration, because events detected in other systems can trigger workflow actions.
• Third-party iPaaS – Other WFAs provide iPaaS through established partnerships with middleware solutions, such as Workato or MuleSoft. Separate subscriptions may be needed.

Administration – The activities related to the setup, monitoring, and maintenance of the automation environment. Includes configuration, data source and systems integrations, access control, and oversight of a portfolio of workflows. Some tools are better suited to either a centralized or a decentralized approach. Mastery can involve a learning curve, so administrator “mindshare” matters.

Document Automation – A number of workflow solutions automate the process of populating variable fields and generating personalized documents efficiently and consistently. Common use cases include the generation of high-volume / low-risk contracts.

• "Bulk" or "batch" processes involved data iteration and parallel execution and should include the ability or monitor for errors and utilize mechanism to take appropriate actions. Use cases including issuing addenda per new regulations or policies, an issuing outside counsel guidelines updates – often on-way processes.
• Campaigns are similar but are multi-step processes involving a variety of tasks and interactions. They provide a structured process (often via a "wizard") for executing and tracking progress. They are well-suited for two-way, iterative process, such as compliance audits and policy management processes.

Compliance and Security – These capabilities typically include audit trails and access controls. Most offer role-based layers of access (user, builder, administrator), while some provide fine-grained permissions using a wide array of attributes – rules, geography, roles, data, actions (e.g. run or edit), and the specific workflow in use. Advanced tools also offer data encryption and compliance reporting to meet industry-specific regulations and data security requirements. Most WFA solutions provide data residency in both the US and Europe, with some offering an ever-broader range of options.

Overall, these WFA tools have extensive capabilities, and we haven’t even gotten to the AI yet.  

How AI Supports Workflow Automation

AI (in its broadest sense) appears extensively across low/no code workflow automation solutions. Under this umbrella, more common capabilities include:   

• Intelligent document processing, using optical character recognition (OCR) and natural language processing (NLP)
• Data mapping and transformation between different formats, making it easier to integrate and process information from various sources
• Decision automation using predefined rules and data inputs, especially helpful for applying conditional logic to determine workflow paths
• Machine learning in low-code WFA to support continuous improvement based on new data and changing conditions

Some buyers may find it interesting to understand what goes on ‘under the hood,’ but focus should be on how the solution overall meets anticipated needs. Given the ‘black box’ nature of many AI solutions, teams typically find greater returns in thoroughly testing solutions and planning for a ‘human in the loop,’ rather than delving too deeply into the technical details of a particular tool. For example, providing a sophisticated paralegal with an AI-based contract redlining tool to provide a first pass analysis rather than rolling out such a tool to the business.

How Generative AI is Enhancing WFA TOOLs

More recently, the enthusiasm surrounding generative AI has driven emerging capabilities based on large language models to interpret queries, requests, and documents (all unstructured) and generate responses. Here are a few examples we recently found:

• Bryter has added “AI Agents” for employee Q&A reflecting handbooks and policies, to run health checks on policies, to search across a defined knowledge base, extract data and review contracts for deadlines and clauses.
• Checkbox has enhanced its “front door” to the legal department with an “AI Legal Chatbot” that automatically responds and routes requests based on company policies and playbooks.
• JosefQ, which uses GenAI for policy Q&A, can be implemented by itself or with Josef’s core WFA tool.
• Neota, in keeping with its modular approach, announced the release of its “Azure OpenAI Building Block” for extracting text, language translations, and drafting context-aware messages.
• OnitX’s “Catalyst Virtual Assistant” answers questions and drafts follow-up emails in the context of stored legal service request, matter, and contract data.
• Mitratech’s TAP announced enhanced features that include text summarization and natural language search.
• Tonkean’s “AI Front Door,” which classifies legal service requests for routing and reporting, includes a “LegalGPT” chatbot and an “AI Concierge” help feature for business users.

Before the advent of GenAI, these tools already used AI in various forms to interact with clients at the legal service request portal and to kick off and intelligently execute automated workflows. Now, with GenAI enhancements, the client interface is more conversational. Chatbots can answer questions clearly and accurately, and lawyers can save time with better search and strong summarization and drafting capabilities.

Legal departments that obtain WFA tools can begin harnessing the power of AI and GenAI within the infrastructure of software solutions that also provide integration, administration, document automation, compliance, and security capabilities. As AI evolves, so will the tools.

Obtain the Key Solution Components to expect, fact-based analyses, and evaluations of the nine Advanced Solutions in the 2024 Hyperion Low/No Code Workflow Automation MarketView™ Report.