As litigation teams prepare for 2027, the central challenge is no longer whether artificial intelligence will affect e-discovery. It is how legal organizations will incorporate AI into defensible, repeatable workflows while managing expanding data volumes, evolving technology, unpredictable costs, and changing talent needs.
Those themes shaped the ACEDS webinar, “What Market Research Reveals About 2027 Planning for Litigation Teams,” featuring Dera Nevin, Managing Director for Digital Insights & Risk Management at FTI Consulting, and James Bickley, Chief Revenue Officer at Contact Discovery Services. Drawing on market research from the 2026 E-Discovery Unfiltered report, featuring perspectives from 34 corporate and law firm leaders, the webinar offered a practical view of how litigation support teams should plan for 2027.
Moving From Experimentation to Operations With AI in E-Discovery
The 2026 E-Discovery Unfiltered report found that 65% of respondents identified generative AI as their biggest e-discovery challenge. That concern ranked above managing new data types and budget constraints. The panelists discussed e-discovery use cases for generative AI, and the importance of integrating them into established practices, workflows, staffing models, pricing structures, and quality-control processes.
Dera Nevin emphasized that legal teams should treat AI as both a source of new evidence and a capability within existing discovery processes. “Prompts, outputs, logs, and AI-generated business content may themselves become discoverable, while AI may also be used for document summarization, classification, chronology building, relationship analysis, witness-interview summaries, and privilege log generation,” she explained.
The panelists noted that rather than simply determining whether an AI tool can perform an e-discovery-related task, legal teams must also understand what the tool did, what information it relied on, whether the result is reliable, and how the process can be explained later. “We need to keep track of how AI is used, for defensibility and to validate that we are achieving the results we expect and intend when we are deploying these technologies within our matters.”
That observation captures a central 2027 priority, which is to document AI-assisted work as carefully as other discovery processes.
Establishing Governance For the Use of AI in Litigation
When considering incorporating AI more fully into its e-discovery process, a provider or internal litigation support team should start with the client’s use case, risk profile, data, and desired outcomes, and not a product demonstration. AI should be combined thoughtfully with established analytics, active learning, and traditional review methods. Different matters and data sets may require different combinations of tools. “The distinction between an AI pilot and a substantial AI program is not the technology itself; it’s the workflow, the governance, and the ability to measure the outcome,” said Bickley.
The panelist recommended that users ask whether the workflow reduces unnecessary volume, identifies important facts earlier, improves consistency, shortens timelines, and reduces costs. Ultimately, the issue is no longer whether a platform includes generative AI. Instead, litigation teams must assess whether the technology produces measurable improvements through a controlled, auditable process.
A Roadmap for 2027
The 2026 E-Discovery Unfiltered report found that data volume in a typical case increased over the prior year for 88% of respondents, while 58% identified new data types as a challenge. “Information governance and data minimization can help organizations understand what they have and eliminate information that no longer has value, and early case assessment can then help teams identify the files that matter before they commit to full-scale review,” she advised.
Bickley offered five priorities for litigation teams planning for 2027.
- Create a data source map to understand where data exists, how sources interact, and what information each source may contain. It should include Slack, Teams, mobile devices, collaboration platforms, ephemeral communications, and AI-generated content.
- Connect legal holds to systems and custodians, so they relate to specific data sources and reflect scope changes, because a notice that exists separately from collection and preservation activity may not provide sufficient operational control.
- Ensure that collection protocols are source-specific and documented because mobile conversations, encrypted messaging, collaboration tools, and AI interactions may present different preservation and collection challenges from email.
- Validate at each important transition. Teams should be able to demonstrate what they preserved and collected, what failed or was excluded, and how they addressed any exceptions.
- Build defensibility into the process, not as an afterthought, by establishing repeatable playbooks, decision logs, and escalation paths before unusual data sources or workflow problems arise.
Talent, Self-Service, and an Emerging Hybrid Model
As routine information processing becomes increasingly automated, it is changing litigation-support roles. Bickley described the trend as a move away from equating capacity with headcount. “Specialized tiger teams may include project managers, analytics specialists, AI workflow experts, subject-matter leads, and senior reviewers; their composition will vary
according to the matter,” Bickley predicted. Those smaller, highly specialized teams will still need strong technology, clear protocols, surge capacity, and defined escalation points.
The same principle applies to self-service. Self-service can be valuable when it puts the right person closer to relevant information, but it can be harmful when it merely transfers complex work to an unprepared user. A professional who can efficiently explore information may benefit from self-service, while one who relies on an AI system to generate unsupported legal conclusions may simply reach the wrong answer more quickly.
For 2027, the strongest operating model will likely be hybrid, with internal teams retaining ownership of strategy, governance, and business context, partnering with external providers that supply specialized expertise, operational capacity, independent validation, and scale.
Budgeting for Uncertainty
Fluid AI pricing, data volumes, labor requirements, and technical workflows make long-term budgeting difficult. The panel discussed token-based pricing and the possibility that current AI costs may not remain sustainable as providers seek to recover their infrastructure investments. Litigation teams should therefore build budgets around scenarios rather than fixed assumptions. They should account for ordinary matters, complex investigations, unexpected data sources, surge review requirements, and potentially higher AI-consumption costs. 2027 planning should prioritize adaptability over standardization.
Human Oversight Is Non-Negotiable
Ultimately, while AI may recommend, classify, summarize, or initiate actions, only qualified individuals should remain responsible for legal and operational consequences. Humans must define objectives, select inputs, test workflows, review results, validate outputs, and set escalation thresholds. “The goal is not to insert a human in every mechanical step; it’s to ensure that informed human judgment remains present wherever the risk and the obligations are affected,” said Bickley.
Conclusion
For litigation teams, planning for 2027 should focus on where technology helps, where human judgment must remain decisive, and how to prove that the entire process was reliable.
