The legal industry’s AI conversation is shifting from whether to experiment with the technology to how it can be embedded into everyday workflows at scale.
The panel, “Beyond Experimentation: What Will AI Adoption in Legal Look Like by 2027?”, was moderated by Shailesh Kumar, Director of Sales, SpotDraft, with Suresh Kalarikkal (Director & Head-Legal, India & Middle East, Thoughtworks), Kavitha Gupta (Director-Compliance Officer, India, SAP), Damandeep Kaur (Regional Chief Legal Officer, India & APAC, Teleperformance) and Panduranga Acharya (CXO, General Counsel, Zepto).
Kalarikkal said sustainable AI implementation begins with a genuine business problem, not the technology itself. Buying a tool and then searching for uses is fundamentally different from spotting an operational problem and designing a workflow around it, the latter is far more likely to survive past experimentation.
Repeat usage by employees, he added, is the clearest sign a solution is working. He stressed embedding AI output inside a defined process with a human owner and escalation mechanism, turning it from a standalone productivity tool into part of the operating workflow.
Kaur said moving into adoption doesn’t mean organisations stop experimenting. Teleperformance began its AI roadmap 18 months ago, building a portfolio at different maturity stages, real-time contract risk tracking, continuous audits pulling data across systems, and an emerging agentic dashboard designed to spot issues like unpaid invoices before they escalate. By 2027, she expects the model to stay hybrid which includes continuous experimentation alongside proven use cases moving into production.
Kalarikkal said AI’s impact extends beyond in-house departments. Law firms are investing heavily in proprietary AI, which will reshape staffing, service delivery and pricing, while in-house teams build capacity to handle work once outsourced. He cautioned against clients simply demanding an “AI discount,” predicting instead a shift toward outcome-linked pricing for repetitive work, even as hourly billing persists for complex matters.
Training, Not Just Tool Access
Acharya emphasised training employees rather than just granting AI access. After finding optional learning programmes drew insufficient participation, Zepto built structured internal training with internal trainers helping colleagues use AI practically. The goal, he said, is helping employees identify repetitive tasks they no longer want to do and determine what can be automated.
Gupta advocated for a multifunctional AI governance committee spanning infosec, privacy, legal, compliance, HR and risk, since responsible AI touches IP, cybersecurity and deployment architecture, not just privacy. Kaur added that governance itself must speed up as AI volume grows; a “mother” framework can set global policy while local procedures handle jurisdiction-specific rules, especially crucial across APAC’s varied regulations.
Governance Should Enable, Not Block
Acharya argued legal and compliance must be involved from product design, not after launch, especially as businesses build platforms combining large volumes of customer data to generate insights or simulated profiles. Business teams may technically build such products without legal, he said, but that doesn’t make them compliant. The stronger model embeds legal, compliance and security into development so innovation proceeds within clear boundaries.
By 2027, the panel agreed, successful legal departments will be those starting with business problems, embedding AI into repeatable workflows, developing tech-literate lawyers, and building governance that moves as fast as innovation itself, making governance an enabler of scale, not a brake on it.
The AI-Powered Legal Transformation Summit 2026 by ETLegalWorld is currently being held in Bengaluru. The summit has brought together general counsel, in-house legal leaders, and technology innovators, to chart the next phase of legal transformation in an AI-driven world.
