Ironman is an iconic movie and Mr Stark emerged as this brat billionaire turned Avenger who saves the planet from deadly attacks with his fellow Avengers. But there is one more character in the first installment of the Ironman that we all were enamored with – Jarvis. An AI assistant to Mr Tony Stark who is pretty much running his life. Back in 2008, it was just an imagination to have an all powerful AI assistant at the beck and call of its employer. But fast forward to 2026, Jarvis like AI assistants are a reality across industries and businesses.
Imagine having an AI assistant that can finish all the mundane tasks in 30 minutes leaving you with time to focus on more strategic cases, manage your workload better and still have time to live life just the way you like it. Sounds too good to be true?
Not anymore. Businesses are beginning to harness the power of AI and legal professionals are not far behind.
Currently legal professionals from enterprise legal teams are using AI platforms at an individual level, completing basic level tasks and completing the assignment and moving/forwarding to the next stage, the working is in silos and disjointed and email is the richest intelligent information repository for any organisation.
Enterprises are a combination of people + process + Systems [PPS] = to achieve a desired outcome collectively. Within enterprises there is interdepartmental dependency on various factors, which is critical for a smooth functioning of the overall business. If the legal department becomes a bottleneck due to never ending documentation and paper work, the whole system runs a risk of a slowdown in the business. To address this, AI led transformation is no more about gaining an edge in the market but it is fast emerging as the very spine of BFSI, fintech and legal industry.
For AI transformation to be rolled out across legal departments and enterprise at large, their integration into the macro work flow, litigation data and much more, the intelligence layer is built on taking into account people + process + Systems [PPS], so there is accountability on every individual to perform tasks and validate AI generated actions. [Text Wrapping Break][Text Wrapping Break]This is DHA in its element, being used across sectors while approaching enterprise AI solutioning – Data that is being analysed by AI with recommendations on actions and Human who executes those actions after validating the accuracy and compliance that is needed to be adhered to.
Lets understand every aspect of DHA and specially when Agentic AI is now being trained and integrated into critical work flows within the law firms and legal departments.
Data: Fuel for AI
We all understand that AI needs large organised data sets to perform its assigned task as close to accuracy as possible. But data in itself is so omnipresent and yet unstructured specially in the legal arena. The contracts, notices, compliance led reminders, record keeping of the customers of the bank, their loan portfolio, repayment and EMI schedules…the list is endless. An intelligent AI should be empowered to access through mountains of unstructured data and convert it into meaningful information which can be used to take time sensitive actions helping legal teams manage their workloads more efficiently and generating business for the company overall.
Data comes into the ecosystem from multiple sources.
Emails are a critical source of data but the data is unstructured and multiple topics get added in the email trail losing track of core agenda. Then there is a question of non standardised formats in which information is exchanged. Email threads with multiple vendors using multiple formats.
Across legal cases again, format changes from one case to the other. Then add multiple states with regional languages, regulatory changes that are ongoing with new clauses and multi-reporting to the mix and you have a cocktail of data but none of it is giving you any intelligent and action worthy output. Let’s not forget informal sources of data like SMS, WhatsApp, Teams and Slack.
For an AI provider, all these elements have to be considered when designing an intelligent tool for data gathering. AI should be able to read this unstructured data, input it in the platform and produce info in such a way that just by looking at the dashboard, a GC should know the critical matters that need top priority attention from the in-house legal teams and the external legal advisors and firms.
Human validation
Humans are creatures of habit so any big scale change is always met with doubts and suspicion.
Just like any other department, even the legal teams often question the value add AI can bring into their everyday workload.
The human VS AI debate will always be at the centre of it. Humans would naturally ask, we have a playbook that we follow and it has given us success over decades, we have built processes and perfected it painstakingly over years so why do we need AI? Our organisation playbook has the secret ingredient that leads us to successful outcomes, can AI replicate this. And when such questions are not addressed, AI is taken as more of a support and not a collaborator which directly impacts the quality of the tasks that AI can perform faster, better and more efficiently than its human counterparts. While the AI layer is designed to perform complex analysis like designing agreements customised to every clause and terms that change from one case to the other, deploying agents to multi-task several aspects of a single case, deep dive in data lakes to find the most accurate precedent that can strengthen a complicated recovery matter. But for many departments AI is assigned basic tasks like convert data into excel with graphics, send agreements with track changes, prepare crisp PPT for the cases update etc. All of these tasks don’t need a cutting edge AI or agentic deployment but a basic MS software or a generic LLM can also perform these tasks.
So humans even now are not fully harnessing the power of AI to make their case load manageable.
The AI advantage
AI output is as good as your prompt and organised data combined, AI/AI agents can work in Small Targeted problems of a Department or at individual level, for Summarisation – Building Key Pointers – Building infographics etc, but when we truly want to live the DREAM DAY, we need Jarvis to be deployed and integrated to accomplish complex tasks while the human can enjoy a quick coffee break or brainstorm with a colleague on devising a legal strategy on a complicated matter. This is how collaboration with AI looks like.
AI Agentic adoption for large enterprises is a journey and we on the path, as Taj Mahal wasn’t build in a day but the leader had the vision for it, AI Agentic adoption is a journey with leaders having a vision for it with lots of bumps but I believe is possible sooner, than we think. [Text Wrapping Break]
My Mantra for AI Adoption is the Only way forward is to not Assume Nothing, Ask everything to AI!
Happy Re-Thinking, Re-Learning, Re-Experimenting.
Authored by Hitesh Jirawla, Founder CEO, Cubictree