Speaking to ETLegalWorld on the sidelines of the inaugural ETLegalWorld AI-Powered Legal Transformation Summit 2026, held on August 7, 2026, in Bengaluru, Shailesh Kumar, Director – Sales at SpotDraft, shared his perspective on the evolving role of AI in legal functions, the importance of governance and human oversight, and how legal teams can move from AI experimentation to measurable business outcomes.
Watch Here: SpotDraft’s Shailesh Kumar on AI, governance and the future of legal transformation
ETLegalWorld: What is your feedback on the First Edition of the AI Transformation Legal Summit 2026 and SpotDraft’s exclusive roundtable discussion?
Shailesh Kumar: I think the answer is something that we have gotten from everyone, right? And they said, the audience at the end of the day, people who are here, felt that this was the need of the hour. And it was not just an overview or a 30,000 feet view of what’s happening.
What was interesting to see is people speaking about insights and opening about how they are doing it at their end. And that shared perspective, right? something that helped everyone. So there quite a few people who are starting that journey or in the middle of that journey. Now, when they hear it from someone who has reached the end point, they now know the do’s, don’ts and how they can navigate pretty quickly.
ETLegalWorld: What are the three key takeaways you would like to draw from your panel discussion?
Shailesh Kumar: Very interestingly, the panel just for everyone’s reference was a discussion topic around how are legal teams going to be ready by end of 2027 with respect to their AI implementation. And one of the key takeaway was that overall there needs to be a hybrid structure, like assembly mechanism where you’re experimenting, doing your pilots, as well as having a flag as to say, when can we adopt this? When can it go into full deployment mode?
The second thing that I understood was for a larger part of the group who think that AI disruption is gonna take away jobs, what will we do? It only highlighted that the job roles are gonna change. And of course, initially there are gonna be lot of work because you have to manage between the pilots as well as your work, but eventually you will start like becoming managers of those systems per se.
And thirdly, most importantly was on the governance piece, Now organizations will have to have their own governance policies, boards, and like committees so as to ensure that they have the overall outlook on AI governance at the same time the ownership needs to lie with everyone who’s using the system.
Watch Here: SpotDraft’s Shailesh Kumar on AI, governance and the future of legal transformation
ETLegalWorld: Including governance, what are the biggest challenges legal teams face today while deploying AI, and how is AI helping simplify the legal function?
Shailesh Kumar: So from a governance standpoint, I think it comes back to evaluation, right? So, how is your system positioned? Where is your data going? Is your data being used for training? Like getting deeper into those parts of like what the tool has to offer. One is you have to do your checks there primarily.
And secondly, at the end of the day, even if the system is not using your data for training, one has to be conscious about what are the inputting into the system. You just can’t input anything and everything because at the end of the day, those systems are also going to have their indemnities around what you input, right? Because the data is purely being owned by you.
So it needs to take into consideration not only like the outcomes, I mean, the output of the system, but also the input because it’s most of like the AI solutions today, even from a learning perspective, are input first.
4. How do you see AI and automation changing the way legal, procurement and business teams collaborate on contract drafting, review, negotiation and execution?
Sailesh Kumar: So one of the major challenges with respect to overall business workflows that has a subset as contracting workflows is collaboration. The very reason why we have legal tools in place is for them to be able to collaborate seamlessly. Now, how the new age systems that are developing are helping primarily is by ensuring that we get to the end outcome faster.
So you are now able to use AI to review contracts based on the playbooks that you have set up. You are able to build your own clause library, right? Right off the bat. So primarily business is getting outcomes faster and it is also incentivizing them to go and do some bit of heavy lift or like parts that they can do in order to make the process even more faster.
Watch Here: SpotDraft’s Shailesh Kumar on AI, governance and the future of legal transformation
5. How do you address the most problematic aspects of AI, such as hallucinations and inaccuracies, to ensure that your product doesn’t face either of them?
Shailesh Kumar: Hallucinations are gonna exist, right? That’s the nature of the AI development, right? So was it with software. Every software had a bug till the time it reached a particular stage of evolution. Just think of this as just another software that’s, and these hallucinations being bugs. Now, what is happening over time is that these learning models are becoming more and more prominent.
And a lot of these bugs are primarily, or like the hallucinations are primarily because they’re being trained on large set of data not specific data. So one way to primarily reduce, at Like what we do Spotdraft, what we do to reduce hallucination is to invest in creating data, phantom data, because we are primarily not using data for training, and then kind of teach the system how to think more like a lawyer and act more like a lawyer. That’s the first part.
The second part is, law is one of those areas wherein you don’t get rewarded a lot for good work, but anything falls flat, you going to be pointed out, right? Things are going to be pointed out. So from that perspective, it has to be a human in the loop.
All of our systems are human in the loop, wherein each and every output ultimately goes through a human level of check. There is feedback that the system receives if it’s not accurate. And if it’s accurate, someone just goes and then validates it and then processes it forward.
So at any point in time, it’s not a purely AI-generated output that gets sent to the requester or the person who needs that. It is going through a proper loop of human validation, as well as the system getting that feedback to say whenever the output is not optimal, we kind of get that feedback and work on it.
6. What are the biggest misconceptions organisations still have about deploying AI? Are they sceptical, or does their optimism not align with the reality of how AI actually works?
Shailesh Kumar: So everyone wants to use AI, everyone wants to get the outcomes faster. So that’s what they think AI is used for. But what I’ve seen more often is that the answer for a lot of organizations to this problem is, let’s get a broad LLM license, and then we will try to figure out what we need to build. No, that will just get us going in circles and circles, and we’ll need the organization to invest more on a continuous basis.
I would say first figure out the problem that you want to solve and then decide how you want to solve it. Now there are certain things that you could pretty much solve using an LLM like pretty much like Claude, you could wipe code at ITRN. But there are problems that have data associations, right? Especially in like a larger business context. Now for those problems, it’s up to like the organization if they want to invest in building the entire tech stack out something that exists or
Can they just adopt it and over on top of it, use their data model slash LLMs to do data processing. So right now the answer to AI is stemming from like an requirement of we need to use AI to become better, right? Rather if it shifts to saying that, okay, these are the problems that we need the AI to solve for that will get like greater outputs slash outcomes.
7. How do you see the short-term impact over the next year and the long-term impact over the next three years in transforming the legal profession? And what innovations is SpotDraft investing in to stay ahead of that transformation?
Shailesh Kumar: So to answer your first part of the question, I think in the next one year, a lot of in-house legal teams especially would have already implemented like base level systems like CLMs, so on and so forth. And in the next three years, we foresee that the conversation is gonna shift from technology AI to more on outcomes context, right? And how much is the legal team able to contribute to the business moving faster. And that is the growth parameter or like the growth metric that they would be tracking.
Now, as far as spot draft is concerned, we have like a fully embedded, like AI team in-house, who is working on like creating the context layer, creating like the entire learning layer, reinforcement layer. So the idea primarily is whatever can be eliminated,as even like a thought to say and if there is data present that would be audible.
Watch Here: SpotDraft’s Shailesh Kumar on AI, governance and the future of legal transformation
