EXL’s acquisition of iMerit, which specialises in model training, fine-tuning, and data annotation, is set to push the data annotation company deeper into enterprise AI.
The $310 million deal, which was completed earlier in August, also brings iMerit Founder and CEO Radha Ramaswami Basu into EXL as Executive Vice President and Head of iMerit. Basu has joined EXL’s executive committee, while more than 3,500 iMerit employees have joined EXL.
EXL says the combined business will provide an end-to-end AI platform that brings together its data, domain, and AI capabilities, with iMerit’s work in AI data, model evaluation, and AI applications in regulated industries.
“The deal is structured with $170 million paid upfront and up to $140 million linked to performance over the next two years, which reflects our confidence in the growth potential of this space and the quality of iMerit’s business,” Rohit Kapoor, Co-founder, Chairman, and CEO of EXL, tells AIM in an exclusive interview.
“This acquisition is a transformational pivot for EXL, deepening our vertically specialised AI capabilities and expanding our reach into high-growth AI technology sectors,” he adds.
He further states that the combination will help clients build, fine-tune, and operationalise AI systems for production environments, particularly in regulated industries where domain knowledge, context, and compliance are important.
Kapoor emphasises that iMerit brings specialised AI capabilities in model training, evaluation, reinforcement learning, red teaming, and expert-led human feedback in-house to EXL.
Basu adds that the two companies are already working to identify the first workflows, use cases, and AI models they can take to clients.
“We’re looking at the top workflows and use cases that will bring the biggest value. It’s about how AI can be used to go from innovation to impact,” Basu says.
She shares that iMerit was initially exploring a fundraise to scale the business when EXL’s approach turned into an acquisition opportunity.
Enterprise AI
Beyond model training and evaluation, Basu believes the next phase will be about making those models work reliably inside enterprises. She notes that it is particularly important in sectors such as healthcare, insurance, banking, and legal services, where AI systems must operate within complex regulatory and domain-specific environments.
iMerit has spent about a decade working with AI, initially in computer vision and autonomous mobility before moving into healthcare and foundation models. Its work now includes fine-tuning models, evaluating them and testing how they perform across specialised domains.
Kapoor says enterprise AI is entering a more demanding phase as companies struggle to turn experimentation into consistent business value. “As the cost of intelligence has fallen, the market consumption of tokens has increased exponentially,” he says, adding that enterprises are increasingly turning to model orchestration and smaller, in-house models.
These models can keep proprietary expertise within the organisation while offering lower costs, latency, and higher accuracy, he says.
25,000 Experts
One of the biggest assets moving into EXL is iMerit’s network of roughly 25,000 subject-matter consultants, or scholars, who work across specialised domains. They include cardiologists, radiologists, and agronomists.
Basu also says AI has made data annotation more important than ever. “It’s one of the highest growth markets today.”
The nature of annotation, however, is changing. Instead of simply labelling images, companies increasingly need experts to identify edge cases, assess model outputs, and test systems operating in the real world.
“A radiologist reviewing AI training data is not following a labelling guide. They are applying decades of accumulated judgement to separate a true anomaly from an artefact,” Basu explains.
For autonomous mobility, she says the requirement is moving beyond image annotation towards testing edge cases, developing test scenarios, and triaging models once they enter production.
The Bigger Opportunity is AI Compliance
Basu sees the combination of iMerit and EXL extending beyond model training and data services into AI compliance products.
EXL’s clients include companies in regulated sectors where AI systems may eventually need to demonstrate that they have been properly evaluated, tested, and audited.
Basu believes the next phase of AI will require deeper adoption across industries, with models being rigorously tested and audited before they are deployed at scale.
She sees services such as red teaming, model evaluation, reinforcement learning, and human feedback becoming part of the infrastructure enterprises need before deploying AI at scale.
“Turning technology into a compliance product is a very important thing to keep in mind,” Basu says. She adds “evaluation and audit readiness as a service” could help enterprises in regulated industries adopt AI while preparing for regulatory requirements rather than reacting to them.
For Basu, scaling AI requires more than speed. “AI just cannot be fast,” she says, arguing that resilience, governance, privacy, and guardrails need to be established before AI systems are scaled.
Kapoor says the acquisition brings complementary capabilities together, with EXL contributing industry expertise, enterprise data, and process knowledge, while iMerit adds model training, evaluation, and expert-led AI workflows. “Together, we can support the full AI lifecycle, from data preparation and model evaluation to workflow integration, governance, and production-scale deployment,” Kapoor says.
Healthcare is a Major Proving Ground
Healthcare is emerging as one of the most important areas for the combined business because it combines large amounts of specialised data with high-stakes decisions.
Basu points to applications ranging from patient triage and diagnostics to insurance authorisation, cancer screening, and clinical workflows. She also expects AI to play a growing role in surgical robotics and other physical AI applications.
She says the same pattern extends to banking, life sciences, autonomous mobility. and precision agriculture, particularly where an unchecked AI action could carry legal, financial, or human consequences.
“Those are some of the areas where we see the major, very high-scaling applications of AI,” she remarks.
The acquisition therefore gives EXL more than a conventional data annotation business. It combines iMerit’s AI model expertise and expert network with EXL’s enterprise relationships, domain knowledge, and operational capabilities.
Kapoor believes proprietary enterprise data could become a major competitive advantage over the next three to five years, as companies look to build specialised models using their own data rather than relying entirely on external providers.
“They will also need continuous evaluation and reinforcement learning to ensure those models remain accurate, safe, and relevant as business conditions change,” Kapoor says.
Basu says the two companies are already building the first set of offerings.
Your reaction