Analyst teams dedicated to repetitive yet operationally meaningful work were once a standard fixture inside large financial institutions. Loan files would be opened, hundreds of figures pulled from each borrower's financial records, verified for correctness, and fed into the relevant lending platforms. Variation in both format and language meant that even standard financial work required considerable time and effort — a complexity that firms such as Evalueserve built a notable part of their business around.

Financial institutions turned to analyst talent based in India partly because the work could be completed there at a lower cost than keeping those functions in-house. Those same companies whose revenues once depended on teams of human analysts are now developing automated software capable of handling a considerable share of that analytical workload without analyst involvement.

Spreadsmart, a proprietary tool developed by Evalueserve — a firm with deep roots in India's knowledge process outsourcing sector — took on the work of automating that workflow. Data extraction from financial records ran 70% quicker under the new system, with an accuracy rate of 99.6%, while connecting the software to the bank's lending infrastructure trimmed overall processing time by roughly a fifth, according to the published case study. A substantial portion of Evalueserve's revenue stream had taken shape around supplying the analysts, research infrastructure, and financial know-how that assignments of this type routinely demanded.

Automation tools that reduce dependence on human labour are now part of the product portfolio at firms across India's outsourcing sector — an industry that built its foundations on deploying skilled workforces but has since turned to developing software capable of handling the very tasks those workers once performed. Concurrently, technology vendors that once built their business models around low-touch product sales are now expanding their engineering headcount to tailor, connect, and run those offerings on behalf of large corporate clients.

For roughly two decades, India's KPO industry ran on an operating model that treated software provision and service delivery as distinct functions — a division that industry participants and analysts now describe as increasingly hard to define in practice. Demand for financial research, market analysis, competitor tracking, forecasting models, investment reports, and business intelligence ran across corporate operations throughout the US and Europe. Building and maintaining full in-house research teams came at a cost that offshore delivery through providers such as Evalueserve, SG Analytics, The Smart Cube, and Netscribes could not match — and those companies structured their entire offering around that labour-cost differential.

Staffing architectures that channel credentialled talent — graduates, MBAs, finance professionals, and researchers placed directly within client operations — are confronting a deeper organisational question now that AI tooling has loosened the once-assumed link between expanding headcount and expanding revenue. Annual reports yield their financial contents within a matter of minutes, while earnings calls lend themselves to transcription, quarter-on-quarter comparison, and condensed summary. Analytical groundwork that once rested entirely with human researchers — sketching out early company profiles, updating financial models, tracking competitor movements, and pulling together preliminary industry reports — has shifted considerably toward AI-powered tools.

Building reusable software tools capable of handling routine research and analytical work has become a core part of how Evalueserve positions itself, alongside the sector knowledge and professional services it has long offered clients. Three tools now sit within the portfolio: Insightsfirst, which handles market and competitive intelligence; Spreadsmart, designed for extracting financial data; and RiskSight, built around risk assessment functions. Designed to handle research automation, financial information extraction, and document analysis, the AIRA platform also extends into industry-specific use cases beyond those core functions.

Evalueserve's CTO, Rigvi Chevala, outlined how the firm had historically assembled a range of applications tailored to the particular needs of individual clients. By around 2019, a shift in direction took hold at Evalueserve — away from building one-off applications tailored to individual client requests and toward reusable products designed to address challenges shared across its customer base. Under the banner of what the company terms Product-Led Services, one illustrative arrangement saw a client channel $2 million per year into a tailored platform drawing on both customer behaviour data and market intelligence. At Evalueserve, the platform is shaped around each client's specifications before being woven together with the firm's own pool of analysts and domain specialists.

SG Analytics broadens its footprint beyond analytics alone, having built out capabilities spanning data engineering, cloud migration, software integration, and enterprise AI deployment. What started as an investment research operation later broadened to take in financial services, market intelligence, data analytics, and AI engineering. The platform pairs proprietary research tools with a team drawn from financial analysis and investment management backgrounds.

According to figures put forward by SG Analytics, adopting its methodology trims the time needed to complete research by 40% to 60%, while investment operating expenses drop by roughly 30%. A September 2026 review of investment-banking materials by SG Analytics put the share of confidential information memorandum preparation that automation could absorb at around 40 per cent, with the remainder of that work still dependent on human judgement.

Under its former name Course5 Intelligence, the company built a tool called Compete that draws on a steady stream of market and rival data gathered and processed on an ongoing basis. Developed under the ownership of WNS — which completed its acquisition of The Smart Cube in 2022 — Amplifi PRO now serves as a dedicated platform for procurement intelligence. Catalogued within that repository by August 2023 were upwards of 1,150 intelligence reports and dashboards, alongside pricing information spanning more than 9,000 individual commodities.

To illustrate how such arrangements work in practice, consider a knowledge-process outsourcing firm holding a yearly contract worth ₹1 crore to deliver financial research services. Workforce reductions among analysts, driven by AI tools, may push operating profit higher in the near term — though that gain faces pressure from the rising cost of software licences and infrastructure needed to support those same systems. When automation enters the picture, pricing discussions at renewal time may lean toward cost reductions on the customer's part. Gains in output efficiency do not translate automatically into lasting margin growth, and a downward revision of the annual fee arrangement could push the KPO's operating profit beneath the threshold it held before automation was introduced.

Coverage from a broadcast outlet in August 2026 noted that India's established IT-services firms were contending with comparable strain. Pressure from buyers was building on multiple fronts — fee structures tied to outputs rather than time spent, a push for lower costs, and the gradual pulling of certain tasks back under their own roofs.

Established in 2013, Quantiphi carved out its niche by taking on enterprise technology work, data analytics, and AI engineering contracts for corporate clients. Among its proprietary offerings, the company counts baioniq — an enterprise-grade generative AI platform — alongside Codeaira, a tool built to handle the automation of software development work. Tredence built ATOM.AI around a library of pre-built data engineering and AI components designed for repeated use across projects. One case study from a private equity firm put the drop in data-gathering and organisation time at around 90 per cent, translating into projected yearly cost reductions somewhere between $8 million and $10 million.

Under its rebranded identity as MathCo — a name that replaced TheMathCompany — the firm built NucliOS, an in-house platform designed to support the creation of AI-driven applications alongside business analysis tools. FY25 filings place the organisation's operating revenue at ₹501.5 crore, a figure that edged only marginally past the ₹499 crore recorded a year earlier, while staff costs alone — totalling ₹374 crore — absorbed roughly 84 paise of every rupee spent across the full expense base. The company's net profit climbed from ₹15 crore to ₹64 crore, driven partly by a 7% trimming of staff-related expenditure alongside broader cost reductions.

Fractal — whose shares began trading in February 2026 — developed Cogentiq, an enterprise AI platform, yet the underlying revenue model still operates along service-business lines despite that proprietary technological depth. Across FY26, Fractal's financials showed revenue of around ₹3,300 crore, with gross margins landing at 47% and adjusted EBITDA margins at 18%, while the company directed 6.4% of its top line toward research and development activity.

Godrej Industries brought Leena AI's workforce-facing platform into its operations, with the resulting digital assistant drawing on the conglomerate's pre-existing infrastructure to field IT-related queries from staff. Job postings from Leena AI in 2026 sought Forward Deployed Engineers — a category of technical staff whose day-to-day work placed them alongside enterprise clients, handling integrations, product configuration, and the handover of completed implementations into live environments. A team of engineers brought on board by Sarvam AI is now embedded with enterprise clients, adapting the company's models to fit those organisations' internal infrastructure and keeping the deployed applications running dependably.

Among the enterprise relationships Yellow.ai has built, Tata Capital, SBI Life, CRED, and LIC stand out as partners, while a separate project for Bharat Petroleum produced Urja — a customer-service assistant capable of handling LPG bookings and support queries across 13 languages. Connecting the work to BPCL's proprietary cloud environment and its suite of internal platforms formed a core part of what the undertaking demanded.

Hiring data from Indeed, covering the twelve-month window ending June 2026, showed job listings in India for Forward Deployed Engineers climbing at a rate roughly nineteen times higher than where they stood a year earlier, with that dataset published in October 2026. For IndiaMART, SquadStack built a voice-driven AI tool operating under the VANI platform name — one capable of reaching out to buyers, gathering their purchasing needs, maintaining follow-up contact on open enquiries, and routing matched leads across to the appropriate sellers. Drawn from published figures, the platform fields upwards of one lakh exchanges each day, with conversion rates running roughly 20% ahead of those logged through manual outreach and the cost attached to each verified lead trimmed by around 15%.

Regulatory filings submitted to US authorities show that Freshworks closed out 2025 with total revenue of $838.8 million, of which professional services brought in $9.4 million — a share of roughly 1.1 percent — while that segment posted a gross margin of 85 percent. A 47% gross margin at Fractal points to the particular balance between product revenue and services revenue within its overall business mix.

For graduates entering India's outsourcing sector, the route into the workforce followed a well-worn pattern: hired into analyst, developer, tester, or support roles, they worked through structured task-based functions before moving into more specialised positions over time. Automation through AI tools is reshaping the work traditionally tied to those entry-level roles. Job postings on Indeed for Forward Deployed Engineer roles across India climbed roughly nineteen times over the twelve-month stretch running from June 2025 through June 2026, based on data the platform released that October.