IGI Logo

Navigating the AI shift

Quick Pointers:

* Near-term growth momentum to continue driven by FS, Industrial and Retail

* No significant impact from industry shift toward output-driven model

We interacted with the management of LATENTVI (LV). As per the management, the changing AI landscape is likely to put pressure on the Diagnostics segment (60% of revenue), especially in technical areas, while the Domain and Consulting segments should maintain a steady state. Traction in Data Engineering remains strong, aided by the Databricks partnership and expanding avenues around Snowflake and GCP, which are expected to accelerate further as enterprises receive boardroom mandates to make their data AI-ready. FS, Retail and Industrial verticals should continue their growth momentum in FY27E, with 25–30% YoY growth, supported by underlying demand and scaling efforts to graduate the first potential FS accounts into the USD10mn+ band. LV is focused on driving revenue through advanced AI horizontal capabilities (~20% of revenue). As advanced AI becomes more progressive, hiring is becoming more selective and domain-oriented. Margins are likely to remain within the guided band of 23–25%, supported by improved utilization, selective hiring, and a higher fixed-price mix (80–85%). With a cashrich (~9% market cap) balance sheet, the company is exploring opportunities in the Healthcare & Life Sciences segment and Databricks capabilities. Following the recent market correction, valuations also appear to be more attractive from a target perspective.

We reduce our USD revenue estimates by 60bps/90bps for FY27E/FY28E, respectively, due to continued weakness in Hi-Tech and higher concentration mix of Diagnostics, but keep our margin estimates unchanged. The stock is currently trading at 26x/21x of FY27E/FY28E earnings after the recent correction. We assign PE multiple of 32x (40x earlier) to FY28E earnings and arrive at revised TP of Rs490 (Rs630 earlier). Maintain ‘BUY’.

Limited AI impact on business: AI-led uncertainty is unlikely to impact LV’s business as enterprises clients are reluctant to expose proprietary data and applications to external AI platforms. While AI is a productivity enhancer, it is not expected to replace jobs extensively

Shift to output-driven models: Although AI has not impacted business structurally, clients demand faster delivery within the same deal sizes and budgets – essentially expecting more for the same cost, driven by AI efficiency expectations. Also, a structural shift is underway from FTE to output-based engagement models. Majority of LV’s revenue is output driven; only 15-20% revenue mix is FTE driven. Also, there is a significant shift toward output driven by FTE & agentic co-worker, which would further reduce the FTE-driven business mode

Data as the foundation: Data readiness and infrastructure are prerequisites for any meaningful agentic or AI deployment. Data engineering is a core LV strength, contributing to 20%+ of revenue mix; LV targets to grow it to ~30% next year. To expand its data engineering capabilities & offerings, LV is expanding its partnership with Databricks and exploring newer opportunities with Snowflake & GCP.

Agentic AI strategy: LV is pursuing an aggressive agentic AI strategy, remaining open to all forms of engagement – investing, building and partnering. The company has already developed 4 agentic solutions, though adoption is still in the early stages. Looking ahead, LV plans to build 2-3 dedicated agentic workflows for clients and plans to invest USD4-5mn in startup investment to gain access to specific solutions. LV acknowledges that the cost of building agentic solutions will decline over time, but recognizes that the talent required is specialized and fundamentally different from traditional IT services profiles. This is where LV sees its long-term edge – by pairing this niche technical capability with its established domain expertise and advanced reasoning and modelling strengths, the company aims to carve out a differentiated position in the agentic AI landscape.

Revenue mix & growth: Advanced AI projects account for 20% of revenue, and the pipeline reflects a similar split of 20% advanced AI and 80% traditional services. FS, Industrial, and CPG segments are expected to deliver ~25% growth in FY27. RGM & Supply Chain represents a significant wallet share expansion opportunity for LV. Many FMCG clients spend USD2-3mn annually in this space, while LV's highest client engagement in RGM stands at USD500k, highlighting the headroom available through deeper client mining and hunting

Segment-wise opportunity: Segment wise, FS is expected to continue its strong double-digit growth in FY27E with opportunity to expand in several clients. One of the clients in FS segment is expected to scale annual spending with the company, which will make it a top client. Tech segment is likely to see single-digit growth in FY27 due to client-specific headwinds, as a top client is pulling work in-house. While the pipeline is healthy within the segment, the year will start with a drag.

Talent & hiring strategy: LV has been selective in hiring, skipping campus recruitment this year in favor of skilled, domain and technology-specific hires. This is due to the shrinking delivery timelines demand ready-to-deploy talent rather than freshers requiring ramp-up time.

M&A: LV has significant cash reserves, which it plans to deploy selectively toward 2-3 acquisitions focused on Healthcare & Life Sciences segment and expanding its Databricks capabilities, both identified as high-growth areas for the company. Additionally, LV is open to pursuing a large transformation deal, the size of which could represent 60-70% of its current revenue.

Valuations and outlook: We estimate USD revenue/earnings CAGR of 20.2%/22.6% over FY26E-FY28E. The stock is currently trading at 26x/21x of FY27E/FY28E earnings after the recent correction. We assign PE multiple of 32x (40x earlier) to FY28E earnings and arrive at revised TP of Rs490 (Rs630 earlier). Maintain ‘BUY’.

Above views are of the author and not of the website kindly read disclaimer