This Stock is Among TD Cowen’s Top Small-Midcap Ideas

TD Cowen highlighted ExlService (EXLS) as a top small/mid-cap idea, citing resilient double-digit organic revenue growth and margin expansion opportunities supported by strong free cash flow. Despite the stock down 32% YTD and 40% over 12 months, the firm argues the selloff is driven by sector AI-cannibalization fears rather than company-specific issues, noting valuation may imply negative terminal growth. The article also flags EXLS’ definitive agreement to acquire iMerit for up to $310M and its Databricks Gold Tier status, reinforcing AI-led demand momentum.

Analysis

EXLS looks like a beneficiary of the market’s current confusion between “AI software winners” and “AI infrastructure/services winners.” The second-order effect is that enterprise AI adoption usually creates more near-term demand for data cleansing, workflow redesign, model validation, and managed operations than for outright software replacement, which supports EXLS’s mix shift and gives it pricing power even if headline IT budgets stay cautious. That also means the cleanest beneficiaries are not just EXLS, but adjacent data/platform ecosystems like Databricks and specialized data-services providers; the losers are legacy labor-arbitrage consultancies whose revenue is more exposed to commodity staff augmentation.

The key risk is that investors may overextend the narrative that AI is automatically margin-accretive for services firms. If clients use AI to compress hours faster than EXLS can reprice contracts, growth can look durable while dollar-based revenue per engagement quietly deteriorates; that is the main falsifier to watch over the next 1-3 quarters. The iMerit deal is important less for size than for signaling where management sees the monetizable layer of the stack: model training and data ops, not generic consulting. If that integration disappoints or margins fail to expand, the stock can rerate back to a low-multiple value trap.

Near term, this is more of a sentiment/positioning trade than a macro call: the stock has already de-rated enough that any clean beat-and-raise plus margin commentary could drive a sharp squeeze over 1-2 earnings cycles. Over 6-18 months, the thesis only holds if organic growth stays low-double-digit and free cash flow converts into steady buybacks rather than deal-driven dilution. Consensus is probably underestimating how long AI implementation spend stays service-intensive, but overestimating how quickly that spend translates into structurally higher EBIT margins.

Key Decisions for Investors

• Go long EXLS into the next earnings window as a re-rating trade; thesis is a 15-25% upside move if management confirms sustained low-double-digit organic growth and margin expansion, with the main risk being decelerating bookings or weak pricing commentary.
• Pair trade: long EXLS / short a more legacy services name such as CTSH or EPAM for a 1-3 month relative-value trade; the spread should widen if the market keeps rewarding AI-enabled mix shift over labor-heavy delivery models.
• Buy a 3-6 month call spread on EXLS rather than outright stock if you want convexity on a sentiment rebound; this fits a catalyst path where multiple expansion happens before full fundamental proof.
• Set a falsifier alert on next two quarterly reports: if organic growth drops below the low-double-digit range or EBIT margin fails to improve, cut the trade — that would signal the AI narrative is not translating into monetization.
• Watch the iMerit integration and any commentary on Databricks-driven pipeline conversion; if either becomes a meaningful revenue contributor within 2-3 quarters, it strengthens the case for a longer-duration position.