Agora heads into its third-quarter report with an awkward split between its business and its stock. Revenue growth has sped up for three straight quarters, management sounded noticeably more confident about conversational AI last time out, and the CEO said he planned a personal open-market purchase of roughly $20 million in shares. Yet the stock has fallen 10.5% since the last report, trailing a weak S&P 500 by nearly 5 percentage points. The report due after the close on November 16 has to show whether the acceleration is durable and whether the AI push can start paying for itself rather than diluting margins.

With no usable Street consensus for the quarter, management's own outlook is the bar. Agora guided third-quarter revenue to $41 million to $42 million, which works out to roughly 16% to 19% growth from the $35.4 million reported a year ago. Last quarter's $40.4 million beat the top of its range and grew about 18%, after 13.5% in the first quarter and 10.7% in the fourth quarter of last year. A result near or above $42 million would extend that streak and suggest the company is again guiding conservatively. A print near the low end would imply growth flattening in the mid-teens, not necessarily bad, but a break in the acceleration story that has supported the bull case. On the bottom line, Agora earned $0.02 a share last quarter against $0.03 a year earlier, so profits remain thin enough that small swings in costs matter.

Gross margin is the most important line after revenue. It slipped from 66% a year ago to 63.7% last quarter as lower-margin conversational AI work grew as a share of sales, though it improved slightly from 63.4% in the first quarter thanks to technical optimization. Another sequential gain would back management's argument that AI dilution is a scaling issue that can be engineered down. A renewed drop would suggest that faster growth is being bought with weaker economics. That ties into the company's stated goal of quarterly GAAP operating profitability by year-end. With a $1.0 million operating loss last quarter and research and development spending at about 38% of revenue, the third quarter should show a narrower loss if that target is realistic.

The AI narrative needs evidence, not just tone. Management said voice agents now match or beat human representatives on some call-center tasks, and it introduced the Gradium voice partnership, developer tools aimed at AI coding agents, a RoboPoint expansion into Japan, and new IoT chipmaker integrations. Still, conversational AI is targeted at only about 5% of revenue as an exit run rate by the fourth quarter, and executives described adoption as a slow, use-case-by-use-case process facing more competition. Signs of proofs of concept moving into production, along with further gains in dollar-based net expansion, which rose to 104% from 99%, would make that target more credible. Cash flow also needs watching after operating cash flow fell to negative $2.1 million, a shift management linked to bonus and buyback timing.

Sentiment has soured going into the print, with the bearish reading deepening to 35.4% from 22.9% a quarter ago. The stock, at $4.20, sits just above its 200-day moving average and in the lower half of its post-earnings range of $3.90 to $4.73. That range is narrower than the prior quarter's swing between $3.71 and $5.49, which suggests the market is waiting for proof.

The central question is whether Agora can keep growth in the high teens while showing that margins have stopped eroding. Doing both would make the stock's recent weakness look like a disconnect from improving fundamentals. Strong revenue with weaker gross margin and cash flow would support the skeptics' view that AI-driven growth is costing too much.