Talk Around Town: Zhibao Claims AI Agents Are Writing Half Its Code — Wall Street Ain't Buying It

What Folks Are Chattering About

Well, butter my biscuit and call it a compliance crisis — Zhibao Technology (NASDAQ: ZBAO), a small-cap Chinese insurance-tech outfit, got itself a stern letter from Nasdaq on November 21, 2025. The exchange confirmed in an SEC filing that the company had missed the deadline to submit its Form 20-F annual report for the fiscal year ended June 30, 2025, putting it sideways with Listing Rule 5250(c)(1). That's about as welcome as a skunk at a Sunday picnic.

According to Zhibao's own announcement, the deficiency notice didn't immediately yank its shares off the exchange, and the company said it had 60 calendar days to hand Nasdaq a plan for getting right. StockTitan's aggregated news feed then reported that Zhibao filed the overdue Form 20-F on January 9, 2026, and that Nasdaq issued a notice on January 12, 2026 confirming the company had pulled itself back into compliance — a resolution the aggregator sourced from a separate Nasdaq communication.

What the Company Says It's Up To

Now, Zhibao didn't just limp back into compliance and sit quiet on the porch. The company describes itself — in its own press materials — as a pioneer of what it calls a '2B2C' embedded insurance model, which Zhibao says it originated in China back in 2020. The way the company tells it, that model tucks digital insurance products inside the apps, websites, and back-end systems of internet platforms, regular businesses, and even government agencies — kind of like hiding a vegetable in a kid's mac and cheese.

Then on February 13, 2026, the company announced — again through its own press releases distributed via Newsfile Corp. — the launch of ten new AI agents covering what it describes as product design, platform operations, promotion, and customer service. The company claims to have built over 40 proprietary digital insurance solutions, and says it has seen 70% year-over-year revenue growth, though none of those figures have been independently audited or verified by any third party.

The headline grabber in Zhibao's own announcement is the assertion that one specific agent — described as an insurance application development tool — is capable of auto-generating more than 50% of the development team's daily code requirements. The company frames this, in its own marketing language, as evidence of an 'intelligence-driven' transformation. That is a heck of a claim, and as of this writing, not a single independent researcher, analyst, or benchmarking body has weighed in to say yea or nay.

What Is Actually Confirmed

The Nasdaq deficiency itself is rock-solid confirmed — it's right there in the SEC's EDGAR database as a Form 6-K filed November 21, 2025. The notice cites failure to file the annual report as the specific violation. That part ain't gossip; that's a federal document, plain as a screen door on a submarine.

StockTitan's aggregation of a subsequent Nasdaq communication provides limited but real corroboration that compliance was regained following the January 9, 2026 Form 20-F filing. Investing.com independently reported — citing observable market data — that Zhibao's stock was trading around $0.86 per share with a market cap of roughly $29 million at the time of the AI agent announcement, down nearly 39% over the prior year. That price information reflects publicly traded market activity, not a company claim.

What Nobody Has Verified

Here's where the mud gets deep, friends. Every single AI performance figure Zhibao has put out — the 50%-plus code auto-generation claim, the 40-plus proprietary solutions figure, the 70% revenue growth number — flows exclusively from the company's own press releases. There is no independent analyst coverage, no peer-reviewed benchmarking, and no regulatory validation of any of these capability assertions. Calling those settled facts would be like calling a fish story a nature documentary.

Separately, Zhibao announced a cooperation agreement with Chengdu's Tianfu Citizen's Cloud platform, a civic app, which the company says gives its subsidiary potential reach to nearly 20 million registered users through embedded insurance offerings, according to StockTitan's aggregated report of the company's announcement. A beta deployment was said to have launched on June 30 under a three-year agreement. However, actual commercial results — premiums written, policies sold, revenue generated — have not been independently reported by any outside party. 'Potential access' to 20 million registered users and actual paying policyholders are two very different animals, as any insurance underwriter worth their salt would tell you.

Analysis: The Gap Between the Pitch and the Pasture

This is analysis, not reporting: the contrast between Zhibao's AI-forward self-description and its actual market standing is striking enough to warrant a raised eyebrow or two. A company that describes itself as an intelligence-driven technology enterprise — the company's own framing — probably shouldn't be tripping over a routine annual-report filing deadline. Back-office stumbles like that tend to undercut the futuristic narrative faster than a busted tractor undercuts a harvest schedule.

It's also worth noting, as analysis, that the market's apparent indifference — a nearly 39% stock decline over the year preceding the AI agent launch, per Investing.com's report — suggests investors have so far found the company's self-reported AI capabilities about as persuasive as a rain dance in a drought. That could mean the market is wrong and Zhibao's technology is genuinely ahead of the curve. Or it could mean sophisticated market participants are appropriately skeptical of unaudited performance figures from a $29 million company with a compliance hiccup on its record. Without independent verification, this publication cannot tell you which interpretation is correct.

The embedded insurance concept itself — weaving coverage products into civic platforms and consumer apps — is a real and growing category globally, and the Tianfu Citizen's Cloud partnership, if it materializes into actual policy uptake, could be meaningful for a company this size. But 'could be' is doing a lot of heavy lifting in that sentence, and the chasm between a signed cooperation agreement and millions of active policyholders is wide enough to lose a combine harvester in.