Most investors likely know three basic things about the enterprise AI software company C3.ai (AI 0.42%). First, its founder and CEO Thomas Siebel previously founded Siebel Systems, and sold it to Oracle (ORCL -0.95%) for $5.85 billion in 2006. Second, its AI algorithms -- which can either be integrated into an organization's existing software or accessed as stand-alone services -- can be used to streamline operations, cut costs, improve employee safety, and even detect financial fraud.

Lastly, C3.ai was growing like a weed prior to its IPO in Dec. 2020, and its stock jumped from its offering price of $42 to $100 on its first trade. It would go on to close at an all-time high of $177.47 that same month. But today, its stock trades at about $13. It quickly lost its luster as rising interest rates drove investors away from pricier and unprofitable growth stocks.

The back of an android's head shatters.

Image source: Getty Images.

But today I'll focus on three lesser-known facts about C3.ai that might change your perception of the company: its constant CFO changes, its symbiotic relationship with the energy giant Baker Hughes (BKR 1.76%), and the recent changes to its long-term strategies.

1. Three CFOs since its IPO

Back in Oct. 2020, David Barter became C3.ai's CFO as it geared up for its IPO. However, Barter resigned last December and was succeeded by Adeel Manzoor, who lasted for a mere three months before stepping down this February and handing the reins over to Juho Parkkinen.

That high turnover rate was already worrisome, but each of these CFOs also changed the way C3.ai reported its number of customers. Under Barter, it reported its total number of customers, which rose 82% year-over-year to 89 at the end of fiscal 2021 (which ended in April of the calendar year).

But under Manzoor, C3.ai replaced that metric with a "revised" customer count, which included all the divisions and departments it served within a single company, as well as a new "customer-entity" metric that only included the parent companies or organizations of those individual customers.

According to Barter's original reporting method, C3.ai's total number of customers increased 47% year-over-year to 110 in the third quarter of fiscal 2022 (ended Jan. 31). But per Manzoor's new metrics, the company ended that same quarter with 218 total customers and 50 customer-entities, which represented respective growth rates of 82% and 28% from the year-ago period.

The bulls might claim that change provided investors with a clearer view of C3.ai's customer base, but the bears will argue that it merely inflated its customer count and year-over-year growth rates.

Parkkinen then stopped posting the total customer-entity count while sticking with its revised customer count, which increased 48% to 223 in fiscal 2022.

2. Its "lighthouse" relationship with Baker Hughes

When C3.ai went public, it told investors it would use a "lighthouse strategy", wherein it locks in large bellwethers of certain industries as lighthouses to draw in smaller customers. But one of those lighthouses -- Baker Hughes -- has been shining far more brightly than all the others.

C3.ai initially provided Baker Hughes with a three-year subscription to AI's software in June 2019. That agreement was extended to five years in 2020, then renegotiated last October and extended for an additional year through fiscal 2025. That deal, which includes a joint marketing arrangement, required Baker to pay annual revenue commitments to C3.ai.

Those commitments accounted for just under 30% of C3.ai's top line in fiscal years 2021 and 2022, and will still generate about a third of its estimated revenue through fiscal 2025. That customer concentration is troubling, because Baker actually renegotiated lower annual revenue commitments last year, divested its own equity stake in C3.ai, and invested in C3.ai's competitor Augury instead.

Therefore, Baker Hughes could simply let its deal expire at the end of fiscal 2025 -- and instantly reduce C3.ai's annual revenue by more than 30%. 

3. Tossing its original business model out the window

C3.ai now expects its revenue to rise just 1%-7% in fiscal 2023. That's compared to revenue growth of 38% in fiscal 2022 and 17% in fiscal 2021. It attributes that severe slowdown to macroeconomic headwinds, which have throttled the enterprise market's demand for big AI software contracts.

To stabilize its business over the long term, C3.ai will intentionally seek out smaller contracts to diversify its customer base away from its larger lighthouse customers, then gradually replace its subscription-based plans with a more flexible usage-based model that only charges fees when its tools are accessed.

That 180-degree turn raises even more red flags, since many cloud-based software companies prefer to pursue larger customers and lock them in with sticky subscriptions. C3.ai's abrupt departure from those strategies, along with its CFO changes and wobbly relationship with Baker Hughes, suggest that Siebel doesn't have a clear plan for his company's future.

Proceed with caution

C3.ai's stock might seem a lot more reasonably valued now at five times this year's sales, but the company is still on shaky ground. I believe its beaten-down shares won't bounce back until it stabilize its sales growth, resolves its customer concentration issues, narrows its losses, and straightens out its business strategies.