Image source: Generated by Unbounded AI

There are only two days left, which is the 2nd anniversary of the launch of ChatGPT, the “originator of general large model applications”.

The lives of many people and companies have been changed by it, but for the company Afiniti, this change may be the most drastic, so drastic that the last time they publicly talked about ChatGPT was in their own bankruptcy filing On the book.

In September this year, Afiniti applied to the Supreme Court of Bermuda for liquidation and liquidation under the Companies Act. They mentioned in the bankruptcy application that the release of ChatGPT led to customer interest in artificial intelligence "almost overnight. "Major customers are focusing their AI budgets on generative AI for customer service rather than Afiniti's predictive AI for optimizing call centers."

Afiniti is definitely a representative of the established AI star companies. In 2017, when AlphaGo caused a global deep learning craze with 60 wins, 0 losses, and 1 victory over Go masters from China, Japan, and South Korea, Afiniti had already received a total of four rounds of financing and was valued at US$1.8 billion. At that time, almost all technology media reported that Afiniti was very likely to become the first artificial intelligence software company to IPO in the United States.

But the good times did not last long. After the E-round financing in 2019, Afiniti has no listing news. As we see today, the rise of generative artificial intelligence (GenAI) appears to have had an irreversible negative impact on the company's business, leaving it a company with $518 million in debt. In the grand symphony of AGI getting closer and closer, large models emerging one after another, AI entrepreneurship becoming more and more common and low threshold, the bankruptcy of this old AI company seems to have played the first heavy note of the replacement of the old with the new.

#01. It also had a moment of unparalleled glory

As an established AI star company, Afiniti has almost all the factors we can imagine for a startup to succeed.

First of all, the founder, Pakistani entrepreneur Zia Chishti, has successful entrepreneurial experience and is also an investor with a Stanford Business School background. Before founding Afiniti, Zia Chishti invented the invisible orthodontic product "Invisalign" that became popular around the world and became a multi-millionaire in 1997.

In 2006, Zia Chishti founded Afiniti in Washington, DC. The problems it wants to solve with AI have very broad market prospects. Essentially, the product offered by Afiniti is software that uses AI technology to match salespeople and customers to improve interactions between the two parties. This algorithm breaks the traditional call distribution method and does not follow the order of calls. Instead, it refers to more than 100 databases to collect caller information, analyzes various characteristics, and predicts which pairs will bring the best results to the caller and the company.

Zia Chishti｜Source: Business Recorder

Although these data have not yet reached the tip of the iceberg of today’s general-purpose large models, they were very extensive and diverse at the time, ranging from the credit company Experian to the data companies Acxiom, Targus and Allant, as well as social networks Facebook and LinkedIn, and even census files from the area where customers call are covered. Next, the Afiniti system is like a magical "matchmaker" (also like Tinder in the sales and customer service center). It will accurately match the staff who is most suitable for this customer based on previous transaction records, and then directly call Transfer it. If a customer has a connection with the representative assigned to them, the likelihood of a successful transaction is greatly increased.

In 2008, Afiniti began filing multiple patents related to matching caller data to computer models. More than 90 patented technologies help Afiniti successfully realize the vision of founder Chishti, analyze and examine data and business information bound to customer identity information, find successful behavioral interaction patterns, and use these patterns in real time to improve health and corporate profitability. capabilities and customer satisfaction.

It is not difficult to see that the key to this model is to obtain large customers in high-quality industries. Afiniti also develops along this path. Starting from Washington, where it was founded, Afiniti first took over major telecommunications companies in North America, and then began to expand into insurance, financial services, hotels, medical and other industries. In September 2017, Afiniti signed an agreement with Huawei, and in April 2018, it reached cooperation with Avaya. In January 2018, Fortune magazine named Afiniti to its list of 100 companies “leading artificial intelligence.” By 2019, this technology has brought billions of dollars in incremental revenue to more than 250 companies in more than 20 countries, covering large companies such as financial services, insurance, hotels, and healthcare. Afiniti has grown to a company with 1,000 employees, supporting 1.3 million interactions and optimized matching every day. It is worth mentioning that Afiniti has also launched business in China. Maybe when you call the bank, the person on the other end of the phone is carefully selected by Afiniti. At the end of 2020, CB Insights, an American technology market information platform, announced a list of the eight AI unicorn companies most favored by capital, and Afiniti was on the list.

At that time, Afiniti’s total financing amount reached US$320 million, and the investor lineup was indeed star-studded, including McKinsey & Company, Swiss asset management company GAM, Vista Equity Partners, and Elizabeth. Murdoch, former CEO of VerizonIndividual star investors such as official Ivan Seidenberg, former Reuters CEO Tom Glocer, and former BP CEO John Browne.

Zia Chishti outlined a grand future vision for Afiniti: "Although the company's main application is the call center, in general, what we are doing is changing the way enterprises allocate internal resources and when contacting customers. employees. So it could be anything: text chat, how salespeople are assigned, in-store employees.”

Afiniti claims to be extremely effective and can lead to significant cost savings or additional revenue for customers. However, even after raising Series E, Afiniti’s overall operations are still losing money.

Picture source: Afiniti

In the last round of AI wave, corporate capital "gold-eating beasts" such as large models have not yet appeared, and Afiniti's losses are likely related to its own charging model. Afiniti adopts a "zero investment" and "pay based on revenue-increasing effect" business model. At that time, no company in the world was doing the same thing as Afiniti, and they also believed that through this business model, they would become the only company in the world that could accurately measure the effects of artificial intelligence applications.

However, this method is actually very dangerous. Since revenue is directly linked to the revenue-increasing effect of customers, Afiniti's revenue is highly unstable. In the early stages of business development or periods of market instability, there may be no income for a long period of time or severe income fluctuations. Once it is difficult to predict future revenue, there will be no guarantee of continued capital inflow to support R&D, operations and expansion in the future.

Furthermore, in order to provide services, Afiniti needs to invest in R&D costs, labor costs and technical infrastructure construction costs in the early stage. While waiting for customers to achieve the effect of increasing revenue, once there is no revenue for a long time, it may lead to The capital chain is broken.

If Afiniti can ride on the last round of AI craze and go public, it will be a good time and it will not end up going bankrupt. But in 2021, the "last straw that broke the camel's back" appeared: an employee made sexual assault allegations against Zia Chishti while testifying before the U.S. House Judiciary Committee, causing him to resign as chairman, CEO and director. The scandal also implicated many political figures on the board of directors at the time. Even former British Prime Minister David Cameron resigned as chairman of Afiniti's advisory board due to these accusations. This incident dealt a heavy blow to Afiniti, which turned from a popular fried chicken to a hot potato overnight.

Perhaps there are more stories under the water, but above the water, the shelved listing plan + the founder’s resignation due to the scandal + the lack of hematopoietic ability, plus the new era that has kicked off with great vigor. The wave of AI, AfiThis is how niti went into decline.

#02. "Heavy bleeding" continues

The news that Afiniti filed for bankruptcy caused an uproar in the U.S. enterprise services industry. Several well-known leaders People and professionals spoke out and participated in the discussion.

Andrew Traba, vice president of product marketing at NICE, was quick to criticize Afiniti, calling it a "fake AI company" that uses "smoke screens." In a LinkedIn post, Traba accused the company of failing to deliver on its promise of an AI-driven customer experience, claiming it favored "flashy demos and bold promises" over tangible business results.

Post by Andrew Traba｜Source: Linkedin

Laurent Philonenko, managing partner of customer experience consulting firm DeepTech Group, tried to defend Afiniti. He believed that the rise of GenAI was harmful to the company. business has been negatively affected.

Andrew Traba further countered that the emergence of ChatGPT has caused companies to increase their overall artificial intelligence investment, rather than shifting funds from other types of artificial intelligence to generative artificial intelligence.

Judging from real-life data, Andrew Traba’s view is obviously more tenable: a study by goldmansachs showed that in 2016, less than 1% of the companies in the Russell 3000 Index reported financial results. AI was mentioned during the conference call. However, since the release of ChatGPT, the frequency of these companies mentioning AI in earnings calls has increased significantly, from less than 1% in 2016 to more than 16% in 2023. About half of the increase occurred after the release of ChatGPT, indicating that generative AI has triggered broader market attention and investment willingness.

The fundamental reason why Afiniti became an "abandoned son" is that the market no longer favors loss-making growth companies.

Its bankruptcy is not only an individual incident, but also part of the adjustment of the entire industry, heralding the gradual decline of some established AI companies and the more obvious reshaping of the market structure by emerging technologies.

With the arrival of big models, the market has reached a critical point. Only companies that can quickly adapt and know how to utilize new AI will become the big winners. On the contrary, other companies will be eliminated.

Afiniti has many characteristics of established AI companies before the emergence of large models.

For example, the application scenarios are relatively narrow and fixed. Decision-making/predictive AI solutions are usually built on traditional machine learning algorithms and smaller-scale data, and are more focused on applications in specific industries, mainly large-scale applications. Customers are concentrated in traditional industry fields, such as finance, medical care, manufacturing, etc.

The current new AI represented by generative AIThe company uses large-scale pre-training models as its core technology, focusing on strong language understanding and generation capabilities, which can quickly adapt to different tasks and fields. New AI companies can take advantage of the versatility and flexibility of large models to have stronger innovation capabilities and market adaptability, and can quickly develop new application scenarios and explore new business models and service areas.

On the other hand, the business model of old AI companies is relatively traditional. They mainly obtain income through software sales, project customization and service charges. Financing channels are relatively limited, and they mainly rely on venture capital and bank loans. Due to market competition and pressure from technological updates, the profitability of older companies may be affected.

In the era of big models, old companies may need to invest a lot of money in technological upgrading and transformation, and face greater financial pressure. The business model of new AI companies is more diversified and innovative. In addition to traditional software sales and service charges, revenue can also be obtained through data licensing, platform cooperation, subscription services, etc. In the era of big models, new AI companies represent future technology trends and market opportunities, and can attract a large amount of venture capital and strategic investment in a short period of time. If it does not have a profitable business to support the application of large-model technology to its own products and services, and cannot obtain more empowerment from the outside world, it will be difficult for an established AI company to transform quickly.

At present, Afiniti said it will wind up its global entities in Bermuda and U.S. courts as part of a restructuring agreement with creditors to form a new company owned by major creditor Vista Capital Partners.

#03. A group of old AI companies fell together

It is worth mentioning that Afiniti is not the only old AI company facing the dual dilemma of debt and the impact of generative artificial intelligence.

Recently, Forward Health, a high-tech nursing cabin that has received more than $500 million in funding and has been operating for 8 years, also announced its closure. The company has not yet explained in detail the reasons for the closure. There is speculation that the story of investing heavily in renting expensive real estate in cities like New York and San Francisco and using AI to do diagnostics is no longer attractive.

Forward Health｜Source: TechCrunch

In addition to shutdown, the sharp decline in value is also another side of the current timeline.

Amelia, an enterprise-level conversational artificial intelligence company, was acquired by SoundHound AI for US$80 million, which is far lower than the US$175 million in financing it received in 2023; the value of the conversational AI solution LivePerson has also shrunk significantly. In 2024, the company's stock price fell below US$1, its market value was less than US$80 million, and it faced the risk of delisting. In previous years, its market value had reached US$6 billion, but with the rise of generative AI technology, its market competitiveness has also declined significantly.

Standing at the end of 2024At the end of the day, this hemorrhage is not over yet, as if any AI technology developed before ChatGPT is irrelevant. In 2025, what awaits us may be an even more drastic change.