The $3.6B Silicon Kolkhoz: How CCaaS and Prosumers are Liquidating Human Capital
The plots of land in the Silicon Kolkhoz are selling for a premium while the C-suite pursues its epic quest for AI-driven ROI.
Did you get the memo?
$3.6B. It is the big number that slipped under the radar of the lukewarm LinkedIn commentary mill yesterday.
And yet!
$3.6B is the handsome sum Salesforce put on the table to acquire a firm, Fin, formerly known as Intercom.
What does Fin, formerly Intercom, do? They produce advanced AI customer service agents. Here, “advanced” means going beyond deceptive chatbots, which are limited to parroting FAQs already available on the websites where the lazy chatbots sit. Indeed, “advanced” means that AI customer service agents resolve complex customer queries end-to-end (we will get back to this later in this insight), across every channel, including live chat, email, WhatsApp, SMS, phone, and Slack already acquired by Salesforce in July 2021 for $27.7 billion (the largest acquisition in Salesforce’s history to create a digital headquarters competing with Microsoft Teams).
The funniest part of deciphering acquisition press releases to deliver strategic insights to our clients is the shell game with adjectives. A true sleight of hand, passing off fool’s gold as pure gold.
But our materialist perspective lets nothing slide, and we love to dig up the reality behind every qualifier. If AI agents claim to address ‘complex’ queries, then ‘complex’ doesn’t function as a lazy filler buzzword. Instead, it signals the new reality of the algorithmic frontline—the very place where the mother of all battles is being mapped out and decided: the appropriation of value!
El Lissitzky, Revolutionary Process (1923) — The interface design for the Silicon Kolkhoz was ready a century before tokens were even invented.
Complex customer queries
“Complex” means that autonomous AI agents can execute workflows. First, it uses multi-step reasoning to break down the customer’s query into sub-problems, think in sequence, and adapt to what it finds. Second, advanced AI agents use cross-app integration to address customer queries. Because the solution rarely resides in a single database, the AI agents act as orchestrators, pulling data from and pushing data to multiple enterprises’ databases via MCP or secure APIs. Finally, AI agents handle human customers who never speak in clean API queries, though there is always a way to bend human customer requests to fit the AI agents’ ecosystem (we will elaborate on this in the next section about prosumers).
In a nutshell:
Before the advent of the agentic platform, “complex” meant the chatbot would say, “Let me transfer you to a human.”
Today, “resolving complex queries end-to-end” means the AI has the authority and capability to complete the job.
Why did Salesforce acquire Fin?
The Silicon Kolkhoz logic of centralized power must extend to every unit of production across the market. The reduction in payrolls—driven by post-COVID adjustments, demographic shifts, and an uncertain economic climate that curbs investment and, in turn, hiring—would have threatened agentic-flavoured SaaS vendors had AI not emerged from its long winter in 2022.
But here is the twist.
AI offers new opportunities to capture value by taxing enterprise operating systems through tokenization. Propelled by aggressive propaganda from stakeholders who have staked their entire nest eggs on the AI gamble, agentic platforms are becoming a reality and redrawing the algorithmic frontline.
AI agents are taking centre stage in the operating systems of businesses, large and small, controlling the production of value, capturing a slice of it by using tokens as a pricing metric, and pursuing their grand replacement—that of human agents, particularly middle management, which is too slow, too cumbersome, and too fussy.
In this cutthroat competition to dominate the Silicon Kolkhoz, Salesforce is trying to deepen customer engagement (marketing fluff for presumption, where customers must produce what they consume) and expand its CCaaS (Contact Center as a Service) presence with Agentforce Contact Center. Salesforce’s acquisition of Fin aims to bring its customer agent platform to companies of all sizes and to expand Salesforce’s ability to deliver Fin’s proprietary AI model, Apex, across enterprises.
The goal is clear: migrate all call centre and customer support infrastructure to the Cloud—a Silicon Kolkhoz blueprint of centralized power over value-producing units. Instead of paying a yearly per-seat fee to access the platform and control its operating system and customer relationship management policies, client enterprises depend on the agentic platform’s will and architecture, which control access to client enterprises’ strategic resources.
The mutation of CCaaS from a human-based platform to an agentic platform is the ground zero of the Silicon Kolkhoz expansion. The forced-march pivot of CCaaS platforms toward pay-per-conversation or token-based pricing is the focal point because CCaaS is the direct point of value capture:
It is within CCaaS that all raw data and the history of customer frustrations converge—the perfect fuel for Databricks or Snowflake models.
Giants like Salesforce are trying to short-circuit pure-play CCaaS vendors (like Genesys) by offering enterprises integrated AI models, such as Agentforce, to directly handle customer calls and update the CRM.
CCaaS is thus becoming the software interface that orchestrates workforce automation and determines which platform will collect the tax on customer interactions.
Customers? You meant prosumers!
In the CCaaS agentic platform model, the term “customer” is misleading and fails to capture the material reality of the new role it endorses. Customers are value producers who consume the value they (co)produce.
They are truly prosumers.
The drift toward a CCaaS agentic platform model does not merely redraw enterprise operating systems and the place of employees within the customer-company relationship. It forces a complete redefinition of the role of the company’s customers. If service, in the traditional sense of the term, implies a social relationship between two human agents—a consumer and an employee who agree to co-create the service—its agentic form leaves little room for the glamour of emotional intelligence and empathy. No. The agentic mode of service production via CCaaS demands the conscription of the consumer into labour, forcing them, from now on, to produce the very service they wish to consume, thereby de facto turning them into a prosumer.
Thus, the prosumer will connect to the platform to update their information, renew their contract, complain about a failing service, or schedule an intervention, bending to the AI agent’s directives (and, meanwhile, performing emotional labour to curve their human emotions to fit in the binary logic of the agentic platform), which will reshape the production flow to suit its technical constraints. The prosumer must therefore adapt their productive labour to the platform’s rigid interface. Much like a colouring book where each number dictates a colour, the prosumer will allow themselves to be guided by the CCaaS platform’s AI agent, which will issue orders in a sequence masterfully engineered to optimize its profits.
This model has been repeated time and again across the service industry over the last thirty years, known as the McDonaldization of services. The mantra—efficiency, predictability, measurability, control—can now be applied to any transaction or interaction of any company in any industry, guaranteeing a permanent and sustained economic rent to the owners of the Silicon Kolkhoz. Furthermore, the prosumer’s labour can be monitored, evaluated, and audited to categorize them based on their capacity to generate value.
Beware of the slackers who cost more than they bring in! The platform will be more than happy to remind you, in no uncertain terms, that a private enterprise has the right to choose whom it does business with!
Yet this prosumer-agentic platform relationship creates a new value conflict between actors with antagonistic interests. Both the prosumer and the owner of the agentic platform feed off the same beast: the client enterprise. The prosumer will want access to a seamless, lightning-fast, and highly efficient service production. Time is money, and this scarce resource must not be wasted on endless prompts and queries designed to feed a platform thirsty for qualified, trained data. On the flip side, the platform—which bills for performance per click, request, or prompt via its token system—has every incentive to drag out the experience to maximize its revenue.
The whole art of the game will thus rest on the platform’s ability to maximize token consumption without ever hitting the breaking point. Customer experience can wait. Consequently, the client enterprises that have made the choice—conscious or not—to trade their good old platform and its customer support teams in Global South BPOs for a gleaming agentic platform will feel the sting of the bill, witnessing an explosion in their operating costs per customer query.
As for the prosumers, they will not fail to notice that access to corporate resources and services—once guaranteed by the mere act of purchase—must now be earned through the sweat of their brow and their cognitive capital, put to work training the platform’s AI model. The prosumer is a click-worker in the making, and it will fall to the client enterprises to invest in their marketing departments to deploy communication campaigns that glorify these prosumers—now supposedly ‘freed’ from the yoke of rigid, siloed, and costly procedures thanks to the new agentic platform, the spearhead of the Silicon Kolkhoz.
In the same way that workers and youth were daily exposed to their own glorification and heroization by the central Soviet power, prosumers will be acclaimed for their loyalty, trust, and enthusiasm in producing free value for the company, and ultimately for the CCaaS platform.
For it is the enterprises, not the owners of the CCaaS platforms, that will be held accountable in the event of a failure, and they alone will bear the consequences of service disruptions, friction, anxiety, frustration, and other disasters these platforms will inevitably generate. This is no longer a dystopian forecast; it is a current reality. Here is a concrete example of this very phenomenon, which I had the pleasure of commenting on for CBC News last week.
A risky operation
Last week, I was interviewed by CBC about the case of Quinn, the BMW’s dedicated “advanced AI customer service agents resolving complex customer queries end-to-end across every channel.” When Zack wanted to sell back the BMW he bought three years ago, he sent a “complex” query to Quinn: How much would you buy back my car? Quinn, oblivious to details despite its grand pretensions, confused the selling price from years ago (CA$27,162.79) with the vehicle’s current value (CA$20,000) and thus made an enticing offer to the client, proposing to buy it back for the exact same price it had sold it for three years prior—effectively allowing the customer to drive a BMW for free for three years.
But reality hit back, and the human workers assigned to babysit Quinn flagged the gross error and withdrew the offer. The client was furious and used the media to leverage his bargaining power. After the media buzz, which I contributed to through my interviews with CBC in French and English, BMW had no choice but to acknowledge the mistake and make it right for the customer.
What could be seen as an unfortunate but regular service failure, requesting BMW to deploy its service recovery strategy, is an interesting case study of value production, extraction, and risk exposure. The few-thousand-dollar gap between the offer and the vehicle’s actual value (around CA$7,000) is not much, provided the mistake is not applied at scale. But the reputational risks are tremendous. And any entry-level marketer knows how much it takes to build a positive brand image, and how little it takes to destroy it. As we say in French, “La confiance se gagne en gouttes et se perd en litres.”
Who bears the consequences of the risk? BMW. Not the platform. Even less so, its superpower AI agents. But we know the drill if BMW complains. The hyperscalers, the AI vendors, and their consulting-group friends will all parrot in unison that BMW must revise its workflow and dataflow, train its workforce, buy new APIs, provide more open access to its data for better, more accurate offers, and deliver a seamless customer service experience. In other words, do not expect AI agents to adapt to your reality. Adapt your reality to the AI agents’ “advanced” capacities. And, meanwhile, bear the cost and the risk alone.
Why does it matter for human capital leaders?
Human capital is a very straightforward term despite the modern management fluff around workplace well-being, engagement, and meaning. It is an asset that is invested to produce value—the fig leaf for the word profits.
For human capital leaders, the temptation to replace human payroll with agentic platforms is marketed as the ultimate cost-cutting panacea. But shifting from controlled employees to an unpredictable army of prosumers introduces systemic risks that the C-suite is fundamentally unprepared for.
Here are the 5 critical points of friction and risk that human capital leaders must consider before diving headlong into the Silicon Kolkhoz:
The agentic platform cannot tolerate sharing its algorithmic authority with human oversight. Consequently, the first casualty of this shift is middle management. However, eliminating middle managers to appease the platform removes the transmission belt between the corporate board and the operational frontline. Without this human layer, organizations lose their capacity for real-time nuance, internal mentorship, and contextual governance—leaving the company exposed when the algorithm hallucinates (and Quinn reminds us that hallucination is hardwired into the very nature of AI agents).
It is inherently easier to monitor, train, supervise, and align paid workers bound by an employment contract than to manage an unpaid, conscripted army of prosumers. Employees can be disciplined, upskilled, and culturally integrated. Prosumers, raw and untrained, owe no loyalty to the firm. Expecting them to seamlessly execute the corporate workflow within the rigid confines of an AI interface is an operational gamble that invites continuous friction. Side effect: this operational shift forces a shotgun marriage between CHROs and CMOs—two roles rarely accustomed to sharing a foxhole. The sparks are a foregone conclusion; I leave the nature of that ‘collaboration’ to the reader’s imagination.
In the traditional economy, friction works in the employer’s favour: a disgruntled worker cannot easily jump to a new job overnight because of market constraints. The prosumer economy reverses this dynamic. A prosumer frustrated by a rigid, token-hungry AI interface can switch to a competitor with a single click (almost). By trading a loyal, captive workforce on payroll for a volatile network of prosumers tied to an external platform, the enterprise completely abdicates control over customer retention. And let’s not be naive: prosumers will not sit idly by while being plucked clean. They’ve long understood that a whole ecosystem specialized in gutting overly greedy corporations—class-action lawyers, advocacy groups, and public regulators, to name a few—is more than happy to rush to the rescue of these ‘exploited’ prosumers. They will gleefully slap heavy fines on the ‘bad’ companies that fail to sugarcoat their unadulterated thirst for profit. They’ll call it “prosumer litigation,” you can mark my words.
When human agents handle customer queries, human capital costs are fixed, predictable, and bounded by salaries and shift schedules. When the frontline is handed over to an agentic platform, costs become fluid and variable. As prosumers inevitably push back against rigid AI interfaces, consuming vast amounts of tokens across circular queries to resolve a single problem, the enterprise loses control over its operational costs. The efficiency gains promised by the platform are eroded by the algorithmic meter running in the background.
When the system breaks down—as it did with BMW and Quinn—the platform does not absorb the blow; the enterprise does. By liquidating the human capital with the emotional intelligence required for complex service recovery, companies lose the ability to patch algorithmic failures. Institutional knowledge cannot be stored in a vector database; it lives in the shared experience of the workforce. Stripping the company of this human buffer means every minor software glitch can escalate into a full-blown, national media crisis.
By Way of a Non-Conclusion
The political economy of AI and the expansion of the Silicon Kolkhoz expose traditional enterprises to risks they are fundamentally unprepared to identify or explain.
At a time when the maps of corporate economic power are being redrawn, the capitalist food chain is being turned on its head, leaving new political friction points in its wake along the algorithmic frontline.
It is time to team up with the consulting firms that do the heavy lifting to map, monitor, and compute corporate platformization risks.
Follow BomaliQ and adhere to the BomaliQ Corporate Platformization Risk Index!
About the Author & BomaliQ
This insight is authored by Mathieu Lajante, PhD, founder and principal scientist of BomaliQ Inc.
BomaliQ provides specialized strategic intelligence for the algorithmic frontline, helping corporate leaders navigate the behavioural and political frictions of corporate platformization.
Nature of Intelligence
The insights provided in this publication are based on the stress-testing of publicly available industry reports, market data, and proprietary analytical frameworks. This content is intended for informational and strategic signalling purposes only. While every effort is made to ensure the accuracy of the analysis, the algorithmic frontline is a volatile environment.
Limitation of Liability
The BomaliQ Silicon Kolkhoz does not constitute professional consulting advice, legal counsel, or a formal business diagnosis. Readers should not make critical strategic decisions based solely on this newsletter without a rigorous, organization-specific assessment. BomaliQ Inc. and Mathieu Lajante shall not be held liable for any business outcomes or losses resulting from the use of this general intelligence.