Contra the Heard: Investing in Cinareo
I don’t recall how it actually happened, but I do remember it was not obvious at first.
Did you ever read the book “That Will Never Work”?
Yeah, it was kind of like that.
It sounded like the kind of thing that only makes sense if you had just read a Howard Marks’ "Dare to Be Great" memo. In the memo, Marks lays out a simple yet profound insight, the road to above-average returns - by necessity - runs through unconventionality, or said differently, you can't take the same actions as everyone else and expect to outperform.
A different type of bet
I’m writing this post in early 2025. In just the past week, OpenAI released Operator and DeepSeek’s R1 has shown that one does not need hundreds of millions of dollars to build an o1-like model — Needless to say, things are moving fast.
One of the verticals in which AI adoption is moving the fastest is that of customer support and contact centers. These hubs of customer interaction are showcasing the power that multiple AI models can have in streamlining operations, to the point of completely replacing humans.
At first glance, Cinareo’s mission to integrate AI in contact centers might seem counterintuitive. Why invest in a company that helps contact centers figure out which humans they still need when AI is rapidly transforming the landscape?
Here’s the thing: I don’t think AI will replace all humans. In fact, I believe the opposite. AI will empower humans—giving them superpowers to accomplish things previously never thought possible or imagined.
The thesis behind Cinareo is that its AI-driven capacity planning software will act as a bridge in helping companies transition from human-centric SaaS tools into agentic solutions that will make humans more productive over time.
AI will reshape contact centers and yes, some roles will be displaced, but most humans will stay and become more productive than even thanks to these tools. Can you imagine a world where calling a 1-800 number is actually pleasant? I (and my team at the Seed VF) invested in Cinareo because it solves a critical challenge across many industries, starting with in contact centers: capacity planning in an era of AI-human collaboration.
To that end, let’s dive deeper into the world of contact centers, how they operate, and the tech stack that powers them. Starting with some history…
The Death of the Contact Center Has Been Greatly Exaggerated
Predictions about the demise of contact centers have been around for some time. In a 2019 Bloomberg featurette titled "Why Call Center Jobs Will Disappear," journalists examined the factors driving call center automation.
At that time, ChatGPT and other advanced AI models were still in their infancy; back-then, “if this, then that” chatbots were the new hot commodity. Before chatbots, email was expected to make call center jobs obsolete, and before that, automated phone systems were seen as the future. And yet, none of these technologies only made contact centers more efficient and attractive to operate.
So, the question becomes, will this time be different?
The AI Revolution in Contact Centers
Contact centers (also known as call centers) assist customers from various industries with a wide range of issues across multiple communication channels, including phone, email, SMS, chat, and more recently, autonomous chatbots and voicebots.
The historical expansion of communication channels (queue) has added complexity to these centers as companies aim to enhance customer satisfaction and reduce churn by improving their customer experience (CX). The introduction of AI has only made this more complex as we are now at a stage where the technology behind non-human channels, like chatbots and voicebots, offer an experience that is often indistinguishable from – and sometimes better – than that of a human.
This is playing out in real-time across industries. For example, recent statistics from fintech giant Klarna highlight the potential impact AI could have on customer service jobs overall.
And while these stats are shocking, what is critical to understand is that an internal support center like Klarna’s is significantly different from that of most contact centers. Not understanding the difference between the two is akin to thinking about forward and reverse logistics as being fungible, which would be silly at best.
Internal support centers often handle queries with low variability and high repeat rates. Contact centers in contrast, deal with more complex issues due to the diverse customer base they serve, the specific service level agreements (SLAs) they must adhere to, and the various software products they use to provide exceptional CX.
For AI-powered solutions (AI-agents) to replace human agents in contact centers, they must not only excel in customer interactions, but also be trusted to effectively interact with the vast number of other software systems that power these contact centers, many of which do not have APIs and do not lend themselves to Computer Use/RPA-type solutions.
Understanding the Al-Agent Hierarchy
When thinking about agentic AI-products for contact centers (or any enterprise setting for that matter), five types of products/solutions generally emerge:
The Router: These products sit on top of traditional contact center software, configured with no code drag-and-drop UX to ingest historical conversational data, company knowledge base data and interaction rules/guidelines to decide where to route a customer's query within the contact center.
The Co-Pilot: These products work in tandem with human agents, providing suggestions an guidance during customer interactions while keeping the human in control of all "write" activities. These solutions aim to enable human agents to become more productive and provide better CX.
The Junior Al-agent: These products can "read" and execute light "write" tasks such as changing an address, booking a technician, processing automated payments, etc.
The Senior Al-agent: These products can theoretically perform any CRUD (create, read, update and delete) operation related to a customer's profile/account. The key here is for the Al-agent to seamlessly interact with the underlying company’s software systems like Salesforce, NetSuite, etc.
The Orchestrator: These products act as the central point of contact, coordinating and managing the interaction between various Al agents and human agents. The Orchestrator ensures that each query is handled by the most appropriate agent, whether human or Al, and optimizes workflows to enhance efficiency and customer satisfaction. Akin to a general contractor in a construction setting where they’re managing the project by balancing and coordinating budget, time, and subcontractors for any given task within the project.
Technologically, the Router, Co-Pilot, and Junior Al-agent are already here. However, for the Senior Al agent and Orchestrator to truly become adopted, the bridge that will need to be crossed is that of trust, especially as it relates to "write” permissions in a world where explainability and reliability of responses are still not guaranteed due to issues: like hallucinations and the non-deterministic nature of generative Al models, not to mention the lack of connectivity and interoperability between systems that often times don’t even have APIs.
The Thing About Technological Advances
A meaningful and important aspect of productivity gains that is under appreciated, is that in competitive markets, productivity gains are generally met with an equivalent or increased response from competitors. For example, let's imagine that Company A adopts an Al product that leads their sales reps to increase their closing rates by 50%. All things being equal, Company A would see this and double down into their sales efforts and hire more sales reps, which they can augment with this new Al product. At the same time, Company A's competitors will not rest on their laurels. Competitors will either hire more humans to keep up with Company A and/or procure that same Al product to equip their sales reps and compete — In either scenario, the edge gained by AI leads to more humans in the mix. Technological leaps make it cheaper for humans to leverage the technological advancement, and, as productivity becomes accessible, more of it is sought until the edge provided is competed away and a new equilibrium is reached.
Now, let's imagine a future in which contact centers have the option to be fully digital and Al-powered.
How would the average customer react when interacting with an Al-agent?
What happens when somewhere in the interaction the customer asks, "Are you human?" Would a human even care?
What happens when the “human” asking the question is not a human at all, but another AI agent acting on behalf of the human?
Technologically, we’re there. AI models and agents are able to functionally execute substantially all of the various roles/processes within the contact center stack; however, I believe human interactions will remain the preferred method of troubleshooting issues over the foreseeable future. What this means is that contact centers will adopt AI in a manner that allows them to make their humans more productive and competitive. AI will give these humans superpowers that will enable them to be much more productive and pleasant to deal with — Less wait times, less transfers between department, less unhappy customers.
Capacity Planning at Contact Centers Post Al-agents
Contact centers have been struggling for years to determine how to effectively plan and resource human agent and non-agent capacity. The complex exercise of figuring out staffing needs required to satisfy customer objectives has been relegated to manual spreadsheets embedded with home-grown formulas and a pinch of guesswork to estimate approximate best answers that are suboptimal, untraceable, and error-prone.
The addition of Al-agents simply represents another channel to manage. The importance of finding a way to integrate this new channel effectively into the contact center environment becomes an even taller mountain to climb as there are downstream effects both at a human and technological level. Moreover, adding Al agents to a mix of diverse communication channels has made capacity planning even more complex than it already is, which paves the way for Cinareo’s software to help bridging this transition.
Cinareo is reimagining capacity planning by replacing manual, error-prone spreadsheets with AI-powered automation. Its software not only forecasts staffing needs across human and AI agents but does so with unprecedented speed and accuracy. This enables contact centers to transition to AI-augmented operations seamlessly, setting a new standard for efficiency and customer experience.
Cinareo stands out because of its proprietary AI-driven software, which can produce enterprise-grade capacity plans in minutes—a task that typically takes humans days. Beyond its technical prowess ( multi-channel, multi-skilling, and explainability on service level and budgetary goals), Cinareo’s founders, Karen Elliott and Dr. Mark Alpern, bring a unique blend of industry expertise and vision, positioning the company to not only dominate the contact center space but also expand into other resource-constrained industries like healthcare and retail.
The best part is that Cinareo’s core technology, is that it has the potential to help other verticals plagued by resource constraints and an overwhelming number of variables to balance in the capacity planning process, including AI solutions.
Looking Beyond Contact Centers
Contact centers are not the only ones facing pressures to optimize labor and staffing needs. Other industries, such as retail, transportation, manufacturing, and healthcare, also experience similar challenges. For instance, the healthcare industry will use Al-powered capacity planning software to better manage human resources and physical assets, reducing long wait times in the ER. By predicting patient arrivals based on historical data and other metadata, similar to contact centers, Cinareo’s software would suggest the necessary staff and resources to support patients throughout their hospital journey. This includes intake administrators, triage nurses, ER nurses, physicians, ER beds, syringes, and more. The software can also estimate the average time needed to treat patients for various conditions and ensure desired service standards, such as having 80% of patients seen by a physician or nurse practitioner within 30 minutes. The same algorithms and workforce management strategies used in contact centers can be applied to hospital administration. AI-first solutions like Cinareo, are well-positioned to address planning issues across multiple sectors, improving both customer and employee experiences.
In Closing
While others are simply focused on AI solutions that eliminate human agents, I (and my team at the Seed VF) saw a different path: invest in AI that stands to augment human performance. Karen Elliott’s and Dr. Mark Alpern’s approach to capacity planning embodies this belief, positioning Cinareo uniquely in a rapidly evolving market.
This is grounded on the belief that Al integration in call centers marks the beginning of a new era of job transformation. Rather than eliminating human jobs, Al will enhance them, making interactions with contact centers as pleasant as seeking advice from a friend. Because of this, I believe Cinareo is on the upper right-hand quadrant of Howard Marks' matrix.
I’m bullish on Al.
I’m bullish on human productivity.
And most importantly, I’m bullish on the future where the two work together.