Startup Founders: Build an Automatic Switchboard, Not More Tools for Switchboard Operators

Amir Shevat
6 min readSep 11, 2023

The automatic switchboard, a monumental invention in telecommunications, was introduced by Almon Strowger in 1891. Strowger was an undertaker by profession. The story goes that he suspected his calls were being intentionally redirected by a human switchboard operator to his competitor, who happened to be the operator’s relative. Spurred by business exigency and a dash of annoyance, Strowger patented the first automatic telephone exchange. This invention didn’t just solve his own problem — it revolutionized telecommunication by removing the need for human operators to connect calls, thereby speeding up the process and minimizing the potential for errors or, in Strowger’s case, sabotage.

Prior to the invention of the automatic switchboard, switchboard operator was a very popular role. In the early 20th century, the role of the switchboard operator was not only crucial but also widespread. According to the U.S. Census Bureau, in 1910 there were approximately 18,000 male and 36,000 female telephone operators in the United States. By 1920, the number had ballooned to over 130,000 operators.

I’m sharing this story because I see too many startups and founders focusing on developing tools for roles that should no longer exist. This strategy is doomed to fail, as the customer base for such products will eventually disappear. On the other hand, many founders don’t aim high enough to create technologies like automatic switchboards that could significantly improve our lives.

For instance, consider the role of first-tier Customer Service Representatives. My first job in tech was as a first-tier Customer Service Representative for an internet service provider. 75% percent of my calls involved password resets, another 15% were resolved by guiding the customer to restart their computer or router, or by verifying router configurations. The remaining 10% of calls were escalated to tier-two support. When I advanced to tier-two support, the calls became more complex and interesting.

First-tier support is a role that should be obsolete, not merely because it’s monotonous, but because AI can perform it more effectively. Unlike humans, AI doesn’t tire from repetitive tasks, doesn’t get frustrated with customers, and doesn’t bring emotional baggage to the job. It also avoids common human issues we suffer from today, like poor phone connections, hard to understand accents, and background noise. Over time, AI will become far more cost-effective than human Customer Service Representatives. It can stay on the line indefinitely to resolve issues, likely doing so more quickly and efficiently than a human in the same role.

The problem of the Minimum Viable Product (MVP)

The Minimum Viable Product (MVP) is an entrepreneurial approach aimed at delivering the minimum set of features that still offers value to customers, allowing startups to learn quickly. While effective in some cases, the MVP model struggles in scenarios requiring extensive research and development for a product to be truly valuable. Sam Altman, for example, conceded that developing OpenAI took an approach diametrically opposite to an MVP. It required a lengthy, resource-intensive effort that only yielded significant value after exhaustive research and development.

Entrepreneurs aiming to address issues in the Customer Service sector may be inclined to develop an MVP that enhances the efficiency of first-tier Customer Service Representatives, such as a template system for customer interactions. Some will say this is a step towards replacing these roles, but I would explain later in this article why that is probably wrong.

Founder should think big, way beyond the MVP. This does not mean that you should spend the next 2 years building a solution and testing it only after you finish, like we used to before the days of the MVP — but it does mean that your strategy (to replace Customer Service Representatives) should lead your tactics. Start with where you want to be in 5 years and work backwards from there.

The problem of picking the wrong Ideal Customer Persona (ICP)

The problem with starting by thinking small about your solution, is that founders often choose the wrong Ideal Customer Persona (ICP). An ICP is essentially a blueprint of your ideal customer and is crucial because it shapes the product, as well as the marketing and sales strategies. Making the right choice in this regard is vital for setting your company on a successful trajectory.

If entrepreneurs focus on Customer Service Representatives as their Ideal Customer Persona (ICP), they may create incremental solutions like better time-tracking systems, online policy training, or social portals for customer service teams. However, these are short-sighted if the endgame is to replace human reps with technology. Moreover, changing your ICP is a complex and painful process, requiring new hiring priorities and skill sets, potentially resulting in a pivot for the company. Ask any founder who’s had to change their ICP — it’s a big pain.

Founders should align their Ideal Customer Persona (ICP) with their long-term strategy. In the case of aiming to revolutionize customer service, the target ICP could be higher up the chain — perhaps the Head of Customer Success, the Chief Operations Officer, or even the Chief Revenue Officer. This approach is more in line with a strategic vision that goes beyond incremental improvements. Which brings us to the third problem.

The problem of Product Led Growth (PLG)

Product-led growth (PLG) is a popular business strategy, amongst startups, that uses a company’s product to drive acquisition, retention, and expansion. Companies like Slack, where I worked at in its early stages, successfully used PLG to grow to become very big companies. This approach works well when users can easily onboard themselves, quickly recognize the value of the product, and transition from free to paid users — all while potentially evangelizing the product to new users. Users of PLG products tend to self-serve themselves onto the product, rather than having sales and customer success representatives convince them to join and onboard the service (usually called product-led, as opposed to motion sales-led motion).

The problem is that it’s extremely difficult to implement PLG for products like automatic switchboards. Products requiring significant organizational change, extensive integration, and a high-ranking ICP tend to perform better with a sales-led motion rather than a product-led approach. While it’s possible to change your go-to-market strategy, doing so is not easy.

Founders should choose a go-to-market strategy that aligns with their long-term vision. Just because a particular approach is popular doesn’t mean it’s the right fit for you. Additionally, it may be possible to make a complex product work within a product-led go-to-market framework.

The problem of Ethics

Founders of startups often strive to change the world for the better. Entrepreneurship is a hard, long process, and requires a lot of grit. Without a strong conviction in your mission, you may lack the perseverance needed for success. So, is building job disrupting technologies a good or a bad thing?

The honest answer is that it depends. Technological disruptions have a dual nature: they can displace workers in the short term but often lead to net-positive outcomes in the long run. While some individuals may experience job loss or find it difficult to secure employment in certain sectors, the broader trajectory of technological progress often yields benefits. For example, the invention of trains accelerated economic growth, washing machines liberated people from tedious labor, and automatic switchboards revolutionized communication. Research indicates that, on balance, technological innovations generate more jobs than they eliminate.

The challenge, therefore, is societal. When industries are disrupted, there is an ethical obligation to mitigate the negative impacts on affected individuals, even as we strive to maximize the technology’s benefits for society at large. In my view, progress in a global economy is inevitable, but its implementation can — and should — be managed responsibly.


Disruptive technologies change things mid-game, and most entrepreneurs (and to a larger extent most of the tech industry) have yet to grasp the order of magnitude of change this technology can create. We’re still thinking in terms of faster horses in the early days of the automobile, still thinking of desktop apps in the early days of the internet, still building tools for switchboard operators.

Those who fail to adapt will be forgotten by history, while those who understand this revolution will build companies that will be remembered forever.



Amir Shevat

Investor in early stage startups. Previously: Head of Product, Twitter Dev Platform, VP product at Twitch, Slack, Google, Microsoft. Author at O'Reilly.