Pivoting a simple messenger into a $50M ARR customer service automation platform

Rashid Khan, built a simple messenger to prove a hypothesis. Then, with some initial traction, he pivoted, doubling down on B2B and AI.

Now, Yellow.ai is making $50M/yr. Here's Rashid on how he did it. 👇

Before the pivot

In 2015, as a fresh college graduate in Bangalore, I noticed something fascinating: People in India and other emerging markets weren't just using WhatsApp — they were living their digital lives through it.

This observation sparked a compelling question: If consumers were spending most of their digital time in messaging apps, wouldn't businesses eventually need to meet them there?

To validate our hypothesis about messaging-based business communication, we decided to run an experiment that would later prove transformative. We launched a consumer app called Yellow Messenger that allowed people to discover and chat with local businesses using geolocation. Within months, we had over 500,000 downloads and 50,000 monthly active users generating thousands of chats daily. The most remarkable part? All these conversations were being manually handled by just two people on our team.

This early traction taught us something crucial about the market: Consumers were more than ready to interact with businesses through chat interfaces, but manually handling these conversations wouldn't scale. This realization led us to develop one of the first production-scale chatbots with natural language processing capabilities in 2016, well before WhatsApp officially opened its platform to businesses in 2019.

Pivoting to B2B

The timing and location worked perfectly in our favor. Bangalore, often called India's Silicon Valley, provided an ideal launching pad with its low cost of living and rich tech ecosystem. Being fresh out of college was actually an advantage – I could afford to be scrappy, living on minimal salary for years while we built the foundation of what would become Yellow.ai.

Our journey from a simple consumer app to a comprehensive AI platform reflects a broader truth about entrepreneurship: Sometimes the best businesses don't come from trying to create something entirely new, but from carefully observing how people's behaviors are changing and building tools to support these evolving patterns.

While WhatsApp didn't open its business platform until 2019, we had already built a robust chat automation system that worked across websites and mobile apps, positioning us perfectly for the messaging revolution that would follow.

What started as Yellow Messenger in 2016 has grown into a company generating $40M-$50M in annual recurring revenue, helping businesses worldwide transform their customer service through AI-powered voice and chat agents.

Every day, our AI agents handle countless conversations across multiple languages and channels, learning and improving with each interaction. It's fascinating to see how a small coding project has evolved into solutions that are reshaping global customer service standards.

Consumption-based model

At Yellow.ai, we operate on a consumption-based business model where enterprises pay for each conversation handled through our platform, whether it's through chat or voice AI agents.

This model has proven highly scalable and aligns perfectly with our customers' success – as they drive more automated conversations through our platform, we grow together. Customers typically commit to an annual volume of conversations, which provides predictability for both parties while maintaining flexibility for growth.

Our revenue journey began in 2016 when we secured our first B2B customer. Unlike many startups that delay monetization, we started charging customers almost immediately after validating market demand. We initiated with modest pricing to gain market traction, but our strategy was always clear – start small and expand through both volume growth and feature adoption.

Equity-free funding

The financial aspect of building our initial product presented another significant challenge, but we found an innovative solution through equity-free accelerators. Instead of immediately seeking venture capital, we strategically participated in three key programs:

Startup Chile provided us with $50,000 in equity-free funding and valuable mentorship in Santiago. This early support gave us a runway to experiment with our product without diluting ownership.

Microsoft Accelerator became instrumental in our B2B pivot, offering not just office space in Bengaluru but also $500,000 in Azure credits. These credits essentially eliminated our cloud infrastructure costs for the first three years, allowing us to invest more in product development.

The SAP Accelerator program gave us crucial access to enterprise customers through their GTM team, helping us understand enterprise needs and refine our product accordingly.

This strategic use of accelerators allowed us to operate for nearly four years without traditional office overhead, as we didn't establish our own office until late 2019. More importantly, it provided us with a powerful go-to-market motion through integration with these large organizations' sales stacks.

Yellow's tech stack

At Yellow, we have always leveraged the latest and greatest tech for building our products. On the front-end, we are heavy users of React and quite recently have started leveraging a lot of NextJS for internal tooling.

The backend at Yellow is actually quite varied and has become technically complex over the years due to the number of customers whom we handle, as well as the types of applications we have to build:

• We run and host our own messaging infrastructure for consumer-to-AI-agent interaction and even consumer-to-human-agent interaction. The messaging infrastructure is built on top of XMPP framework on Ejabberd (Erlang).

• The core workflow engine for business logic for each customer runs in a virtual environment created on top of NodeJS where we have built out our own executor or run time.

• The core micro-services which handle authentication, session management, business logic management, agent management are written in combination of Java and Go.

• On the database front, we started with MongoDB, then over a period of time, we started to leverage Postgres for some internal workloads.

• AI Agents and AI Stack is built on top of Python.

• We have architected an analytics engine on top of a time series database Druid.

• The platform is cloud agnostic and we run seven availability sites across the globe over AWS, Azure and GCP. The platform gets deployed on Kubernetes and has continuous integration and continuous delivery of code.

The single most important success factor

Building Yellow.ai has taught me that the single most critical factor in a company's success lies in "people decisions" — specifically who you bring onboard, how you develop them, and how you integrate new leadership. Let me share our biggest challenges and the valuable lessons we learned along the way.

The most significant challenge we faced was making the right hiring decisions during different growth stages. At Yellow.ai, some of our biggest setbacks came from hiring the wrong leadership talent, particularly when we were scaling rapidly.

We learned that bringing in senior leaders without proper context and integration can be more damaging than having gaps in the leadership team. For example, we once hired a senior executive who made sweeping changes to our sales strategy within their first month – changes that took us nearly a year to course-correct because they didn't align with our core market understanding and customer needs.

This experience taught us our first major lesson: The immense value of developing internal talent. We discovered that team members who joined us early in our journey, when we were still a small company, often became our most effective leaders. They had a deep contextual understanding of our product, market, and culture that external hires typically take 6-9 months to acquire.

However, this doesn't mean we completely avoid external leadership hires. There are times when bringing in outside expertise is crucial, particularly when entering new markets or developing capabilities that don't exist internally. The key is in how you integrate these leaders into the organization.

Our second major lesson was developing a structured onboarding process for senior hires. We now implement a 90-day integration period where new leaders focus on learning and understanding before making major strategic decisions. This approach has significantly improved our success rate with external hires.

Growing to $50M/yr

Attracting users and growing Yellow.ai has been deeply rooted in building strong partnerships and creating product-led growth mechanisms.

Partnerships

In the early days, one of our most successful strategies was partnering with Microsoft Azure. This partnership wasn't just about using their cloud services – we deeply integrated with their marketplace and actively participated in their ecosystem. We became early adopters of their experimental features, which gave us both technical advantages and visibility within their network.

This partnership opened doors to enterprise customers who were already using Azure, significantly reducing our sales cycles and providing us with credibility in the enterprise space.

Product-led growth

Product-led growth has been another cornerstone of our strategy. Being an end-consumer facing product, we made a strategic decision to include our branding within the chat interfaces we powered.

This created a powerful network effect as more businesses adopted Yellow.ai, more end-users were exposed to our technology, leading to organic inquiries from other businesses. This approach helped us grow from handling 100,000 conversations per month in 2017 to over 150 million monthly conversations by 2019. And we're now processing more than 1.5 billion conversations monthly across our platform.

Sales

Our sales approach evolved significantly over time and varied by region. In the early stages, particularly in India and Southeast Asia, we focused heavily on direct sales through cold calling and in-person meetings at industry events. We made an unconventional, but effective, choice to hire ambitious early-career professionals for sales roles, pairing them with technical solution consultants. This combination of hungry sales talent and deep technical expertise proved crucial in converting complex enterprise deals. Our customer base grew from 50 enterprise customers in 2018 to over 1000 by 2024, with our average contract value increasing from $15,000 to $150,000 during this period in the enterprise segment.

Thought leadership

We've also leveraged thought leadership and PR extensively, particularly during technological shifts in our industry. When new technologies like voice assistants or generative AI emerged, we positioned our leadership team as industry experts through media appearances and thought leadership content.

This approach helped us secure coverage in major technology publications and business newspapers without spending on paid PR. The strategy particularly paid off during the COVID-19 pandemic when businesses were rapidly digitizing their customer service operations. Our inbound leads grew by 300% during this period as a result of our established thought leadership position.

Focus on distribution

For aspiring entrepreneurs, my key recommendation would be to focus on building distribution advantages early. Whether it's through strategic partnerships, product-led growth mechanisms, or thought leadership, having strong distribution channels is often more crucial than having the perfect product.

And always align your growth strategies with market waves – our biggest growth spurts came when we correctly anticipated and rode major technological shifts in our industry.

Accessing an existing market

Ignoring the market trends is probably one of the fastest ways to die in the startup world.

It is super hard to create a market; it is much easier to surf a large wave. For Yellow, there have been few major market trends which have pushed our growth to the next level. Key trends where we had tremendous growth:

• Meta/Facebook announcing M (Messenger Platform)

• Launch of Amazon Alexa, Microsoft Cortana & Google Assistant

• Launch of WhatsApp API

• Adoption of ChatGPT/Generative AI

As Steve Jobs said: “You can't connect the dots looking forward; you can only connect them looking backwards. So you have to trust that the dots will somehow connect in your future.”

The key aspect with waves/market trends is to keep an eye on what is changing around the world and keep experimenting, the dot will automatically connect in the future, your job is to keep taking the bets and experiment with what is out there.

Revenue growth

Several factors have contributed to our consistent revenue growth:

1. Our strategic shift to focus on enterprise customers has significantly increased our average contract value.

2. The advent of Generative AI has allowed us to introduce more sophisticated capabilities, enabling us to adjust our pricing to reflect the enhanced value delivery.

3. We've successfully introduced new products and moved towards more outcome-based pricing structures, directly tying our success to our customers' results. This has not only improved our revenue but also strengthened customer relationships and reduced churn.

One of our key learnings in growing revenue has been the importance of keeping price simple, while maintaining a laser focus on sales metrics. We've found that complex pricing structures often become barriers to purchase decisions

For entrepreneurs building similar businesses, I'd emphasize the importance of having a clear monetization strategy from day one. While it's tempting to focus solely on growth and worry about revenue later, we've found that early monetization provides valuable feedback about product-market fit and helps build a sustainable business.

Additionally, invest time in understanding your unit economics and customer lifetime value early – this knowledge will be crucial as you scale and need to make strategic decisions about product development and market expansion.

Parting advice

Trust your instincts

An essential principle that has served me well is trusting my instincts, particularly when evaluating opportunities or making critical decisions. In the startup ecosystem, you'll frequently encounter situations that appear too good to be true – whether it's a potential partnership, an investment opportunity, or a promising hire. I've learned that these initial gut feelings often prove accurate, and taking the time to carefully evaluate decisions rather than rushing into seemingly attractive opportunities has prevented numerous potential missteps.

No silver bullet

A critical realization that came to me over time is that there are no silver bullets in building a successful business. I frequently see founders searching for that one perfect hire or magical solution that will solve all their growth or product challenges.

The reality is that sustainable success requires founders to deeply invest their own time and energy in understanding and solving core business challenges. No single person or solution can replace the founder's role in driving the company's vision and execution.

Financial discipline

Perhaps the most practical advice I can offer is about financial discipline. Regardless of how well your business is performing or how much funding you've raised, maintaining low costs and focusing on profitability should be constant priorities.

This has helped us navigate through various market cycles and maintain our ability to make independent strategic decisions.

Resources

As the famous saying goes “People who don’t learn from history, are doomed to repeat it..”

I'm a big consumer of long-form content to learn about everything, from picking up new programming paradigms to company building. In no order, listing down the key resources which have helped shape my thought process:

• The Hard Things About Hard Things – Ben Horowitz

• The Great Mental Models – Farnam Street

• Rework – Jason Fried, David Heinemeier Hansson

• Paul Graham’s Essays

• Masters of Scale – Reid Hoffman (Podcast)

• Indie Hackers

What's next?

I'm incredibly excited about the opportunities that lie ahead both for Yellow.ai and personally. On the company front, our primary focus is on pushing the boundaries of what's possible with Generative AI. We're working on developing more specialized, vertical-specific products built on top of the Yellow Platform.

The goal is to offer more than a generic AI platform, and create deeply integrated, industry-specific tools that solve unique challenges in sectors like healthcare, banking, and retail. This specialization will help us maintain our competitive edge in an increasingly crowded market.

One of our key initiatives for growth is investing heavily in content and community building. We're particularly excited about expanding into video content, which we see as an underutilized medium in the B2B space.

On a personal level, I'm working on finding better balance. After years of intense focus on building Yellow.ai, I'm making a conscious effort to prioritize my health and well-being. This includes dedicated time for motorcycle rides, which I find incredibly therapeutic, and plans to explore more countries – both for personal enrichment and to understand different markets better.

You can follow along on X, and LinkedIn. And check out Yellow.