Yogesh Sharma is the founder and CEO of Geeks2Connect, which helps businesses grow through software development, AI solutions, cloud technology, Salesforce expertise and technology talent services. He talks with Laurent Cochet about building the company, the sales and marketing problems he sees in every industry, and Celestra, the AI sales and marketing product he chose to launch on the show.
From WordPress sites to AI agents
Sharma is a computer science engineer who says he was always keen to learn new things in IT. He started his company in 2016 or 2017, beginning with small projects such as WordPress websites. Today the team works on AI agents and Salesforce, and has now developed its first product of its own to solve a problem it sees in the market.
The early challenges were marketing the brand and winning sales, and above all, he says, capturing the right market and the right clients. The company gradually won clients in India and abroad by understanding the basic problems most companies face.
Geeks2Connect starts from a simple observation: many companies struggle to build their own product, rely on partners and are sometimes unhappy with the result. His team sets up a process in which clients can follow progress and check that the right product is being built. Clients range from startups to large enterprises, and most now want to know how AI could help them without knowing exactly which problem it should solve.
Why sales and marketing break down
Sharma sees the same problem in every kind of business: sales and marketing. Companies pay sales teams and agencies but still do not get the leads, partly because they cannot monitor whether the work targets the right people. A prospect who emails in the middle of the night expects a quick reply, and people cannot be available around the clock.
Marketing has its own version of the problem. Teams use many tools to generate content, videos and images without knowing whether that content fits their product, and every three or four months, he says, the platforms change their algorithms. Follow-up is the other gap: it is very hard to remember data on thousands of contacts and keep following up with them for years.
Selling internationally adds the question of trust. In his experience, a prospect rarely converts after one or two calls or a few emails; you need to stay in touch for at least three or four months. Client expectations have shifted too: they once wanted a product to enter data into, and now expect AI to remember everything from the first day, produce analytics and remind them when to reach out again.
Celestra, an automated sales and marketing team
Celestra automates the sales routine Sharma describes: gathering prospect data, sending bulk or personalized emails and following up. Users pull contact data from the platform or through a LinkedIn extension, for example to target healthcare companies in California, and with two or three clicks Celestra builds a sequence of 10 to 20 emails. It checks whether each email was opened and sends a different next message accordingly.
After two or three emails, an AI agent starts calling the prospect to ask whether they have seen the message. For marketing, users describe their product in a few steps and the platform generates posts, images and videos, publishes daily and can run Google and Meta ads targeted by region, gender and age group. If an ad underperforms, it reports the analytics and proposes a new one.
Incoming leads get an automatic call or email, and a WhatsApp chatbot and AI calling agent answer customers at any hour. Sharma says users need to give Celestra only 30 minutes a month. He built it first for IT companies and says it also suits real estate businesses and companies selling beyond their own region. His first-year goal is to help teams that cannot afford 15 or 20 salespeople, with the product sold on a subscription basis.
Competing on understanding and delivery
Making the brand visible is the hardest part of scaling, Sharma says, when 10 other companies offer the same thing, often at a lower price. His answer is to understand the client's problem first, including how long the work will take and which engineers it needs, before proposing a solution. Other agencies, in his view, just want the project.
Delivery is his other edge. Because markets change quickly, he aims to deliver what the client wants within the agreed timeline, and credits customer loyalty to keeping promises. AI has sped this up: his developers now work more like testers who prompt the AI, and code that once took a month or two takes a day or two. Clients who once waited six or eight months now want a product in a month, and he says his team delivers in 20 to 25 days.
AI skills and advice for founders
Sharma rejects the idea that AI is taking jobs. Each period has had its key skills, from HTML and CSS to JavaScript and then Python, Salesforce and DevOps; now the skill is understanding how to use AI. People lose jobs, he argues, when they stop learning. Even Celestra needs someone who understands it and checks that it works properly, and people who can write good prompts will be in demand.
His advice to founders and operators is to keep pace with technology without rushing. First understand the challenges and pain points of customers in the industry you want to enter, which need not be IT. Everyone wants AI, he says, but few know which AI fits their service or product and how it would help the business.
The habit behind his success is to dream about the impact you want to make and then work toward it. He would tell his younger self not to trust too quickly or make fast decisions, and his guiding mindset is to stay positive, not regret past choices and learn from mistakes without repeating them.
Watch the full conversation
Watch the full conversation above. Yogesh Sharma says his details are on LinkedIn and on the Geeks2Connect website.
AI InterConnect hosts conversations with founders, investors and operators building with AI. Explore more episodes on the AI InterConnect podcast page.