Learnings from the founder of Observe.ai

Doing sales as an engineer

Swapnil and I met when he was raising Series A. He had several term sheets when I reached out to him, but he gave me the time of the day since I knew their voice tech stack. I found Swapnil via their transcription provider back then, the startup Deepgram. I met Swapnil when he was just finishing founder sales and starting to build the sales team. Swapnil was an engineer just like myself, and in a couple of years, he led his startup from low single-digit to high double-digit in annual recurring revenue and impressive customer logos. 

Observe.AI’s customers use AI to gain insights from their interactions, enhance frontline team performance, and achieve outcomes based on the industry’s first contact center LLM. While it’s an AI-first company, Swapnil is not a sales-first founder. He is an engineer at heart, having graduated from IIT Delhi. So how did he do it? Does an engineer founder stop doing sales after hiring a GTM team? When should they step back? These are common questions among engineer founders.

"Every time I say this is the last year I'll be involved in sales, I end up staying involved," says Swapnil, CEO of Observe.AI, which now has 300 employees.

“I am not good at sales, I’ll just hire someone” - when it works and when it doesn't

In the early days of a startup, you don’t have a playbook, you don’t know the ideal customer profile, and you don’t have product marketing, so getting and expecting a CRO or VP of Sales to come in is very hard. As a founder, you will be involved in sales from the early days and continue to be, even after you find product-market fit. Swapnil says, “The good news is while my involvement in sales has remained the same, the reasons I’m getting involved have changed.” By series A, Swapnil and his cofounder did cold outreach on LinkedIn, did the demos after, etc. Swapnil had a GTM cofounder, and they did only founder sales by $2M ARR, no sales team. Then, they got an SDR to get meetings for the pipeline, act like an assistant for sales, and quarterback the process. Since then, the company has had several iterations of the go-to-market teams and hired a director to the VP level for partnership, sales, and an SDR with Series A money. They would do the first meeting to qualify, but Swapnil would still be involved in the product discussion and the use case but would not be involved in commercial and negotiation. He was involved in the middle of the sales process for the company's second phase. Then, for the third phase, they got the VP of Sales and CRO, and Swapnil’s role changed to that of executive sponsor. 

Selling outside the Bay Area: wine and dine, look professional 

People won’t sign a $200k deal by talking with someone in sales of a random startup. They need the executive connection. The founder has to do a lot of wine and dine. If you’re selling to Bay Area startups, maybe you’re okay, but outside of the Bay Area, people want to meet you to see if they can trust you and that you’re innovative, etc. A big part of selling is relationships, wine, and dining; customers want engagement, and only you can do that. You have to show up dressed up formally. Customers don’t expect you to look formal in the Bay Area, but if you want to sell to the rest of America, you must look professional.

Show up and make your board do the work.

As an exec sponsor, you up-level the conversation. Your sales rep can get to a director level or junior VP, but a senior VP on the customer’s side won’t want to talk to a sales rep; that’s how it works. They want to speak with a C-level person or the CEO. Nobody cares about titles in the Bay Area, but the rest of the world does. If the deals are getting bigger, the board is flying out to wine and dine customers to up-level the conversation, show commitment, and show the company will be around. Make your board members do work, especially for more significant deals. They can be impactful.

“Are you going to the Google conference in Vegas? My natural engineer self says no, but this customer pays me $1M, he’s a CIO, he wants to hang out, and I’ll fly for a day just to see him.” This is the reality of selling: You have to show up for your customers.

How to navigate procurement - give and get

Who loves working with procurement?! But if you get to a point when procurement is asking you for a significant discount, take a moment and appreciate it’s a good problem to have. You’ve reached the negotiation point. Procurement’s job is to beat you down, to get a discount, and come very cold on the call. The best thing to do is understand how much leverage you have, which comes from a business buyer. Procurement asks for 20% off, saying, “My board will be mad, etc. Can we do it for 10%”? All these functions come in isolation to attack you; you have to go to the business buyer, and they’ll work for you internally.

Another tactic is to give and get. If procurement says give me 20% off, you cannot say let me go back and check if you’re the decision maker. In negotiations, you always have to clear a layer; you can say, “I can try, but I need to go check with my CFO.” If you don’t have a co-founder, put it on the board or someone else above you: “Let me talk with the board and come back to you.” “Let me think about it. I think it could work, but can you get this deal done in April?” - give and get, negotiations are that vs. I beat you down, and you provide me with everything and don’t ask for anything in return.

Business happens by text, not email.

You have to have an excellent relationship with the buyer. Buyers who buy your software are putting their jobs on the line. You have to help them look good and get promoted so they want to have a relationship with you. If it’s not on text, it’s not a relationship. Business happens via text and quick calls, not via email. Do you have 5 min for a quick call? Procurement, business buyers, and users run all that in parallel - multi-thread.

How to speed up a deal to close faster

The typical timeline in mid-market sales right now is three months and 6 to 9 months for enterprise and strategic. Doing a proof of concept speeds up the sales cycle, also ROI modeling and value selling speed up the sales cycle. You have to understand the person you talk to, to whom they go for approval, understand the chain of command, and what those people care about. When Swapnil does POCs, they are very bound, not only from a time perspective but also from success criteria, which is what we want to achieve in the POC, and the buyers are okay with that. Limit the POC to 45 days, and communicate the success criteria to transform into a contract. Bring some fiction in the deal; if it’s free, you’ll waste time. 

When you talk price - it’s not in the first meeting

In the POC process, customers need to know what they sign up for, so high-level pricing is given. You don’t negotiate; you negotiate after that when they want to convert to a contract from POC. You share the pricing late. 

How much of the increase in deal size came from products having more capabilities or you getting better at sales? 

How do you get over the line without a massive discount if you miss some significant product components and the friction inside the company this creates? Sales and product friction can happen in series A and B because,  in the end, the products are just not there and are not ready. Swapnil and I spent some time in 2019 talking about the product and sales friction. Swapnil leans on the side of committing to features vs not. Observe.AI didn’t build customer-specific features, but he let the customers and their priorities dictate the roadmap. They once hacked a feature in Retool over the weekend for a big customer. Swapnil is known to push the team for customers; he is customer first. Customer commit and customer-facing functions are great ways to expedite the roadmap. To avoid the discount, one can say, “We don’t have this; we’ll commit to the contract and build it in this timeline.” This helped lessen the pressure on negotiation and discounting. In good faith, so and so will deliver this by this date, but if they run into issues, we’ll grant six months to resolve them. A strong CFO should have all the wording to get the protection the company needs. 

Are you trying to increase the ACV or do a DNA shift for your company?

Product maturity leads to being able to sell to bigger customers, so the ACV grows like that. You market to customers, up-level the company, and make the hard call not to sell to SMBs. What many founders don’t realize when they move upmarket is a shift in DNA. It’s not just increasing the deal size; your entire company needs to learn to work with larger customers vs. small ones; it starts at the top: it’s how do I now market to larger customers, how to close larger customers, how do you build for larger customers, etc.? To build that team, rewrite the processes, and decide goals. It’s not about what tactic you can use to increase the price; it’s about changing your DNA. It’s telling customers if they can’t pay our deal sizes, we need to move on. 

Doing partnerships before product market fit - generally a terrible idea

Even you can’t sell, now you expect someone else to sell? A better approach is for founders to sell first. You build the playbook, you build repeatability, you know how to deliver, etc., and then you go to a partner. Observe.AI did it early on because it was a referral motion, and that’s easier since the company owns and runs the sales cycle.

Go and sit in your customer's office when you do strategic sales.

When selling to a large company, map all the people involved in the sales cycle: who they report to, their incentives, and what they need to test in the product. It’s a strategic sales cycle. Talk with as many people as possible at BigCorp, even if they aren’t in the sales cycle. The only way to discover what’s happening in the account is through tremendous customer discovery questions. Hey, I’m near your office. Do you have 10 minutes? I want to see if we can help. Watch the movie Air, put in the landmines. Watch Air again; that’s who you want to be. Nike used to be 3rd. The thing they did is your sales cycle. He does all the crazy stuff to get Jordan to work with Nike. Have people sit in the customer's office to get some intel. If you are not doing this, someone else is doing this right now: your competition, someone who has asked the board to reach out, someone who has a relationship from the past, or someone who is building a new champion internally.

As an investor in primarily engineer founders, it is one of my greatest joys to see the founders I get to work with go from trying to defer the uncomfortable sales to enjoying and talking with their customers and extracting the value they worked so hard to create in the first place. If you are hesitating to start a company because you think you’re not good at sales, don’t. Nobody is born good at sales. You can learn it. If you want to chat more about this topic or pick my brain about starting a company, please reach out. I love talking with people before they are ready to raise and contribute when they need that extra hand. 

I am grateful to Swapnil for being the first founder I saw progress from engineer founder to successful product seller, and for the many lessons he shared with me over the years. Observe.AI is a category leader in the contact center space and has a world-class team. I am humbled to be part of the journey and cheer on. 

Oana