Sales reps spend most of their time not selling but caught up in data and admin. RevOps startup Scalestack, powered by MongoDB Atlas on AWS, is seeking to solve the problem with AI.
If you are asked what a sales rep does, you might think the answer is simple: sell. The rather alarming reality, however, is that sales reps spend only 28% of their time actually selling, according to a Salesforce study of more than 7,700 sales professionals from around the world.
The rest of the time, the research found, sales reps are “bogged down” in non-selling tasks such as researching prospects, data input and other administration, which holds them back from their principal function and results in just three in ten professionals expecting to hit their sales quotas.
Co-founder and CEO of Scalestack, Elio Narciso, has observed the sales productivity gap first hand during his career working in business development within both small and large companies.
Having founded and exited other companies which utilised automation to solve business challenges, he spotted an opportunity.
“Seeing just how inefficient the job of a salesperson is was a stunning realisation for me,” Narciso says. “I worked in a lot of start-ups and of course when you start a company, not everything is perfect.
“But seeing up close in even larger companies that salespeople spend most of their time trying to research and make sense of data about prospects made me realise there is a tremendous opportunity to improve productivity in the sales and marketing space.”
Scalestack: Bringing order to sales chaos
In 2021, Narciso set out to solve the problem with co-founder and former colleague Alex Prioni, by using the power of AI to build a data orchestration and activation platform which helps companies manage structured and unstructured go-to-market data. RevOps startup Scalestack was born.
With a mission to bring order to the often chaotic blizzard of go-to-market data which sales professionals must contend with, Scalestack automates a lot of the research that sales and marketing professionals need to do in order to better understand their targets. It serves to produce clean, orchestrated and rich data everywhere sales or marketing people live, from the CRM platform to the tools for outbound emails to third-party sources of sales data like LinkedIn.
“Through the Scalestack platform we give information that is constantly up to date, clean and rich, so you always know who to target, when and why,” Narciso says. “You know what you should know about targets in order to be better at presenting whatever product you want to sell.
“We serve larger sales and marketing teams with sizable outbound motion. If you have a wide addressable market, you sell your product or service to many, you need to rely on fast access to accurate data to become better at your go-to market. We excel at making sense of a lot of data.”
Best in class
To live up to this promise, Scalestack needed a formidable technical backbone able to deal with an enormous amount of data. The foundation of its tech stack is threefold: AWS for computing infrastructure and storage, MongoDB Atlas for the underpinning database (with Atlas Vector Search for its powerful query capabilities), and the use of LLMs like OpenAI and Anthropic.
Scalestack utilises LLMs in an “agentic” infrastructure which Narciso credits with making the platform better and more customisable. Its AI agents can very efficiently scan the web and bring back valuable data, properly categorised and enriched for whatever use case a customer has.
Meanwhile the decision to use MongoDB Atlas on AWS was an “obvious choice” inspired by Narciso’s previous role managing the AWS Global Startup Program, which powers the go-to-market of fast-growing enterprise tech startups. Time after time he saw customers successfully implementing MongoDB as the best in class database solution for large datasets.
“We were already familiar with both AWS and MongoDB when we started the company, and when we faced the first version of the product, and we realised we’d have to manage hundreds of millions of data points for our customers, it was an obvious decision to choose MongoDB.
“While there are other solutions that are native to the cloud, we wanted to go with the best in class solution for that component and that's why we chose MongoDB. And that’s also why we integrated tightly with AWS, because they have such a tight relationship. MongoDB has since become a customer of Scalestack, although the decision processes were completely separate!”
Mission accepted
To generate the most accurate responses from the LLMs, a key technology for Scalestack’s solution is retrieval-augmented generation (RAG), which augments LLMs with additional data to generate the best insights. Rather than use a standalone or bolt-on vector database, Scalestack has been able to build its AI agents, or bots, on the retrieval capabilities of Atlas Vector Search.
The bots that power Scalestack are programmed to complete “missions” for customers. For example, a mission could be to find everything it can about a company, such as how it has grown its revenues and employees over time and what kinds of customers it has, and then not only bring back the data but outline the reasoning, the steps it took and the data sources it used.
Other use cases for the bots include synthesising data in a document, or simply cleaning it up. It might seem trivial but data needs to be clean because when fed into AI the results are amplified.
One of the biggest problems in sales and marketing data is when salespeople inadvertently create duplicate data, for example one inputs a company in the CRM as “AWS” and another as “Amazon Web Services”. AI helps identify these kinds of events and also creates relationships between data, recognising for instance that one company is a subsidiary of a parent company.
“You and I can search the web, but there is no comparison to an AI agent browsing and searching the web,” Narciso adds. “The ability to search data at massive scale and speed and bring back the relevant answers is incredible. We can now enrich 500,000 company objects.
“Imagine 500,000 companies in your CRM need to be enriched with 50 data points that you really care about. Consider the number of reps and territories you have to cover – we’re talking hundreds of millions of data points. We do this enrichment now in a matter of a couple of hours, when it used to be a month or more, and at a much higher accuracy because Atlas Vector Search, with its search and RAG capabilities, helps us do that at an incredible scale and speed.”
Delivered in partnership with MongoDB
