Deckmatch pioneers richer insights for private market intelligence

Remember the early days of the internet? Opening to the soundtrack of a dial-up modem, we’d head to a search engine, type out a query, and wait for results. These early tools were like directories, allowing for basic searches and displaying information in a straightforward, simplistic format.

The first search engines appeared in the early 1990s. Three decades later, while AI and algorithms have improved search engines in general, the process of gathering and analyzing information in many sector-specific virtual spaces remains outdated. “As advanced as we are in all things tech, that same directory structure that dictates how data gets transmitted across private markets is basically just a directory that you have to navigate,” says Leopold Gasteen, Co-Founder and CEO at Deckmatch. Deckmatch has a vision to transform this process and, working with AWS, it’s already well on its way to doing so.

The company set out with a very specific aim: to “build the world's foremost and most useful natural language search engine for private markets,” explains Gasteen. The result? AlphaLens, a market intelligence platform that maps the competitive landscape by analyzing product-level data.

Deckmatch partnered with AWS earlier this year, while it was experimenting with different AI models. It sought a solution that could index website information and help it pioneer a new approach to private market intelligence. As a startup, Deckmatch was working with limited resources and the AI models it was evaluating “just didn't do a great job relative to cost,” says Walid Mustapha, Co-Founder and CTO. To address this challenge, Deckmatch utilized AWS Activate Credits, designed to help startups bring their ideas to life by offsetting the cost of AWS solutions. And to realize its aim, Deckmatch turned to these solutions and expertise from the AWS team, leveraging Amazon Nova Pro and Amazon Bedrock to build and scale its generative AI-powered platform.

Natural language search for superhuman insights

Both Gasteen and Mustapha are repeat founders and, says Gasteen “passionate about startups and the venture space.” This space is a crowded one; “it’s never been easier nor more viable to start something yourself and it’s never been cheaper to get off the ground.” The thriving startup landscape generates significant opportunities for investors, however, discovering and analyzing these opportunities can be complex and cumbersome, yielding poor results.

“Essentially there are millions of new companies and millions of new products,” says Mustapha, and the databases that analysts use are typically full of company descriptions which are “quite glossy and marketing-oriented as opposed to being objective.” Lacking a product-centric approach and adequate search functionality, legacy tools were not delivering.

Enter AlphaLens. The solution offers product-level intelligence using semantic engine maps which understand specific products and their features, offering up rich descriptions and unpacking their key offerings. In doing so, AlphaLens frees up time and “empowers the analysts that are actually doing the work to get other work done right,” says Mustapha; work such as “deep thinking” and “evaluation.”

AlphaLens automates the discovery process, so instead of spending days or weeks identifying companies and their products, analysts can “do a couple of searches in a couple of minutes,” says Mustapha. While AI-enabled automation is a key element of AlphaLens, Deckmatch is keen to emphasize the technology’s ability to enhance rather than replace critical human skills. Deckmatch “doesn’t believe in a future where all of this is automated,” explains Mustapha. “We believe in a future where we enable people to make the best decisions possible, and that's about finding all the data that they care about, presenting it in a meaningful and useful way, and allowing them to make decisions based off that data.”

A model example of cost-efficiency

This is all possible due to the platform’s underlying architecture. Amazon Nova Pro generates the product descriptions from company websites which are so critical to the search process for Deckmatch’s customers. A newer tool, which the company is still developing, allows users to pick and compare products, as well as discovering insights such as funding rounds and company headcounts. By gathering and analyzing unstructured data—much of which isn’t available anywhere else online—and serving this up in a rich yet digestible format, AlphaLens is empowering investors to make more informed decisions.

Amazon Nova Pro also offered benefits to Deckmatch itself. When the company evaluated the solution against other models, differences in performance and accuracy were negligible, however, Amazon Nova Pro offered a significant advantage in a key area: cost.

“When you compare the cost, you can see that the input token cost and the output token cost is much higher for the frontier model than Nova Pro,” says Mustapha. It was also “much faster,” he continues, “so that made it an easy decision for us; with those metrics we could justify the use case.”

In addition to cost, Deckmatch also needed a solution that would support its growth. Today, AlphaLens indexes around 12 million products, but with the crowded startup scene ever-expanding and product launches increasing, scalability was key. “We need to refresh websites, revisit them in the future, ensure that our data is always up to date,” explains Mustapha.

Amazon Nova Pro is enabling this expansion, offering the latest foundation models and allowing Deckmatch to expand the scope of information it can serve its customers. To do so, “the workflow has to be as cost-efficient as possible,” says Mustapha. As such, Deckmatch is planning on migrating all its workloads onto AWS later this year, allowing the company to scale its solution and better serve customers as its business grows.

Also powering this growth is Deckmatch’s recent listing on AWS Marketplace, the online channel for companies to browse and buy solutions running on AWS. First and foremost, says Gasteen, Deckmatch benefits from increased “exposure”, and “a sales force within AWS pushing our products across the world.” Deckmatch’s customers also benefit: “a lot of our customers are going to be in the financial markets space, who have committed to spend a certain amount with AWS and independent software vendors. So it’s a great way for AWS to add value to those customers.” Finally, vendors selling through AWS Marketplace must meet certain requirements to prove the competency, security, reliability, and trustworthiness of their solutions. For Deckmatch, the listing therefore provided “a stamp of approval,” and marked “the first milestone of a very flourishing and beautiful partnership between AWS and AlphaLens,” says Gasteen.

One partnership, an entire ecosystem

Deckmatch’s early partnership with AWS included dedicated expertise to support technical and buying decision-making, which “has been really helpful along the journey,” explains Mustapha. “We found them to be by far the most helpful of the cloud vendors that we've ever worked with.” Since then, the partnership “has only grown in quality,” he continues, leading to a number of speaker sessions and joint presentations at major AWS events. Deckmatch is still in a relatively early stage of its startup journey and “pretty early on in our relationship with AWS in terms of exploring all the ways in which we could collaborate,” says Gasteen.

Even so, the rapid deployment of Amazon Nova Pro and subsequent launch of AlphaLens demonstrate what can be achieved over a short period of time, and the growth ambitions this unlocks as a result.

Deckmatch is looking to “really change the way in which people discover,” says Gasteen, evolving its solution and “turning it into an entire ecosystem that sort of drives itself.” Much like the (albeit slow) transformation of early search engines into multi-layered propositions, Deckmatch envisions “turning this into a valuable ecosystem play; a platform that not only connects the industry through unique data and a novel way of discovery, but also a place where there's a lot of user generated content over time.”

Equipped with the tools and expertise to enable rapid and cost-effective growth, Deckmatch looks set to follow a far faster transformation than that of traditional search engines. This sets the startup in strong stead to realize its ambitions and continue to expand its offering into a multi-dimensional toolset that delivers value for users and the growing number of companies utilizing Deckmatch’s services.