How CB Insights uses AI
I write about businesses using AI and today I’m excited to show you how CB Insights is using AI.
CB Insights is a tech market intelligence platform that provides data and analytics on private companies and emerging technology trends.
Thank you to CEO Anand Sanwal, CPO Nichelle Dekeyzer, and CTO Mike Ruggieri for spending time with me and sharing their thoughts.
Here’s a quick breakdown:
- Creating a cross-functional AI tiger team
- Encouraging adoption through Hack Days & Lunch and Learns
- Launching new AI-powered products
- How AI boosts their current offering
- Ways in which AI can transform the customer experience
- Reimagining how they capture and create data
CB Insights has been using machine learning for several years.
This 2014 post mentioned that 70% of their data collection was from their machine learning software, with the remaining 30% from direct submissions from investors.
We’ve built machine learning software which parses unstructured and semi-structured data sources and programmatically extracts the key pieces of structured data we care about. Things like company & investor/acquirer name, amount of funding, valuation, date, stage, board of directors, patents, etc.
The core of this data is extracted from crawlers that analyze 150,000+ sources on a daily basis. In the last 8 months, we have crawled and analyzed nearly 16 million unique articles and information sources. The list of sites we index grows regularly but includes the following mostly unstructured data sources:
Regulatory filings – form Ds as an example
Investor websites (press & portfolio pages)
Company websites (press pages & blogs)
Acquirer websites (press pages or investor relations sections)
Press releases
Social media (Twitter primarily)
A select group of local, national and international news and trade publications
This was pre-generative AI. Now they’re integrating AI internally and within their product in several ways.
Forming a Cross-Functional AI Tiger Team
CBInsights formed a "Tiger Team" that brought together top performers already doing innovative AI projects across the company.
The tiger team was born out of a desire to innovate ways of working quickly across multiple teams. Like many companies, we found that we had pockets of genius throughout the organization where really incredible and innovative work was happening with GenAI, but that work was being done in a silo.
The tiger team framework allows us to bring together very cross-functional groups that were already innovating in their functional areas to work together on high-impact, company-wide problems.
Nichelle Dekeyzer, CPO
The Tiger Team is split into three workstreams.
- Content creation
- Research
- Sales automation
But they continue to bring the whole group together weekly to share learnings which influences work happening in other projects.
One of the first initiatives was a dramatic overhaul of the content creation process. They turned long-form research briefs into easy-to-digest content for customer outreach.
Traditionally, it would take four days.
Now it’s less than 1 hour.
They used ChatGPT (GPT-4) to condense research briefs into < 60-second scripts. Then, they used WellSaid Labs for lifelike voiceovers and Jitter for the video. Here’s the output - “Why Flexport’s valuation might have dropped by 80%”.
"Our content creation team focused on creating short videos... The entire process was done using generative AI and took video creation down from 4 days to less than 1 hour from start to finish,"
The customer insight group developed a Slackbot that delivers relevant market and competitor research to customer success reps in under 30 seconds instead of 2-10 hours of manual search.
The bot makes individualised recommendations on relevant markets, companies, or competitive insights.
For example, our customer success team can input a user into the bot and receive back a Market Report or Feed recommendation that would be beneficial for that specific customer.
The sales automation group also benefits from this research tool.
On the flip side, the research team or sales team could feed the slackbot a piece of research content and get a list of users who are likely to care about this research.
Encouraging AI Adoption Through Hack Days & Lunch and Learns
CBInsights held an all-company "Hack Day" focused on AI and regular "lunch and learns" demos to encourage AI experimentation.
Hack Days have been successful internally for ~10 years, started by Anand (the CEO) and the early engineering team.
The entire company hacks, not just the tech teams.
After ChatGPT was released, they hosted a Hack Day entirely on Generative AI.
This was another intentional move to get teams comfortable with and exploring using LLMs in their day to day work. Historically, the aim of Hack Day was to generate great ideas and hopefully quickly get them into production to impact the business. The GAI Hack Day was no different, but I'd say that my primary goal was to get people excited to "play" and to find out what they could build quickly.
Mike Ruggieri, CTO
The goal of this Hack Day was to create a real impact while also allowing the teams to understand the potential of using AI.
The outcome included multiple product ideas which were ultimately launched to customers - summarizing business partnerships, automatic writing of company descriptions, scorecards comparing competitive tech companies.
CBInsights hosts regular Lunch and Learn sessions where team members show interesting things they’ve been building.
Some developers were early adopters of AI, primarily using it for coding tasks.
One senior engineer was asked to highlight his project in a session. Having a well-respected, senior engineer show that level of excitement and the power of GPT for coders was a game-changer.
a senior engineer working on our data pipelines team. He had recently adopted a number of components coded in Java. He was not a java guy. He fed the java code to ChatGPT and asked it to rewrite in Python, a language he was proficient in. Moments later, he had code he understood. Further, he had asked ChatGPT to write comments in the code. His mind was blown.
The goal was to excite even AI sceptics about its possibilities and destigmatize using AI. Now, AI assistants are seen as essential tools rather than "cheating" among CBInsights' development teams.
Launching New AI-Powered Products
CBInsights is releasing its first customer-facing AI features, like automatically-generated company descriptions that still undergo human review (with an approval rate of over 99%).
"Our first AI-powered features were focused on back-end data types of processes...We are about to release a full set of GAI-powered features within a tool we call the CBI Analyst."
New tools like CBI Analyst will provide executive Briefing-style summaries, or “scouting reports”, of companies by condensing quantitative and news data to 1-2 pages. This will drastically reduce the time customers need for insight before key meetings or decisions.
This was tested with human sampling until they were comfortable with the results to roll out more broadly.
They’ve used AI in the site's search functionality to test if the user gets the right results.
The Future of AI at CB Insights
Anand envisions a future where AI is integral to every aspect of CB Insights. He sees Generative AI not just as a tool for enhancing current offerings but as a fundamental driver for reimagining data capture, creation, and utilisation.
It was clear from the first moment we saw chatGPT that we could use GAI to transform how our customers make company defining decisions about
Which markets to enter
How to understand and outflank competitors
Who the right partner, investment and/or acquisition targets are
The goal is to:
move beyond helping our customers search for information into an age of answers.
He sees massive potential in combining the company's data assets with AI.
The combination of our proprietary quantitative & qualitative information and Generative AI, which is an unbelievable information compression engine, is allowing us to dramatically collapse time-to-insight for our customers.
The long-term vision is to train tools like CBI Analyst to be a knowledgeable "co-pilot" that make users "bionic" in assessing markets, competitors, and partnerships.
We’re also looking at how GAI can help us reimagine how we capture and create data as well as how it can allow our go-to-market, research and engineering teams to do their best work.
If you found this post valuable, share it with a friend, and consider subscribing. Feel free to suggest new topics for me to cover.
Cheers,
Yikes, a paywall!
70+ tutorials, courses and case studies wait behind it. No subscription, $150 paid once.
✅ Case studies on companies using AI
✅ Private community access
✅ No subscription, $150 paid once
✅ Expense it using this template. Or get a team account.
.png)
More like this
Start learning today
If you scrolled this far, you must be a little interested...
Start learning ->Join 3,107 professionals already learning
