In a matter of weeks, AI technology has gone from sci-fi trope to seemingly the topic of every other business meeting and conversation.
I’ve been through similar technological shifts before: the microcomputer, the internet, the smartphone – though those were dismissed early on as toys or niche solutions. That’s not the case with AI. Almost daily, it bounces between being the savior of our failed education and healthcare systems to posing an existential threat.
That high-level talk is fine, but my priority was to determine how this tech would impact my company, a retail venture capital firm. I needed to decide how my company should approach AI. Should I wait and see? Cautiously experiment? Or try something more aggressive?
I chose to go “all-in.”
Here’s my story of why I made this move and how we went about it.
Why we went all in
Why did I opt to go all in? Partly my background. As a student of Clayton Christensen, I’ve studied the power of disruption, and lived the ‘before and after’ results of powerful technologies that simultaneously wipe out and spawn careers, companies and whole markets.
AI is at least on this level of transformation, with the potential to be the most disruptive of all. I’ve learned – sometimes gratifyingly, sometimes painfully – that you want to be on the right side of epoch-defining tech disruptions.
I care deeply about our customers, owners, teammates and portfolio companies (including many in the AI sector). It’s my job to ensure we’re on the right side of this historical trade. Reflecting on my 35-year career, I realized my best decisions have been carefully considered “all-in” moves. That’s the approach I’ve taken regarding AI for our company, Alumni Ventures.
Another fundamental reason: I believe that AI can make a huge difference given the context of our company. In the world of venture investing, Alumni Ventures is unique in volume and complexity. We are retail VC (versus institutional), with almost 10,000 individual customers. We also invest a lot – 1,100-plus portfolio companies, adding approximately 300 per year. In fact, PitchBook ranked us as the third-most active VC on the planet in its 2022 global league tables. Another marker that charts our explosive growth is that we have grown from a single $1 million fund in 2015 to raising over $1.2 billion.
All that translates into lots of data to process, analyze and use, plus many repetitive processes that need to scale. Spread that need across 50 investment professionals and approximately 100 support staff from various departments (engineering, finance, customer service, OOI, marketing, P&C, legal) and you can see why AI seems heaven-sent to us.
Specifically, here’s the opportunity we saw for our company:
- Get insights to make us better venture investors;
- Make our business more efficient and scalable by automating hundreds of repetitive workflows and tasks;
- Reduce errors;
- Improve the experience of our customers.
And here’s the opportunity we envisioned for our people:
- Decrease their more routine, mundane tasks. Early data suggests that workers who use AI as part of their jobs are finding more satisfaction;
- Help with higher-level, higher-value thinking tasks;
- Equip our teammates for an inevitable future: I’m in the camp that says AI isn’t going to take jobs, but AI-powered people will.
Setting a clear vision and goal
Decision made: I had to rally our people. I generally believe in a barbell approach to management, which balances top-down and bottom-up approaches. And building consensus is often the best approach to align a company around a goal. But in this case, I knew the decision, goals and plan needed to be unequivocally communicated by me.
Within 12 months, our firm would become the most AI-powered VC in the industry. The statement was ambitious, provocative and a definite stake in the ground. My goal was to have every single person – from me to our interns – use AI daily in meaningful ways.
In support of that message, we created several special Slack channels, plus hosted numerous brainstorming and planning sessions. We also put resources behind the drive, ensuring that everyone in the company had access to ChatGPT, tutorials and consultants. Initially, we just wanted folks to play with AI tools.
Next, we began sharing experiments (things that worked and those that didn’t), prompt suggestions and industry developments. Leadership also modeled use cases, offered guidelines and cautioned folks about tool limitations and safety issues. Employees were starting to find the tools not only fun but also helpful.
After getting teammates acquainted with the tech, we wanted to home in on very practical ways it could impact daily work. Our focusing mechanism was the creation and launch of AV’s AI Challenge. This contest split the company into teams – 43 in all – based on departments. The contest charged our people to come up with AI applications relevant to their jobs, workflows and deliverables.
Expectations were clear. We laid out the scorecard (process 40 percent, results 40 percent and presentation 20 percent) and a deadline. Carrots were prizes of cash or equity ($100,000-plus in total), plus obviously seeing the best ideas backed and developed by AV resources. We made tech and consultants available to help, but largely left the teams to innovate independently. Critically, we walled off deep work time blocks for them to huddle, discover and build.
Another key decision was to keep the bosses out of the way. This contest was to bubble up great ideas and automations from the field. We wanted to empower team members to discover and think, creating true ownership of solutions.
We chose to do a contest with prizes for several reasons:
- Dramatic signaling;
- Distributed and practical;
- Learning is often better in a peer-to-peer setting;
- Fun and interesting;
- Competitive juices;
- Hard deadline on results.
Six weeks later, the judging took place over three days. Each group had 10 minutes for their pitch. Judges were asked to hold questions until teams finished their presentations, then we privately scored and conferred. I made the intentional decision to watch every session.
The results: we found the ideas to be astounding in terms of variety, applicability, impact and sheer cleverness. To my surprise, the floor was incredibly high – ranging from good to ground-breaking.
Here’s a small sampling of those ideas, some proof of concept, some nearly complete, all leveraging AI:
- Customer service: Automation that generates personalized, on-demand portfolio reports for all 10,000 customers, showing their investments, performance and diversification;
- Legal: Automating 80 percent of the most time-consuming legal tasks, such as comparing term sheets to legal documents;
- Investing: Rapid identification and evaluation of rising venture professionals, enabling early partnering with attractive deal sources;
- Investing: Detection of ventures in the process of raising a round (usually a very private process) by spotting faint data signals in public communications;
- Sector research: Efficient compilation of information on an emerging sector and prep of a white paper to quickly deliver rich, topical content in one-twentieth of the time;
- Portfolio monitoring: Monitoring, interpreting and alerting investing teams on key changes at the 1,100-plus companies in our portfolio.
Ultimately, we chose 20 winners from across the company and provided all teams with constructive feedback.
Besides walking away with many impressive and actionable AI applications, the judges and I learned what worked about our approach. My takeaways:
- Insights into problems: Understanding the problem being addressed was as critical as solving it. Team members walked us through workflow diagrams, task mapping and identification of roadblocks. This process not only gave the judges key insights but also allowed team members to detach from daily operations and understand the bigger picture.
- The power of teams: Working in teams proved hugely beneficial. The challenge brought teams closer and allowed those lagging in AI adoption to bridge the gap. Teams inspired each other and found unity and pride joining with all employees in working on an innovative, challenging initiative.
- Balancing top-down and decentralized approaches: Our combined top-down and decentralized approach was highly effective. While directives, vision and resources were provided from the top, employees were empowered with the tools and autonomy to explore and innovate, encouraging ownership.
- Framing problems and solutions: We suspected and confirmed that a one-size-fits-all approach wouldn‘t work for implementation. After the contest, we developed a tailored framework that evaluates the level of development and resources required for each idea, along with its potential impact. Our evaluation involves identifying problems that can be solved with ready-made solutions and others that require in-house development for proprietary advantages.
- Significance of corporate commitment: We walked the walk. A strong corporate commitment in terms of messaging, time, resources and funding encouraged employee engagement and buy in.
- High-tech and high-touch: We learned that optimal solutions were both high-tech and high-touch. Most of this technology is not ready to be unsupervised, so our teams carefully considered human QA and “last eyes” processes. However, by automating routines or gaining new insights, we will free up time to think at higher levels and to do things that humans do best, like build relationships, connect far-flung ideas in novel ways and lead.
- Not letting up: Post-challenge, we found it was essential to keep communicating regarding results, development plans and implementation. In addition, continual information sharing about AI has helped engrain the technology into AV’s culture and our way of thinking.
We’re already deep into phase two, sorting through contest ideas, determining what teams can implement by themselves vs what needs collaborative development and priorities for our work. Our cross-departmental implementation team has been evaluating, bucketing and prioritizing contest ideas, focusing on the highest-impact solutions first. We already see that developments on one solution can be applied to others, so we expect development to accelerate as we proceed.
Beyond development plans is our commitment to continue experimenting, sharing and implementing AI. Undoubtedly the tech needs to improve in some areas and challenges like bias, security, surveillance and other issues loom. But this is the worst the tech will be, and it’s already amazing. Given AI’s stunning evolutionary pace, improvements are likely already on the horizon. Welcome news and our company is ready for the challenge.
Every great technology wave has been accompanied by handwringing and catastrophizing. But the net results of powerful new tools are generally positive. The AI genie is not going back into the bottle. For the sake of our society, company, customers and teammates, AV plans to stay all in on AI, making the best use of the tech.
Mike Collins is founder and CEO of Alumni Ventures, which has raised more than $1.1 billion across more than 30 Alumni and Focused Funds, serving a network of more than 9,000 investors and 600,000 community members. The Harvard MBA began his career at VC firm TA Associates. In addition to being a venture capitalist, Collins has started multiple companies, including Kid Galaxy, Big Idea Group (partially owned by WPP) and RDM. Connect with him on LinkedIn.