Avid Larizadeh Duggan, senior managing director at Ontario-HQed growth-stage investor Teachers’ Venture Growth, is one of Europe’s first operator-turned-VCs.
Her career in tech began with stints as a product manager at places like eBay and Skype, and she’s switched between operational and investing roles ever since.
In 2010, she cofounded ecommerce site Boticca, in 2014 she joined Google Ventures and in 2018 she took up the COO role at music startup Kobalt Music. Then in 2019 she cofounded The Operator Network — a group of C-level public company execs who angel invest and offer advice to early-stage entrepreneurs — and in 2021 she joined TVG, which is part of the Ontario Teachers’ Pension Plan, one of the world’s biggest pension funds.
Its investments include German AI translation company DeepL, German battery startup Instagrid, Swedish healthtech unicorn Kry, British chip scaleup Graphcore and French health insurance unicorn Alan. Her personal angel investments, meanwhile, include French AI unicorn Mistral, Italian unicorn app developer Bending Spoons and Dutch biotech Cradle.
It’s a CV which has given her a fairly unique insight into what makes a good (and bad) board member, and about how many more AI hype cycles we have ahead of us.
We get into all of that on the latest episode of the Sifted podcast. Listen to the full conversation here — also covering M&A, TVG’s investment focus and what it’s like working for a pension fund — or read some lightly-edited highlights, covering AI and board members behaving badly, below.
When you’re sitting around those board tables, do you act differently from some other investors who haven’t ever run a company or had an operational role themselves?
I think so, particularly those who’ve been in the industry for less than a decade, who haven’t had enough cycles or been with enough companies to see what it takes to build a business. As an operator, or as somebody who’s been in VC for two decades, you really get a sense of what it takes to build a business. Partly, it’s the empathy for the entrepreneur’s journey; just how tough it is, how you know your highs are much higher, but your lows are really low, and the resilience that you need to build… Whether it’s a macrocycle like we’ve seen, or because you have a falling out with your cofounder, or your CFO decides that he needs to go back to the US, or whatever it is… Every day brings a different type of crisis, and that wears you down and so managing yourself through that is incredibly important.
A lot of this is about human relationship and communication; being able to get the founder to trust, to communicate — and that’s always easier if you’ve been in somebody else’s shoes before.
Would you say that’s often a challenge — getting founders to really be open? Is there a general wariness towards letting on when stuff’s really going wrong, or asking for help on particular issues?
I think so, especially for first time founders. As a second time founder, somebody who’s gone through this before or seen the movie a number of times, you know nothing goes up and to the right all the time. You’re always going to expect something to go wrong. So many successful companies have had near-death stories, and they’ve managed to pick themselves up.
As an investor, if you know this, you’re calm when the shit hits the fan. And that calmness helps the entrepreneur. It’s almost like a parent-child relationship. And in this particular example, if your mom or dad is yelling at you because they believe you’ve done something wrong, then the next time, you’re probably less likely to be comfortable telling them what’s happened. It’s a similar dynamic, and it’s human nature: your investor is somebody who’s given you money, and you’re a steward of that money, and so you may be slightly more reticent to tell them about things that are not going right, thinking that you’re going to upset them. And so the role of that investor is to say, ‘No, I’m your partner. I’m here to build this with you. I’m not just giving you money, I’m giving you my time, my experience, and we’re going to go through this together.’
What are some great attributes of a board member — not immediately freaking out when stuff doesn’t look like it’s going that well… What are some other things?
Being prepared! Just reading the board packs and not just reading them five minutes before going into the boardroom, but preparing them, thinking about the business, coming with the right questions and the right answers to the board… I think that’s incredibly important — and makes a massive difference.
Part of also being a good board member is to have an entrepreneur who can manage its board well. And as a board member, you can help them do that. Your board is not there for you to summarise the metrics and report to them; you need to come to the board with the 1, 2, 3 questions that you want answered, the critical aspects or the stuff that’s not working that you need help with, and then engaging the board to do that.
Being present — simple things — being present at the board, not being on your phone, not being on your laptop, really being engaged in that, is incredibly important. As is the way you communicate; one of the traits of people who haven’t been in operations, who sit on boards, is they think that they tell the entrepreneur ‘You need to hire X person’, or, ‘You need to go and fix X thing’ and they expect that to happen tomorrow or in two weeks, and it just takes time to operationalise that, to execute on it. And so having that understanding of that journey and the right expectations…
A good board member is also a coach. You’re really there. You can’t do; you need to know you cannot do. You’re there to influence and to coach, and you need to be able to take time outside of the board to do that. So whether it’s going on a walk with the entrepreneur, where it’s easier, one-on-one to air out some of the challenges, to get into the details of things. So making time outside of that board meeting is critical.
Where do you think we’re going to see AI really make a difference? Where will it live up to the hype?
Over the next 10 years, I think most everyone will be using it in one way, shape or form. I think we are at the beginning of that cycle. If you compare it to cloud, it took about a decade or more for the big cloud providers to reach a billion in revenue. It just takes time, from an enterprise perspective, to adopt. Enterprises have processes; they have thousands of people, you need to change the tools that they use, the way that they work. There’s compliance, there’s security… Enterprises really don’t like adopting version 1.0 and we are today at version 1.0. Version 2.0 will come out, and people will feel a bit more comfortable; the companies that are selling the AI will have been in the market longer, understand the sales cycle, know how they want to sell to these enterprises… And so that will, over time, accelerate.
I do think that we will all be using AI, just like another form of software, in five or 10 years, and we will need to adapt the way we work to it. I think AI will make us smarter in many ways. I’ve noticed in some of these applications that we’ve been looking at that, at least today, the best way to interact with that AI is to understand a process you’re trying to make it learn and adapt to. So when you are prompting it, you know what are the prompts that you need to ask it to get the answer you want. So if you’re trying to, for example, automate a process, you need to know that process so well that you can actually break it down for your AI. So it requires someone to be excellent at their craft, to truly understand how they think about what they do, whether it’s how they respond to a customer or support query, whether it is how they think about the steps of leading to a great design or leading to identifying where fraud comes from. And so I think those people that truly understand these methods will be in a much better place as AI takes over.
You’re an angel investor in Mistral, which many people consider one of the forerunners in Europe, creating LLMs. Can a European startup win against the likes of OpenAI, which just has so much money and so much attention gathered around it already?
One of the very attractive attributes of Mistral is that they are very frugal. I don’t think I can disclose it, but the amount of capital that’s gone into developing the models that they have now, that are competitive with these other models that you’ve described, is a fraction of the cost of what these other organisations have spent. And the team, the three cofounders, are responsible for that. That’s what they brought to the table. Partly, it’s good to not be the first mover and to have learned how to optimise that cost. Partly, it’s also their strategy, what they’re doing. They started initially with the developer community, as opposed to going after large enterprises. Now they also have an enterprise offering, but they really listened to the developer community and built a brand for the developer community, built tools because they are part of that community, which is different from OpenAI’s approach — and their open source approach also caters to that.
Mistral had an enormous valuation at its last round… Are we starting to see the valuations cool off a bit, or are they still pretty astronomical?
I think the multiples relative to others are pretty astronomical. And even if some may have cooled down a bit, wait until you get GPT 5.0 or some other leap in the progress of the LLMs, and you’ll probably get another boost in valuations and hype and excitement. So I don’t think we’re out of the woods in terms of these types of cycles. If you look within the 10 to 20 year time frame, I think there are companies that don’t exist today that will be billion dollar companies. Google came several years after the internet boom; Netscape, Yahoo, all of those came before Google and Amazon… So I think we still have a long way to go, and a lot of value to be created, and along the way, a lot of companies will fail, whether they become acqui-hires or simply fail. And the entrepreneurs might start a new company.
So if I’m a PhD student at Oxford or TUM in Munich or wherever, it’s potentially not too late for me to leave and start my billion-dollar AI company?
Absolutely not; I think you can do that, but I would say to that researcher, ‘Make sure you know what you’re getting into’. The challenge is, researchers like to do research; they don’t necessarily understand or want to or get excited by building product with commercial value, answering customer requests and catering to that, which is a process that takes a lot of minutia, a lot of listening, a lot of execution, and you’ve got to love it. And I think that’s part of what we’re seeing in the ecosystem, in times where it sounded cool to go and start a company, because people are throwing money at any smart researcher. But it still comes down to product, market, TAM, and you can’t shortcut that.
Read the orginal article: https://sifted.eu/articles/avid-larizadeh-duggan-podcast/
