What's 🔥 in Enterprise IT/VC #343
How JPM Spends $15B on IT + collaborative doc template for super efficient board meetings
As you guessed it, what’s 🔥 this week is like most weeks, every enterprise startup and incumbent going all in on on AI. Below you’ll find a story from The Information on how Satya drove the OpenAI relationship, a slide from Palo Networks on how it’s infusing AI into many of its products, and how a bank like JP Morgan Chase spends $15.3B on IT - yep, there’s lots of data scientists and AI researchers.
Following last week’s post, a number of folks reached out for more details and frameworks on board docs (read 🧵).
While no set rule, one thing I can tell you is to please tailor by stage as in the early days there will be much less to report and as your company grows, you will need to cut some information out and focus on what really matters. If done right, those monthly emails you send to your investors can also be reused for some of the board material. From my experience, super early stage companies have docs around 3-5 pages long and as the company grows, different leaders take responsibility for sections. Here’s a great article from Harvard Business Review on “How Netflix Redesigned Board Meetings.”
At most companies, directors have a less complete understanding of the company than executives because of their limited exposure to day-to-day activities. The format of the information they receive does little to overcome this information deficit. The typical board book of a large corporation is a dense PowerPoint presentation spanning hundreds of pages in length. Some directors find these presentations heavy on data but light on analysis.
Furthermore, boardroom dynamics impede information flow, particularly in settings where the CEO maintains strict control over the content presented, when presentations are carefully scripted, and when presentations are made by only a limited number of executives.
Netflix takes a radically different approach. It incorporates two unique practices. First, board members periodically attend (in an observing capacity only) monthly and quarterly senior management meetings. What’s more, communication with the board comes in the form of a short, online memo that allows directors to ask questions and comment within the document. Executives can amend the text and answer questions in what is essentially a living document. We believe these two innovations meaningfully contributed to Netflix’s extraordinary performance in recent years.
The collaborative document IMO makes it easy to get the basic update and reporting out of the way, to level set all board members on the current status of the company, and allows the actual board meeting to be spent clarifying any comments or questions from the document and diving into more strategic topics. This only works if you send in advance, and the board members do their work which unfortunately is not a given.
Given that framework, here’s an example of some of the headline areas that I’ve seen covered for an early stage investment. Whatever the material you put together, one point I like to emphasize is that you should have most of this information accessible and trackable by you and your team internally so there is very little new information you should have to create. These written docs also remove the need to spend hours making a slide beautiful and just presenting the facts as they are. You only get real value if you don’t sugarcoat progress and look to learn from each meeting.
Not a must have, but I do like when founders put this upfront. This is a reminder of why the company exists. Emphasize how the user found your product, how they used it, what they learned, etc. As your customer base grows, it becomes even more interesting who you select as your one user/customer to highlight. This sets the tone for the rest of the meeting.
CEO TL:DR - highlights/lowlights/headlights
Summarize what is top of mind for you, what went well, what didn’t go well, what are top priorities going into the next month/quarter. This is also where you want to highlight the key discussion topics for the board meeting. Think about what you want to get out of your board while you have them there for this precious time.
Cash on balance sheet, cash burn for period vs. budget, how it’s trending, months of runway - any commentary on why spend was over/under budget - examples could be 1 time expense or did not meet hiring goals.
Great opportunity for demo time! Share screenshots, loom video, or do live in meeting. Review product roadmap from last quarter and delivery against those goals. What new features released and why do they matter for customers. What new features or product enhancements are you contemplating based on your own knowledge or customer feedback. How does product delivery compare to plan? What are blockers if any.
Sales Update (pipeline review)
For super early startups, you may not be at this phase yet, and this where I would shift more to product marketing (see below). If you are in sales mode, share your pipeline, what goals are for quarter and how you’re tracking against them, and highlight any key deals to land. I’d also highlight any key losses and reasons why.
If in early days and not ready for customers, I’d spend time on laying out the definition of your Ideal Customer Profile (ICP), the key value proposition for your ICP, and the work and conversations you are having to nail that. As you mature, this will change over time with more and more conversations with prospects. If you’re more mature, you can provide funnel analytics and other lead generation activities, comments on trends, and what you plan to double down on or remove.
What is current team size by department and how it’s trending, hires made, voluntary or involuntary attrition, what key hires to make in future - any commentary on morale, where you feel like you may have to beef up staff, etc
Remember these board meetings are for you to improve and accelerate the outcomes and increase the value of the business. Don’t forget the asks for the board - these include intros to potential customers, advisors, hires, etc. If you request intros to customers, it’s always nice to have a forwardable email with a one pager that your investor or board member can easily add commentary to and share with their network.
This is an appendix where you share the same stats every board meeting, your KPIs for your business. For devtools cos, these could include downloads over time, # of registered users, time spent using product, etc.
If you just want to use a deck, here’s a simple enterprise board deck template I put together which covers much of this ground. IMO, you may just want to use that as your framework for a collaborative doc. Try it, it will work wonders, and remember to use most of your time for discussion!
As always, 🙏🏼 for reading and please share with your friends and colleagues. For those celebrating, have a wonderful Memorial Day Weekend and please do remember those who served our country.
🎙️must listen w/friends @HarryStebbings + @snyksec founder @guypod, 🙏🏼 to have partnered on day 1 w/both his cos. Guy share how he 👀 around corners using 1st principles, simplifies problem to essence + anchors in future: will product be > or < necessary w/time - a great listen for investors as well as founders!
Guy also lays out the story of the struggles to get to $100k ARR before scaled to $650k to $4.5M ARR
Excited to represent all of those unsexy enterprise investors and founders! Seed 100 list here and thanks to Darius Rafieyan on the profile - also found here without paywall.
How do you secure OpenAI and other LLMs?
This week’s MLSecOps Podcast from Protect AI (a portfolio co) covers “Indirect Prompt Injections and Threat Modeling of LLM Applications.” The conversation with esteemed cyber security engineer and researcher, Kai Greshake, centers around the concept of indirect prompt injections, a novel adversarial attack and vulnerability in LLM-integrated applications, which Kai has explored extensively.
Great story on how Satya willed the OpenAI partnership to happen(Aaron Holmes - The Information)
But Nadella abruptly cut off Lee midsentence, demanding to know how OpenAI had managed to surpass the capabilities of the AI project Microsoft’s 1,500-person research team had been working on for decades. “OpenAI built this with 250 people,” Nadella said, according to Lee, who is executive vice president and head of Microsoft Research. “Why do we have Microsoft Research at all?”
How is $15.3B spent on IT by one company every year? Read the 🧵 and also check out Lori Beer’s (CIO of JPM Chase) deck from JPM’s annual investor day.
Adobe Generative Fill - simply insane - power of incumbent when it comes to AI and so many deepfakes to come…must watch how easy…
AI Canon from a16z - all the resources you ever need to start learning + building
Palo Alto Networks crushed earnings and shared its AI everywhere strategy.
from earnings call:
"This is a new baseline," Arora said. "We think there is continued opportunity from here and we haven't even factored in the potential impact of generative AI. As you've been hearing all the conversation in the industry, we're still working on it, we're understanding it, we're really looking at processes, but we believe there is a there there. We think there will be an opportunity in the future to get more efficiency from generative AI as we go ahead and implement some of the capabilities through our organization."
And another 50 infra cos go Poof as this week Microsoft launched Build your own Azure OpenAI Studio to build your own CoPilots…(watch here)
But I’ll create a model monitoring and performance co - well Datadog recently announced it’s own model monitoring and performance for OpenAI (VentureBeat)
Once up and running, the Datadog-OpenAI integration automatically tracks GPT usage patterns, providing teams with actionable insights into model performance and costs via dashboards and alert
For performance, the plugin looks at OpenAI API error rates, rate limits and response times, allowing users to identify and isolate issues within their applications. It also offers the ability to view OpenAI request volumes — along with metrics, traces, and logs containing prompts and corresponding completions — to understand how end customers are interacting with the applications, and to gauge quality of the output generated by their OpenAI models.
From Jamin Ball on Snowflake earnings call on data and cloud consumption trends - will there be a reacceleration? - full tweet here
Snowflake consumption trends (mentioned on earnings call):
Big headwind - customers changing their data retention policies. ie Instead of storing 5 years of data they store 3 years. This reduces storage costs, but also compute (queries run faster as they run against smaller data sets).
They called out strong consumption in Feb / March, but then a slowdown from Easter through today. Not entirely clear what caused this change in consumption patterns. They're modeling similar consumption trends like they saw from Easter to today for the rest of the year
The way they guide (as described on the call) - they look at 4 weeks of daily consumption leading up to call, and model that daily trend for rest of year. And in April they saw "no week over week consumption growth." So full year guide based on 4 weeks of week consumption relative to Feb / March
Nvidia shares spike 26% on huge forecast beat driven by A.I. chip demand (CNBC) - will this be like Zoom or are we getting ahead of ourselves again? From earnings call, enterprises also coming on strong with AI deployments
For example, Meta has now deployed its H100-powered brand Teton AI supercomputer for its AI production and research teams. Third, enterprise demand for AI and accelerated computing is strong. We are seeing momentum in verticals such as automotive, financial services, healthcare, and telecom where AI and accelerated computing are quickly becoming integral to customers' innovation road maps and competitive positioning. For example, Bloomberg announced it has a $50 billion parameter model, BloombergGPT, to help with financial natural language processing tasks such as sentiment analysis, named entity recognition, news classification, and question answering.
Auto insurance company, CCC Intelligence Solutions, is using AI for estimating repairs. And AT&T is working with us on AI to improve fleet dispatches so their field technicians can better serve customers. Among other enterprise customers using NVIDIA AI are Deloitte, for logistics and customer service, and Amgen, for drug discovery and protein engineering. This quarter, we started shipping DGX H100, our Hopper generation AI system, which customers can deploy on-prem…
Our collaboration will let ServiceNow create new enterprise-grade generative AI offerings with the thousands of enterprises worldwide running on the ServiceNow platform, including for IT departments, customer service teams, employees, and developers.