Posts

Selling software will evolve to selling agents, AI that acts on behalf of users.

The efficiencies for rote work are too massive to ignore for many uses. As the technology rapidly evolves, so too will the sales strategies.

Sellers & the startups they represent will need to re-imagine roles. In a sense, selling AI agents is analogous to category creation.

About a decade ago, Nick Mehta & the Gainsight team created the customer success category. In 2014, I remember joining a panel at the Gainsight annual conference Pulse, then in its second year.

Should software assist humans or act on their behalf?

In 2016, the question was easy to answer : sell Ironman not Robocop. Technology hadn’t reached the level of sophistication we have attained today where AI is 90% as capable as a high-school student, the MMLU benchmark for AI is precisely this.

The next generation of software startups have a strategic question with different terminology & potentially a different conclusion.

To be or not to be an agent, acting on behalf of workers?

Will AI sofware companies operate with better or worse profitability than a classic SaaS company?

Initially, I thought worse since the expense of serving AI as a product is signficantly higher.

But now I’m not so sure. AI SaaS may be much more profitable than the -10% average net income margins of the current crop of public businesses.

Yes, AI inflates the cost to serve the product. Google queries may be 10x more expensive than standard search results. That’s an unfair comparison since Google has focused on classic query cost optimization for more than 20 years.

Within data teams, a tension exists. Centralize the data analysis to ensure accuracy or enable end-users to analyze their own data directly which is faster & more direct.

The pendulum between these two states started with centralization during the 2000s with BI products from Microstrategy, Cognos, BusinessObjects, & Hyperion. In 2004, Tableau emerged from the Stanford campus to deliver their application to the users.

If you were to watch three videos on YouTube Shorts - one on Italian cooking, one on chess openings, & a third on crypto trading, YouTube Shorts’ recommendation algorithm combines the video descriptions with your dwell time.

Watching the osso bucco video to its end would trigger more Italian cooking specialty videos in your feed.

We believe every LLM-based application will need this capability.

Combining text & structured data in an LLM workflow the right way is difficult. It requires a new software infrastructure layer: a vector computer.

The database is being unbundled. Historically, a database like Snowflake sold both data storage & a query engine (& the computing power to execute the query). That’s step 1 above.

But, customers are pushing for a deeper separation of compute & storage. The recent Snowflake earnings call highlighted the trend. Larger customers prefer open formats for interoperability (step 2 & 3).

For those of us who love logic, the paradoxical title of this post should catch your eye, just as it did mine.

In Alchemy, the founder of a major brand agency describes the way many of the major consumer companies in the world created brands.

We call it breaking out : a double entendre which means both growing faster than competitors but also in a different way than their competition.

Rory Sutherland, the author explains it this way :

On March 15th at 9:30 Pacific time, Office Hours will host Colin Zima, CEO of Omni Analytics.

Colin is no stranger to business intelligence & data analysis. He worked on search quality at Google, founded a dynamic pricing company for the restaurant industry, then ran data at a HotelTonight before becoming Chief Analytics Officer at Looker through its acquisition by Google.

In the past, the bigger the AI model, the better the performance. Across OpenAI’s models for example, parameters have grown by 1000x+ & performance has nearly tripled.

But model performance will soon asymptote - at least on this metric.

Every successful startup faces critical marketing decisions from day one. How do you establish your value proposition? Which customers should you target first? What channels deliver the best ROI with limited resources?

During a recent session with a former executive Kady Srinivasan from Lightspeed Commerce, Dropbox, and Klaviyo, several powerful insights emerged on how early-stage companies should approach marketing. Here are three key takeaways:

1. Define Your ICP Before Anything Else

The most fundamental decision for any startup is determining your Ideal Customer Profile (ICP). Without clarity on exactly who you’re targeting, marketing efforts become diluted and ineffective.