AI in search of (enterprise) customers

Writer Lauren Goode succeeded in generating the catchy title, ‘The Unsexy Future of Generative AI Is Enterprise Apps’ for her article in WIRED magazine. Why are we suddenly talking about AI and unsexy at the same time?

It appears that economics and absence of strategy are catching up with few AI startups. What originally sounded good on paper (or on PowerPoint slides) has a hard time in the real world.

A few quotes from the article:

'Tome, a San Francisco startup that makes presentation software juiced with generative AI. The company launched its product in early 2022 with a healthy cushion of $32 million in venture capital funding, and successfully surfed the ChatGPT hype wave after that, raising even more funding in early 2023.'

'Tome had one problem, though: It wasn’t generating meaningful revenue. And AI startups like Tome, which build their services on top of both open source and proprietary language models, pay significant fees to companies like OpenAI to power their apps.'

'They also announced a new focus: Their app, which is often described as PowerPoint-on-GenAI, would be aimed squarely at enterprise customers. They would now charge three times what they were charging premium users.'

'.. let’s pick a segment of customers that not only have a lot of presentations to build but also have clear outcomes, like whether they closed a deal or not. And that is salespeople.'

Reading these statements, you wonder:

• How many millions do you need to create PowerPoint-like presentations?

• Did you ever, instead of PowerPoint, open Excel and run the numbers to see how much it will cost Tome to generate one slide deck?

• Do you know how to sell to enterprise customers?

• What were the previous assumptions about your potential customers?

Let's keep scrolling.

'Perplexity, another buzzy startup that offers an AI-powered search engine, announced Perplexity Enterprise Pro (along with another massive round of funding)'

'OpenAI says the company has spent the 18 months since ChatGPT launched in 2022 building out its software sales and go-to-market team, growing it from 15 people to 200 employees. The group now makes up one-fifth of OpenAI’s current workforce.'

Yes, the panacea of making real money. Selling to large customers.

Since this newsletter is called Recurrent Patterns, you know there is a story from the past.

It was the early 2000s, and Google, an emerging behemoth, had search technology that was leaving its competition in the dust. The company realized it could make even more money by selling the same technology to enterprises. Just imagine — the same algorithm put on a server and shipped to the customer can generate an untold amount of revenue. Millions of customers. Billions of dollars.

WIRED magazine ran a story about it in 2012.

In 2016, Google advised its business partners and customers it was sunsetting the product, and it shut it down in 2018. How is it possible that a company generating billions in revenue and profit can pass on such a huge opportunity?

A glimpse of the answer can be found in the article from ComputerWorld, where at the end you can read: 'While the Search Appliance’s functionality still isn’t as sophisticated as that of high-end products from Autonomy and Microsoft’s Fast Search, it has now moved upstream, said IDC analyst Susan Feldman.'

But you can find a better post mortem in 'Lessons from the Death of Google Search Appliance' and the headlines for each paragraph outline the perils of working in the enterprise environment:

• Enterprise Search Is Hard

• Google Doesn't Do Customization

• Application-level Search May Matter More

• Google Doesn't Do Corporate Customers Very Well

• Consumer Tech Needs to Be Adapted for the Enterprise

Unlike the Internet, where the search engine can index anything which is available, within corporate settings, you can't do that. Not every document is available to everyone, and it’s based on your access privileges. You might see only a subset of all the available information. If the previous sentence sounds simple, let me assure you that it is an incredibly difficult problem to solve. And it’s one of the reasons why Google left.

Back to the present with OpenAI and its 200-person sales team. What are they going to sell and to whom? How long does it take to sell anything to a large company? These people will learn about corporate budget planning, about competing interests between various departments and how to navigate the org chart. And just as they are about to close the amazing deal, the executive sponsor leaves the organization, and they must start from scratch.

Or, the deal is completely dead since the other people really didn't like the past executive. I guess it would also be premature to warn them about the procurement department, seconded by legal, and finish it with the introduction to the IT people.

All that will be happening, while the enterprise sales department is burning through $20 million (my rough estimate) a year on payroll with an equal amount spent on travel-related expenses. In no time, the sales people will be complaining that the marketing department is not supporting them enough and not providing enough leads to sell to.

But this is still the fun part of this process. The real excitement starts when the sales people start coming back with 'almost' signed deals dependent on whether certain features are in place. In the beginning, the product managers might accommodate small requests. But before they know it, they are creating a customized version of their base product for each new customer.

The killer? Ongoing support and maintenance. A simple product, which was the same for everyone and accessible through a browser and credit card on the internet, has become a multi-headed monster.

Don't let yours truly bring down the festive atmosphere when the contract is signed, congratulations are passed around, dinners are eaten and bonuses paid.

Once all the salespeople have done their jobs and the customers have the required licenses, it’s time to turn all this into a gold mine or money-printing machine.

Before the customers get their gold nuggets, they either have to hire new people to implement the software, or bring in outside consultants. In both cases, these newcomers won’t know the organization enough to understand how to properly integrate the software into the enterprise. A steep and expensive learning curve for everyone involved.

Would be too late to mention the study by MIT Sloan Management that found 'only 10% of companies obtain significant financial benefits from artificial intelligence technologies.'

As you can see, I can go on and on about the wonders of working with enterprise customers. It takes years to build an organization to support this type of effort and deliver real value to the customer.

If you have never been part of a meeting of an enterprise sales team discussing what and how to sell, this is a must for you to watch.

The recurrent pattern? I am sure you already know the future of the enterprise sales teams in these companies. They will be the victim of bad strategy.