How to buy an AI solution the right way: 7 questions new customers should consider
Here's a high-level guide for evaluating an AI-based solution as a potential customer — an enterprise buyer scorecard, if you will.
Embrace these FinOps best practices to ace your cloud strategy
Here are some best practice takeaways you can consider to help ensure your cloud transformation is successful across every metric
Ask Sophie: What’s the wait time for EB-2 and EB-1 green card categories for those born in India?
Back in 2018, the Cato Institute estimated it would take 151 years for a person born in India to get a green card in the EB-2 category.
4 ways to show customers they can trust your generative AI enterprise tool
Here are ways your generative AI organization can build as an open book while building trust with potential customers.
The 6 most important things to know about SaaS+ product architecture
Whether you want to build a SaaS+ company from scratch or turn an existing company into a business that can monetize embedded products and services, these are the six key concepts you need to know.
These concepts have technical implications but are as much business logic decisions as architectural ones. A founding team should have a shared perspective on these six issues. Being aligned on these concepts will drive product roadmap, core technical architecture, pricing strategy and product marketing.
You can 10x the revenue of your SaaS company by putting the right building blocks in place from the start. If you build the foundation of your SaaS+ house correctly, you can remodel the interior fairly easily in the coming years.
1. Everything revolves around the transaction
Shopping cart functionality and flexibility at the transaction level are two of the critical technical elements in SaaS+ because a high percentage of revenue typically revolves around the flow of funds on the platform. There are a couple things to think about when building transaction technology:
- Multimerchant cart: Building a shopping cart can get complicated when you’re taking into account more than one merchant in a single transaction, but architecting the cart to handle this from day one will pay dividends when you look to sell multiple SaaS+ products at the time of checkout. Specifically, this technology allows the user to see a single transaction, while behind the scenes there are actually multiple transactions occurring simultaneously, with each vendor individually. This is especially critical if you hope to sell regulated products such as embedded insurance.
- Split transaction/payout tech: An alternative to the multimerchant cart is the split transaction and split payout technology. This capability allows a truly single transaction at checkout, but then carries the burden of instantly associating net amounts due with each interested party and then settling with them by dynamically distributing the correct funds to recipients following the transaction. This is often initially viewed as the more elegant solution, but it doesn’t work for regulated products like insurance where the original transaction has to be with the actual insurance company. Realistically, you need to build both a multi-merchant cart and split transaction capability from day one.
For the end user, it all boils down to the checkout experience. Whether a given transaction is leveraging multimerchant cart technology or split transaction tools, the choice should be invisible to the end user while also accommodating a myriad of different SaaS+ products sold at checkout.
Being aligned on these concepts will drive product roadmap, core technical architecture, pricing strategy and product marketing.
Example: At SportsEngine, we built a commerce system where customers use one shopping cart to check out, but each item is actually being bought separately. The customer enters their payment information only once, after which the platform initiates multiple transactions on their card behind the scenes. So, for example, they can register for Minnesota Hockey and USA Hockey in one step, while also buying insurance and their uniform in the same transaction. One cart, one checkout . . . four independent vendors.
Etsy also allows customers to buy from multiple merchants in a single transaction. The marketplace then splits the payment between the different vendors and themselves. DoorDash’s single vendor shopping cart allows you to add multiple dishes from a single restaurant, but to order from two different restaurants, you need to make separate orders.
2. Single instance of a human
Don’t silo your data. You want to create one profile per person and use it everywhere on your platform. This means: one profile, one payment method, one background screen and one rating system for each human. Build your data model so each person’s profile and information is available across the entire platform. This is the power of the platform.
Making AI trustworthy: Can we overcome black-box hallucinations?
Like most engineers, as a kid I could answer elementary school math problems by just filling in the answers.
But when I didn’t “show my work,” my teachers would dock points; the right answer wasn’t worth much without an explanation. Yet, those lofty standards for explainability in long division somehow don’t seem to apply to AI systems, even those making crucial, life-impacting decisions.
The major AI players that fill today’s headlines and feed stock market frenzies — OpenAI, Google, Microsoft — operate their platforms on black-box models. A query goes in one side and an answer spits out the other side, but we have no idea what data or reasoning the AI used to provide that answer.
Most of these black-box AI platforms are built on a decades-old technology framework called a “neural network.” These AI models are abstract representations of the vast amounts of data on which they are trained; they are not directly connected to training data. Thus, black-box AIs infer and extrapolate based on what they believe to be the most likely answer, not actual data.
Sometimes this complex predictive process spirals out of control and the AI “hallucinates.” By nature, black-box AI is inherently untrustworthy because it cannot be held accountable for its actions. If you can’t see why or how the AI makes a prediction, you have no way of knowing if it used false, compromised, or biased information or algorithms to come to that conclusion.
While neural networks are incredibly powerful and here to stay, there is another under-the-radar AI framework gaining prominence: instance-based learning (IBL). And it’s everything neural networks are not. IBL is AI that users can trust, audit, and explain. IBL traces every single decision back to the training data used to reach that conclusion.
By nature, black-box AI is inherently untrustworthy because it cannot be held accountable for its actions.
IBL can explain every decision because the AI does not generate an abstract model of the data, but instead makes decisions from the data itself. And users can audit AI built on IBL, interrogating it to find out why and how it made decisions, and then intervening to correct mistakes or bias.
This all works because IBL stores training data (“instances”) in memory and, aligned with the principles of “nearest neighbors,” makes predictions about new instances given their physical relationship to existing instances. IBL is data-centric, so individual data points can be directly compared against each other to gain insight into the dataset and the predictions. In other words, IBL “shows its work.”
The potential for such understandable AI is clear. Companies, governments, and any other regulated entities that want to deploy AI in a trustworthy, explainable, and auditable way could use IBL AI to meet regulatory and compliance standards. IBL AI will also be particularly useful for any applications where bias allegations are rampant — hiring, college admissions, legal cases, and so on.
Ask Sophie: What are the visa options for a startup founder with family?
Here’s another edition of “Ask Sophie,” the advice column that answers immigration-related questions about working at technology companies.
“Your questions are vital to the spread of knowledge that allows people all over the world to rise above borders and pursue their dreams,” says Sophie Alcorn, a Silicon Valley immigration attorney. “Whether you’re in people ops, a founder or seeking a job in Silicon Valley, I would love to answer your questions in my next column.”
TechCrunch+ members receive access to weekly “Ask Sophie” columns; use promo code ALCORN to purchase a one- or two-year subscription for 50% off.
Dear Sophie,
I’m a startup founder in Berlin. I just returned from a visit to Silicon Valley where I met with a new customer. On the trip, I realized I need to be based in the U.S. to grow our base with U.S. customers.
What are the best visa options for me and my family? Will any of them allow my husband to work and continue his career?
— Seeking Scale
Hey there, Seeking!
Kudos to you on your business successes so far — and for your courage to take the next big leap to relocate to the U.S.! I’m honored that you reached out as you and your family begin your journey. I’ve got you!
You may be able to avoid having to go through an in-person consular interview for L-1 or O-1 visas if you apply now because until the end of this year, the Department of State has given consular officers the discretion to waive the visa interview requirement for certain work visas if the beneficiary was previously issued a visa and has never been refused one.
Consult an immigration attorney who can guide you to the best immigration options for your and your family based on your circumstances, timing and goals. There are a variety of options that might apply to you, based on various factors such as having a co-founder in a specific role or your citizenship in certain countries, but for now let’s dive into two of the visa options for you and your family so you can compare the general pathways!
L-1A is a top option
If you have worked for your startup for at least 12 continuous months in the past three years and can document your employment through payroll slips or tax documents, your startup can file for an L-1A visa for intracompany transferee executives or managers for you to come to set up an office in Silicon Valley.
To get an L-1A visa to open a new office in the United States, your company will need to sponsor you for the visa and show that you’ve secured a physical office location. Your company may also submit business plans, growth models, and organization charts. If you’re setting up a new office in the U.S. and are approved for an L-1A, that type of visa will can be initially valid up to one year. To extend the L-1A beyond that, you need to show that your U.S. business met your growth models and that the business is viable.
If your startup applies for an L-1A on your behalf while you’re in your home country, once the petition is approved, you will need to apply for a visa at a U.S. embassy or consulate. Consular posts have the discretion to waive interviews on a case-by-case basis at least through the end of 2023.
Some people visit the U.S. first on ESTA or a B-1 business visitor visa to secure an office and meet with prospective customers. It’s crucial to keep in mind that the B-1 is not a work visa, so while working in the U.S. is not allowed, you can perform some business activities, such as participating in meetings and signing a lease or other agreements. When you have an office and meet all the other requirements of the L-1A, your startup can petition you for an L-1A.
The B-1 visa is good for six months initially (ESTA is valid for only 90 days at a time) and can be renewed once from the U.S. for another six-month period if necessary. Premium processing is available for the L-1A, which means for a fee, U.S. Citizenship and Immigration Services (USCIS) will either decide on your case or issue a request for evidence within 15 days.
If your husband wants to accompany you to the U.S. to simply apply and interview for jobs while you scout for office space, he can enter on ESTA or apply for a B-1 visa as well. The B-1 and the B-2 visitor visa for pleasure are issued together, so it’s crucial that you and your husband let the U.S. immigration officials know, particularly at the airport, that you will be conducting business while in the U.S. Failing to do so may put your ability to stay in the U.S. and any future visas and green cards at risk.
The spouse and dependent children of L-1A visa holders are eligible for an L-2 visa. As an L-2 visa holder, your husband will be eligible to work. Since 2021, individuals who have an L-2 visa no longer have to apply for or renew their employment authorization document (EAD), otherwise known as a work permit. The USCIS will issue Form I-94 listing “L-2S” when granting your husband a status change to an L-2. That’s considered equivalent to an EAD card and it’s a great benefit for families!
The maximum stay in the U.S. on an L-1A visa is seven years. The L-1A offers a path to the EB-1C green card for multinational executives and managers. The requirements for the EB-1C are similar to those of the L-1A: Your company must sponsor you, and you must have been employed in the U.S. as an executive or manager for at least one year.
O-1A is an option, but . . .
If the L-1A is not an option for you, we’ve had a lot of success helping founders get an O-1A extraordinary ability visa. But keep in mind that unlike the dependent spouse of an L-1A visa holder, the O-3 dependent spouse of the O-1A visa holder is not eligible to work. However, your husband can work if he finds a job with his own employer willing to sponsor him for a work visa.
If you pursue the O-1A, it’s easier for your company to qualify for an L-1A, but the overall bar for your accomplishments is higher. However, I often find that most startup founders with a product, perhaps some funding, and some initial traction can easily qualify. To qualify for the O-1A, you must demonstrate at least three of eight criteria, such as receiving international or national awards; exclusive, invitation-only membership in organizations; and being featured in professional, trade or major media. Check out this previous Ask Sophie column in which I dive into how to meet each of the eight O-1A criteria. Premium processing is also available for the O-1A.
Because the EB-1A extraordinary ability green card has many of the same criteria as the O-1A, the O-1A is a fairly easy reach to an EB-1A or some founders pursue an EB-2 NIW if their wait time is acceptable.
E-2 is also an option, but . . .
The E-2 treaty investor visa enables international founders whose home country has a trade and commerce treaty with the U.S. — as Germany does — to live and work in the U.S. while investing substantial capital to build a business here. (The U.S. Department of State maintains a list of treaty countries.) But keep in mind that at least half of your U.S. business must be owned by people or companies from your country of citizenship to maintain E-2 status, which gets tricky particularly if and when your startup begins raising funds.
Although the E-2 requirements don’t specify how much capital you must invest to build your U.S. business, immigration officers look for large investments in office space, equipment and inventory, somewhere in the $100,000 range. That can make it difficult — but not impossible — for startup founders to qualify for the E-2. While the E-2 does not specifically require job creation, immigration officials may consider your U.S. business to be too “marginal” without it.
Another major factor is that the E-2 visa application process occurs directly at the consulate, and there is no option for premium processing if you are seeking a multiple entry visa in your passport.
The spouse of an E-2 visa holder is eligible to apply for an EAD. Like the L-2 visa holders, E-2 dependent visa holders automatically have work authorization with their visa and will receive a Form I-94 that serves as proof they are authorized to work.
Like the O-1A, there’s no limit on the number of times the E-2 visa can be extended. However, for the E-2, immigration officials will want you to demonstrate that you still maintain a residence and ties to your home country and intend to eventually return there. This is called non-immigrant intent, and immigration officials will want to see that you do not intend to and have no desire to remain in the U.S. permanently. In contrast, the L-1A and O-1A visas allow you to pursue a green card (permanent residency).
Immigration officials heavily scrutinize both the L-1A and E-2 visa applications, so I want to gently remind you how important it is to work with an immigration attorney to present a strong case whatever route you decide to take.
Enjoy your journey!
— Sophie
Have a question for Sophie? Ask it here. We reserve the right to edit your submission for clarity and/or space.
Sophie Alcorn, founder of Alcorn Immigration Law in Silicon Valley, California, is an award-winning Certified Specialist Attorney in Immigration and Nationality Law by the State Bar Board of Legal Specialization. Sophie is passionate about transcending borders, expanding opportunity, and connecting the world by practicing compassionate, visionary, and expert immigration law. Connect with Sophie on LinkedIn and Twitter.
Sophie’s podcast, Immigration Law for Tech Startups, is available on all major platforms. If you’d like to be a guest, she’s accepting applications!
How this VC evaluates generative AI startups
The launch of ChatGPT in November of 2022 propelled our world into the Age of AI, and the tech industry will never be the same.
Nearly every pitch deck I’ve seen since December has had AI on the front two pages.
As with any emerging technology, however, venture capitalists like myself have had to quickly develop a strategy to separate the high-potential startups from those that are mostly hype or are likely to face insurmountable challenges that will prevent them from achieving venture scale.
Understanding that distinction requires fluency in the various layers of the generative AI value stack, determining which are ripe for investment and creating a due diligence strategy to evaluate the risks and opportunities of a given startup.
Specifically, generative AI is composed of:
- Data.
- Middleware.
- Fine-tuned specialized models.
- The cloud and infrastructure layer.
- Foundational models.
- The application layer.
Within this tech stack, there are a few areas that we think are especially investable and others that are more challenging for a seed-stage company to compete in. Here’s how we break it all down.
Areas we’re interested in
Data
One of generative AI’s greatest challenges — and thus one of its greatest areas of opportunity — is the accuracy and reliability of the information it provides. Today, generative AI models are built on massive datasets, some as wide and as broad as the internet itself, containing both relevant and useful information, and a whole lot of everything else.
We believe that the galaxy of generative AI applications that will emerge in the coming years will be composed of more precise data, or bits and pieces of different, more specialized models. Rather than casting a wide net, these specialized models will utilize proprietary data specific to a domain, which will help to personalize the output of the application as well as ensure accuracy.
There are a few areas that we think are especially investable and others that are more challenging for a seed-stage company to compete in.
Having proprietary data to infuse with foundational models — combined with the right middleware architecture — will result in these specialized models, which we believe will power the application layer that consumers and businesses interact with.
Middleware
Accompanying the data layer of the generative AI stack is middleware, which we define as tooling and infrastructure that supports the development of new generative AI applications and is the second part of our investment thesis in the sector.
Specifically, we are bullish on infrastructure and tooling companies that evaluate and ensure safety, accuracy, and privacy across model outputs; orchestrate inference across multiple models; and optimize incorporating proprietary data into large language models (LLMs).
How any SaaS company can monetize generative AI
If you work in SaaS, you’ve likely already been part of a conversation at your company about how your customers can benefit with increased value from your products infused with generative AI, large language models (LLMs) or custom AI/ML models.
As you hash out your approach and draw up the product roadmap, I wanted to call out an important aspect — one that I couldn’t help but draw an analogy to the good ol’ California Gold Rush. Don’t show up to the gold rush without a shovel!
Similarly, don’t overlook the monetization aspect of your SaaS + AI. Factor it in at the outset and integrate the right plumbing at the start — not as an afterthought or post-launch.
Two years ago, I wrote about the inevitable shift to metered pricing for SaaS. The catalyst that would propel the shift was unknown at the time, but the foundational thesis was intact. No one could have predicted back in 2021 that a particular form of AI would serve to be that catalyst.
SaaS + AI — what got you here won’t get you there!
First thing to realize is that what is required is not merely a “pricing” change. It is a business model change. Traditionally, SaaS pricing has been a relatively lightweight exercise with a simple per seat model and a price point set sufficiently high above underlying costs to attain desired margins.
Don’t show up to the gold rush without a shovel!
A pricing change would be a change in what you charge; for example, going from $79 per user/month to $99 per user/month. A monetization model change is a fundamental shift in how you charge, and with AI as a consumption vector, it inevitably requires a need for accurate metering and usage-based pricing models.
There’s already a handful of great examples of companies leveraging usage-based pricing to monetize AI, including OpenAI and all companies that provide foundational AI models and services, and the likes of Twilio, Snap, Quizlet, Instacart, and Shopify that are integrating with these services to offer customer-facing tooling.
Why usage-based pricing is a natural fit for generative AI
One challenge of monetizing generative AI is that the prompts and outputs vary in length, and the prompt/output size and resource consumption are directly related — with a larger prompt requiring greater resources to process and vice versa.
Adding to the complexity, one customer may use the tool sparingly while another could be generating new text multiple times daily for weeks on end, resulting in a much larger cost footprint. Any viable pricing model must account for this variability and scale accordingly.
On top of this, services like ChatGPT are themselves priced according to a usage-based model. This means that any tools leveraging ChatGPT or other models will be billed based on the usage; since the back-end costs of providing service are inherently variable, the customer-facing billing should be usage-based as well.
To deliver the most fair and transparent pricing, and enable frictionless adoption and user growth, companies should look to usage-based pricing. Having both elastic front-end usage and back-end costs position generative AI products as ideal fits with usage-based pricing. Here’s how to get started.
Meter front-end usage and back-end resource consumption
Companies leverage prebuilt or trained models from a plethora of companies and may further train them with their custom dataset and then incorporate them into their technology stack as features. To obtain complete visibility into usage costs and margins, each usage call (be it API or direct) to AI infrastructure should be metered to understand the usage (underlying cost footprint).
4 steps founders can take today to improve team recognition tomorrow
A talented team is arguably one of the most valuable assets of any company. For startups and smaller enterprises, this statement usually rings even truer.
But it takes plenty of time and effort to recruit the right people. And, once on the team, making sure that your talent sticks around is an even bigger challenge. While some degree of staff turnover is natural, it takes time and energy for new hires to learn the specific ropes of your company, build relationships with colleagues and begin delivering results. If these new hires leave, it not only puts unnecessary strain on your operations but could also point to a larger issue.
Retention initiatives often look at how competitive compensation rates are, or whether work-life balance at the company could be improved. Employee recognition is often left out of the conversation but it is something that has a significant impact on any retention strategy. That’s because 36% of employees cite a lack of workplace recognition as the main reason for quitting, with 60% being more motivated by recognition than money.
And it’s not just retention rates that are at risk, but the performance of those employees who stick around. Research shows that employees on the receiving end of fair and consistent recognition from their leaders generated twice as many ideas per month as those who don’t receive the same close attention.
In a tight labor market, every founder can benefit from taking a look at how the individual achievements of each team member are currently being celebrated and being honest about where there may be room for improvement. Here are four ways that founders can strategically improve employee recognition to better retain their best performers and boost motivation across their companies.
Share your network with star performers
Founders and senior executives are often overwhelmed by requests to act as speakers or mentors at industry events. At times, it may be appropriate to consider sending a rising star from the company in your place. This not only helps to increase company representation at such events but also offers a highly valuable way to recognize your top performers.
For employees at an earlier stage in their career, getting named as a board advisor or mentor is an accolade that adds prestige to the individual’s résumé while also validating the presence of the company.
It’s important to communicate exactly why a certain employee has been chosen to represent the company.
For example, the former co-CEO of Salesforce, Bret Taylor, was also chairman of the board of Twitter in addition to his role at Salesforce, which helped draw attention back to Salesforce.
Equally, speaker roles at industry events are highly valuable opportunities for aspiring leaders, helping them to fast-track their visibility among peers and forge valuable connections on behalf of themselves and the company. Rather than hoarding such opportunities, leaders who choose to share the benefits of their established networks with star employees for the greater good are set to win in the long run.
Here it’s important to communicate exactly why a certain employee has been chosen to represent the company for such opportunities. In this way, other team members will know that it’s in relation to a specific achievement rather than a case of favoritism and gives others something to aspire to in the future.
Don’t limit recognition to internal comms only
Many managers just aren’t good communicators and are often what could be called “emotionally stingy.” Although employees report a general lack of recognition, it usually does happen in some form. However, in my experience, smaller wins and general team achievements are limited to a shoutout during company meetings or buried within performance reviews.