The AI Hype: It’s all about Domain Specific Applications – Financial Services

This article is co-authored by Sanjay Gandhi + Ryan Janssen.

The responses to our The AI Peak Hype post have been both polarized and engaged! We’ve had a number of requests for more detail and real-life examples; here, we’ll do just that.

  • To recap the previous post and provide the setting for the next phase of commentary:
  • AI Peak Hype core arguments:
  1. The current technology for AI isn’t there yet.
  2. Current “breakthroughs” in AI are operating in very specific domains.
  3. Even if progress exceeds expectations, there is no way it can live up to the hype.

What should an AI entrepreneur do?

  1. Raise fast – hype cycles tend to crash fast too.
  2. Have a real business.
  3. Utilize proven technologies (e.g. conventional neural networks and LSTM).

What should an AI investor do?

  1. Focus on the next-gen of AI technologies (specifically, reinforcement learning + adversarial networks).
  2. Avoid AI solutions looking for a problem.
  3. Stay domain specific.

The overarching theme from our first piece is that domain specificity is key in understanding current strengths in AI and how to think about potential commercial applications that can develop into viable business models. More overarching approaches are still very much best left to academic and research disciplines. Companies and investors need to keep their heads focused on very specific domain applications that have the ability to add value and take advantage of well understood AI dynamics. To that end, we’ve focused this follow-on post to very specific domain and sub-domain applications of AI and dissected what’s on the cutting edge, what’s on the “bleeding edge” and what remains pure hype.

In looking at specific domains where AI shows promise, one feature to consider is the extent to which the industry has invested in technology to drive productivity gains and the level at which the industry has become digitized—meaning specifically that activity has shifted towards digitized formats and transactions handled by computers. This creates data, oftentimes reams of it, which gives AI an opportunity to come in, “eat the data” and automate decisions or do things more efficiently or intelligently.

The financial services sector is an area which has consistently been one of the biggest investors in applied technology spending. Most trading in the world has shifted to the electronic realm and could no longer take place without computers/digitized data and investments in machine-driven trading processes have accelerated over the past decade—a portion of which was highlighted in Michael Lewis’ book Flashboys which recounted the world of high-frequency trading. Trading is an area where vast amounts of data, both current and historical, are accessible to prime AI learning algorithms. In addition to trading, numerous aspects of the financial sphere have become increasingly data and algorithm dependent, from insurance to lending to traditional corporate finance. And, separate from customer-facing/trading situations, financial institutions have also invested in technology to handle increasing amounts of internal processes and functions, which provides further opportunity for AI-driven applications.

We’ve chosen four areas of the Financial Services sector for examples. In each, we’ll highlight some leading AI companies in the subsector, discuss what we think the technology can achieve now and in the near-term, and what we feel is further out.

Subsector 1: Insurance

The cutting-edge: Chatbot-powered user interfaces. For example, Lemonade, the NY-based insurance startup, has raised a huge war chest and is using tech to better the slow world of insurance. One of its key innovations is the one-line chatbot interface for handling everything from customer service to claims. This is in place and working today.

The bleeding edge: Actual AI-led chatbots. Most of that interface at Lemonade still has a person behind it. However, natural language processing, powered by convolutional neural networks, continues to become more effective. We think it’s safe to say scripted chatbots will be leading insurance interfaces within months or years.

The hype: Image recognition for claims handling. Some companies are testing this now. For example, multinational insurer Agaes is using technology developed by fintech startup Tractable. The insurer is leveraging technology that seeks to analyze images of cars which have been involved in accidents and make a decision whether a claim is valid, typically within seconds. There’s a lot of optimism that just snapping an image will allow instant confirmation of claims. But this is a very high level of abstraction. It will be years for image recognition to advance from distinguishing chihuahuas/muffins to actually making high-level (and high-value) judgment calls.

Subsector 2: Asset Management

The cutting-edge: Alternative data. Smart companies are using big data technology to consolidate, process and ultimately supply data that professional investors use to make decisions.

The bleeding edge: AI powered stock-picking. It’s no secret that the machines are eating Wall Street and investing better, faster and a greater scale than traditional traders. New York’s own Numer.ai is a tech-take on quant investing, creating a crowd-sourced, AI-powered hedge fund. Generally speaking, some of the most common techniques used by AI-powered funds are multivariate regression and supervised machine learning. So far, it’s unclear if quant funds are over or underperforming vs. traditional funds, but the trend is strongly towards AI-powered investing (and soon).

The hype: Generalized anything. AI stock picking works well because AIs perform better on narrow domains. For instance, it’s (without hyperbole) millions of times easier for Facebook to correctly identify a photo if it only needs to choose from your several-hundred friends. Stock picking is a narrow domain by itself—developing cross asset macro hypothesis about the world is not. So while machines are already beating humans at certain narrow tasks, we’ll need some general intuition running the bots for a long while.

Subsector 3: Lending

The cutting-edge: Predictive credit scoring. There’s no shortage of companies using AI to provide an improved version of the credit score for both individuals and businesses. Some examples include Zest Finance, Underwrite.ai and Demyst Data. This is actually quite hard to do (the technical challenge is mastering the design of the very subtle features that predict credit risk), but in our opinion AI solutions are now doing a more accurate job than traditional credit scoring techniques. Expect to see much more of it in the very near future.

The bleeding edge: Expanding beyond scoring. Some of the world’s best consumer lenders are using AI to decide what action is most cost-effective if you miss a payment.

 The hype: A human-free loan process. We see isolated examples of this, and a lot of boasting that it will be happening soon. But in practice, there’s too much at stake to relinquish the human underwriters any time soon.

Subsector 4: Corporate Finance

The cutting-edge: Natural-language processing of financial data. This essentially enables users to ask questions in plain English that can establish connections, or correlations (obvious or obscure), that can influence financial markets. For example, “When Netflix beats earnings, how do Amazon shares perform the next day“, or “Which Apple supplier’s share price goes up the most when the company releases a new iPad?” (answer: OmniVision which makes the sensor in the iPad camera). Kensho, a Cambridge-based AI startup with a long list of top-tier investors including Goldman Sachs, NEA, Google and Accel Ventures, has staked a leading role in this field. It sells itself as giving “the masses” access to the kind of computing power heretofore reserved for the likes of top hedge funds like Bridgewater Associates and Renaissance Technologies.

The bleeding edge: This is the perfect example of a Centaur at work— the combination of human mind and machine analytical firepower, also called “AI-human hybrid” intelligence. As companies like Kensho continue to increase in efficacy, we’ll see a number of the typically time-consuming analytical roles be automated and enable greater productivity.

The hype: The deal machine. There’s been some rumbling about AI being able to identify good M&A targets, etc. This is admittedly closer than the other “hype” sections in this article, but we’re still talking broad domain, generalized AI to achieve things like this with anything close to human-level precision.

Let’s not lose sight of what AI is. It’s really just statistics. The principles being applied aren’t that much more advanced than what you learned in STATS 101. They’re just applied iteratively and at scale.

Are there some groundbreaking engineering achievements being made? Absolutely. But it will be quite some time before we figure out how to assemble a bunch of linear regressions into stronger forms of intelligence. And then there is the issue of whether or not machines can achieve “consciousness.” …we’ll leave that to another day.

For more information contact Sanjay Gandhi at sgandhi@oxfordvp.com. Look for more funding advice by visiting the knowledge section of the Oxford Valuation Partners website.

Disclaimer: The companies we’ve chosen here are just illustrative examples of leaders in their field.  We don’t have any contact with them or non-public information on how they operate.



5 Ways Technology Will Change the Manufacturing Sector

This article was originally published here on October 17, 2017.

Accounting for almost 12% of the United States’ GDP and 9% of the workforce, there is no denying that the manufacturing industry is massive – $2.18 trillion to be precise. Additionally, a recent Forrester report projects the sector’s budget for technology will increase to 2% of revenue in 2017, thus putting overall spend at more than $40 billion for the year. Given the undeniable influence the industry holds over the economy as whole, we’ve decided to take a look at the top 5 ways technology will change the manufacturing sector over the next few years.

1. 3D Printing

3D printing, also known as additive manufacturing, is defined by ASTM International as “an additive technique that uses a device to create physical objects from digital models.” Previously considered only a niche market due to technology constraints, new developments in hardware and software are pushing 3D printing into the mainstream. Gartner projects that by 2020, 10% of industrial operations will incorporate 3D printing into their manufacturing processes, 75% of manufacturing operations worldwide will be using 3D printed tools or objects in some form, and total spending related to 3D printing will grow 66.5% CAGR to reach $17+ billion.

Why is this technology so impactful? By providing the ability to print customizable, physical objects on the spot, companies will produce parts and products closer to the location of purchase, severely reducing the cost of shipping and a need for extra inventory. This decentralization of production could mean the end of low-cost manufacturing labor advantages, altering international supply chains and the global economy.

2. Internet of Things

The manufacturing sector has already invested heavily in the Internet of Things and those numbers will continue to grow at a fast pace. BI Intelligence suggests the number of IoT manufacturing devices will jump from 237 million installed in 2015 to more than 900 million installed by 2020. In dollar terms, that’s an approximation of $70 billion in spending on IoT.

Why is the industry so excited? The bottom line is the technology will change the manufacturing sector by making it smarter. Through connected IoT devices, manufacturers will be able to collect massive amounts of data that can be be analyzed (discussed more below), insights can be extracted, and improvements can be made by pushing actions out through the IoT ecosystem. That sounded pretty vague so let’s run through an example: predictive maintenance. Whereas preventative maintenance – which is widespread today – relies on scheduled checkups of machinery based on a specific time period (ie. every x number of days), predictive maintenance relies on real-time data captured and transmitted by IoT devices to determine the exact moment when maintenance should occur. This can result in major cost reductions because a) unnecessary maintenance is costly, and b) failing to provide maintenance when necessary is also costly. By understanding the real-time performance of a machine and whether it’s within certain acceptable limits, IoT ensures that measures are taken when needed – no more and no less. This is just one of many examples of how IoT technology will change the manufacturing sector.

3. Machine Learning

We’ve already covered the troves of data now being captured and transmitted through connected IoT devices. But to continue with our previous example, how can companies analyze this large amount of data and turn it into meaningful insights for workers?

That’s where machine learning comes in. Machine learning software, such as that produced by Bowery Capital portfolio company Fero Labs, analyzes the data collected from factory sensors, enabling non-technical workers to discover low-performing metrics, identify the causes of the poor performance, and ultimately, make predictions about future performance.

This is just one example of how machine learning technology will change the manufacturing sector. Other applications may include price optimization, demand forecasting, increased production yield, improved quality, and more.

4. Edge Computing

When dealing with thousands of connected devices, sending large amounts of data to a centralized infrastructure (which is prevalent in today’s world) can be both costly and time-consuming.

In the manufacturing world, real-time analysis can make the difference in preventing a breakdown in production, and thus a major headache and loss of revenue. Edge computing allows the necessary analysis to take place through data processing power at the edge of the network, closer to where the data is collected – as opposed to undertaking the time and costs needed to transfer the data to a central warehouse – so that insights can be generated faster and the resulting actions can be taken in time to achieve the necessary goal.

5. Robotics

Finally, we come to robotics, which has already made its mark and will only continue to grow in size. Between 2000 and 2010, there was a loss of over 5.5 million manufacturing jobs in the US; 85% of those were due to automation – mostly, robots. Despite this loss of jobs, the manufacturing sector is producing more real value than ever before and at lower costs. It’s no wonder then that according to a recent PwC survey, almost 60% of manufacturers are already employing this technology in some way.

Whereas previously robots were used mostly in isolated circumstances and for complex projects, developments in technology have allowed them to become more independent, smart, and cooperative. It’s not hard to see a near future where connected robots communicate with each other and smartly adapt their actions together as needed to whatever the circumstances call for, leading to even higher efficiencies and cost gains.

If you liked “5 Ways Technology Will Change the Manufacturing Sector” and want to read more content from the Bowery Capital Team, check out other relevant posts from the Bowery Capital Blog. Special thanks to Jake Kupperman for helping contribute to this post!

Navigating the Future of IoT (Panel Discussion)

This podcast was originally posted here on August 23, 2017.

UILabs recently hosted a panel of IoT experts to discuss the future roadmap for IoT.  I moderated the discussion and the panel included:

  • Ben Forgan, CEO of Hologram
  • Brenna Berman, Executive Director of City Digital, UI Labs
  • Jenny Fielding, Managing Director, Techstars IoT
  • Jim Gagnard, CEO of SmartSignal (acquired by GE) and Chairman of the Illinois Technology Association
  • Ty Findley, Senior Associate at Pritzker Group Venture Capital

During the discussion, we addressed the following:

  • The “inning” or stage of maturity that IoT is at
  • The smart home and what the winners will look like in the space
  • How investors assess consumer vs enterprise IoT and suggestions for entrepreneurs with applications for each
  • Investor thoughts on the metrics they evaluate for IoT vs. other horizontals like SaaS
  • The key opportunities at different levels in the IoT tech stack
  • The key barriers to mass IoT proliferation and what’s limiting IoT
  • Thoughts on security for IoT and the role of government
  • Wrap up w/ Q&A from audience members

Listen to the podcast

How to Hire the First Salesperson for Your Startup

This article was originally posted here on Early Growth Financial Services’ blog.

Few things are as important or as challenging to startups as landing that first salesperson to take your vision forward. Often you start as a solopreneur or as a small inner circle of founders wearing all the hats necessary to keep your firm lean and aggressively hungry. As you grow and scale operations, it becomes necessary to hire a strictly sales-focused team member.  It can be the thing that really lets you shine and concentrate on the vision forward, or it can become the internal bottleneck that pauses everything.

Early Growth Financial Services is a company that aids startups in their financial management and actively advises its clients through the myriad of challenges during the rapid expansion years.  Therefore, we have been fortunate to see some of the best practices, and as seasoned pros, we have the intuition to know what is needed. There are some basic ideas and reflective questions that will aid you in setting your course. The essential goal when taking on new people is maximum ease and minimal challenge, and these practices can help with this.

Define the level & style

For starters, you want to consider where your firm is at in sales.

Is the company still experimenting and fine tuning its offering or is the product ready to go? Will the sales person need to work internally to make those changes? Are you still experimenting with who is the ideal client, or are you clear on who your leads are and what your market is?

A salesperson needs to know what they’re selling and who they are selling to. If your offering has been through revisions, especially significant ones, it is important to make sure your sales team can appreciate those changes and reinforce the new offering as the best choice. Previous clients may need to be resold and re-evaluated, or new markets may need to be explored. Mapping out how much of the role is servicing existing sales, how much is growing new clients, and how much is raw business development is crucial.

Are you best served by someone with B2B experience or someone with a B2C sales background?

Style of communication can vary greatly in sales. What works best for those who service consumer needs can be problematic if you are a B2B focused entity. The reverse is true as well. Having the mindset and dialogue abilities for personal versus bureaucratic corporate speak makes a world of difference. Conversation is the beginning of everything, so make sure your salesperson can talk the right talk.


Your initial hire in sales needs to be comfortable with certain challenging aspects of lead generation, client management, and enthusiastic engagement. The ideal candidate should be eager, hungry, open, and willing to try new things. They must demonstrate that they are not afraid to cold call and can be ready and ambitious while doing it. The right candidate may be straight out of college especially if they are experienced in cold calling and have shown an eagerness to build the sales skills needed.


This is the internal evaluation of your success as a company that will be important in guiding your search.

Have you raised a seed round or are you Series A and beyond?

How deep your pockets are and, if applicable, what burn rate you are at will give you a strong start in picturing what kind of salesperson you are ready to take on. With your first hire in this area, you will want to be sure to minimize any false starts and do overs.

Compensation can vary widely from a fixed salary to totally variable pay based upon results. A good rule of thumb, in either case, is for the salary to be commensurate with the earned MRR (monthly recurring revenue) they bring in.

Following these guidelines, focusing in on quality and the cost to benefits ratio, making that successful first sales hire should be achievable. If you have additional tips we would love to hear them.

Evaluating SaaS Startups As Potential Vendors

This article was originally posted here on September 12, 2017.

Today, even the largest enterprises are working with startups. While F500 procurement processes are still a tough go for young businesses, corporations are recognizing that adopting cutting edge technology early—while a potential risk—is a form of innovation in its own right. Inking a good deal with a new best-of-breed vendor years earlier than the competition can go a long way. That is why evaluating SaaS startups as potential vendors is important. “Buy,” moreover, is simply going to make sense more often than “build” in a world where software touches every business function under the sun. Accordingly, enterprise stakeholders are both democratizing the SaaS buying process and finding ways to test early-stage products org-wide in a more agile way.

However, working with early-stage vendors still presents real risks. Security can be some of the most concerning. Luckily, a range of standards and testing procedures have long been in place to gauge vendor security. Corporate lab environments and gradual rollout plans designed to limit initial exposure to things like data loss, improper PII handling, and backdoor system breaches have been commonplace for years. I find that there is another type of risk that is just as common, yet much more nebulous and hard-to-gauge: the general “maturity” of a vendor. Different companies and IT buyers have different ways of measuring it: age of company, financing raised, size of reference customers, or even benchmarks based on financial metrics. While they see startups as a potential source of innovation, more F500 companies than not miss out on opportunities to work with startups because they have outmoded or overly rigid definitions of what constitutes a vendor that is sufficiently “mature” to support their business.

So how should businesses best think about evaluating SaaS startups as potential vendors or partners? I think the question of whether the vendor is solving a big problem is much more important than spending a ton of time probing financial stability. Of course there are baseline questions you might ask, like which or what type of other customers they have. You can do basic research on the stage of the business or whether or not it has institutional backing. But weigh the value of the product to you vs. the effort it will require to put it in place once you give a verbal. A big benefit of SaaS is that it’s relatively easy to adopt. So while larger companies have understandable needs for reliability, I think more companies should view working with early stage SaaS vendors as a more of innovation in itself. If you draw artificial lines like “company must have $10MM in revenue” or pick your vendors exclusively from Gartner Magic Quadrants, you will be systematically years late in taking advantage of best of breed solutions.

Are there specific metrics, measures or vendor characteristics IT buyers should consider closely? Most “SaaS metrics” aren’t relevant from a procurement perspective, even if the startup would share them (which they won’t and shouldn’t). Basic financial metrics will be held equally close to the vest. Some more relevant measures or diligence points will be contract-specific: average length, discounting, pricing dynamics (as an account scales), billing terms, customer success/support resources, and SLA details. For more qualitative vetting, learn about or even talk to other customers.

Finally, IT buyers should remember: the biggest indicator of future success of a startup is the people involved. An even better indicator, of course, would be product-market fit in which the product is defensible, the problem is growing, and the market is big. But you can’t really know you have this until you’re there. Therefore, the best leading indicator is the quality team steering the ship. This rule holds true even when you’re evaluating SaaS startups as potential vendors. Look for trust and skill, not just in the CEO, but also in the people you’ve dealt with throughout the sales cycle. Look for a customer-first culture and an agile product and engineering team. In exchange for not churning, IT buyers in the era of subscription SaaS should look forward to continual product growth and evolution. So look for product momentum and a roadmap that suits your own needs. The very best teams continuously test and iterate to improve outcomes in all parts of the business, and early customers should be the biggest beneficiaries of this rapid growth and innovation.

If you liked “Evaluating SaaS Startups As Potential Vendors” and want to read more content from the Bowery Capital Team, check out other relevant posts from the Bowery Capital Blog.

Square 1 Bank Announces Credit Facility to MapAnything

Square 1 Bank, a division of Pacific Western Bank, today announced that it has provided a revolving line of credit to new client MapAnything, Inc., a leader in geo-productivity and intelligence solutions for business. The $6 million facility is available to support the company’s future working capital needs and complements its $33,100,000 Series B equity financing which closed in January 2017.

Headquartered in Charlotte, NC, MapAnything’s geo-productivity software for Salesforce and ServiceNow drives productivity across multiple use cases including sales, service and operations. Founded in 2009, MapAnything provides a suite of location-aware applications to drive revenue and productivity for field-based work, while tracking all business activities, decisions and procedures through the CRM workflow. MapAnything is backed by leading investors including Columbus Nova Technology Partners, Greycroft Partners, Harbert Growth Equity, Salesforce Ventures and ServiceNow Ventures.

“Building a high-growth SaaS business requires access to capital to fuel growth,” said Chris Rosbrook, chief financial officer of MapAnything. “Our partnership with Square 1 and this new credit facility will ensure we deliver the best possible shareholder returns while we continue to grow our business rapidly.”

“MapAnything’s software increases client productivity by enabling users to leverage their Salesforce data to schedule more appointments and boost revenue,” added Dhruv Patel, vice president in Square 1’s technology banking practice. “With an experienced leadership team and solid investor backing, MapAnything is poised for significant growth and success. We look forward to serving as their financial services partner during this exciting time.”

About Square 1 Bank

Square 1 Bank is a division of Pacific Western Bank, a Los Angeles-based commercial bank with over $21 billion in assets. A full service financial services partner to entrepreneurs and their investors, Square 1 provides clients flexible resources and attentive service to help their companies grow. Square 1 offers a broad range of venture debt, treasury and cash management solutions through offices in top innovation centers: Atlanta, Austin, the Bay Area, Boston, Chicago, Denver, Durham, Los Angeles, Minneapolis, New York, San Diego, Seattle and Washington, DC. Pacific Western Bank is a wholly-owned subsidiary of PacWest Bancorp (NASDAQ:PACW). For more information, visit www.square1bank.com.

About MapAnything

Combine today’s global economy with a mobile workforce, and you get a lot of moving parts. CRM alone can’t help you answer the critical business questions: Where is my business? Where do I need to go? Founded in 2009, MapAnything is an innovator and pioneer in Geo-Productivity Software. With more than 1,800+ customers globally, ranging from Large Enterprises to Small Business, we believe that “Where Matters.”

MapAnything is a Salesforce Gold App Innovation Partner, and a ServiceNow Technology Partner. They have received Ventana Research’s 2016 Technology Innovation Award for Location Analytics and been named a Customer’s Choice – Highly Reviewed App by users of Salesforce. To learn more, visit www.mapanything.com.


Media Contact
Square 1 Bank, a division of Pacific Western Bank
Dee McDougal

Even on the Way Out, Investors Deserve a Manager’s Empathy

This article was originally published by Industry Ventures here on August 24, 2017.

We were taught as children about the golden rule: treat others as one would expect to be treated. Sometimes we forget this rule in business—particularly in the venture business. While we could take a sentence like that in several different directions, today the focus is on limited partner transfers.

In many cases, nobody is at fault when a venture fund investor decides to move on. General partners should bear this in mind when dealing with LPs, whether the latter are in it for the long haul or headed for the exits. Indeed, they should treat LPs the same way they expect to be treated by portfolio company CEOs. That is, with the understanding that while the relationship they have is undoubtedly important, it is only one of a broad mix that must be managed in its entirety.

GPs set a high bar for founders and CEOs—they insist on accuracy in reporting, efficiency in communication, excellence in execution, and patience in sponsorship. While portfolio company managers are scrambling to scale and achieve their missions, fund managers are intensely focused on their own marketing, financing, operations, and product development efforts. Under the circumstances, VCs expect empathy from portfolio management teams, yet they may forget that the shoe is on the other foot in their interactions with LPs.

This is particularly pertinent in the current environment. Venture-backed companies are staying private or independent longer than in the past, heightening demand for alternative sources of liquidity. The result: a robust secondary market in leading private company shares, as well as increased trading in private equity and venture capital fund LP interests, which, as noted previously, amounted to more than $40 billion in 2016. Specifically, Campbell Lutyens advisory firm expects this trend to accelerate in 2017, with a mix-shift of more transactions being initiated from endowments and foundations—as well as from less traditional LPs like corporate pensions and insurance companies (see Figure 1.)

Figure 1. LP-led Transaction Activity by Type of Seller [1]

No Fault Transfers - LP Secondaries graph

The trend is not solely limited to LP-initiated transactions, as GPs are becoming increasingly accommodating of transfers. Indeed, Campbell Lutyens also reports 22% of 2016 transactions were GP-led secondary transactions (non-directs).[1] Despite this, some GPs haven’t quite gotten the message. Although many willingly accommodate LPs looking to sell their interests, others make matters difficult, to the point of dismissively shutting down any such request. While a lack of empathy can be related to the position a fund is in—an oversubscribed GP with a waiting list for new offerings can afford to be more aggressive than one managing a first-time fund that has yet to prove itself—it does not seem to be the primary reason.

In fact, an unhelpful response likely stems from the perception—or, perhaps, misperception—that such a move reflects nothing but a lack of confidence in the manager. In our experience, this is misguided. There are numerous circumstances where the desire to move on has nothing to do with a soured relationship or, for that matter, an investor default. In fact, there are a variety of situations where LPs have little choice but to exit—and where it makes sense for GPs to allow and encourage others to take on the commitment. We’ve detailed eleven of these below.

  1. Macroeconomic Change. Macroeconomic downturns and disruptions can impact GPs and LPs alike. While the real deal may lead to LP defaults (and another outcome we discussed here), deteriorating perceptions alone can spur investors to preemptively batten down the hatches and shift funds into less risky assets. We briefly witnessed such a development in early-2016, when markets anticipated a reversal of fortunes (that never came). Otherwise, some LPs, such as pension funds, will readily sacrifice long-term gains to avoid near-term losses.
  2. Portfolio Rebalancing. LPs may decide to free up capital for other purposes, regardless of how well (or poorly) a fund is doing. In fact, some may have little choice but to liquidate their interest because of other developments, including margin calls and the kind of idiosyncratic volatility that has been seen in the energy patch in recent years. While private equity and VC markets may appear to be holding their own, unrelated factors could force the hands of any number of LPs. Ironically, resiliency in PE funds could flag them as attractive sources of liquidity.
  3. Return Management. There doesn’t necessarily need to be problems elsewhere; too much of a good thing can also encourage LPs to take money off the table. Significant gains in a venture investment or a PE allocation can stimulate efforts, perhaps automatically, to pare down the exposure in question. Many pension funds and asset managers, as well as other types of LPs, have portfolio limits they must adhere to by charter or mandate. Under the circumstances, it is not unusual to see them selling winners and buying laggards based on strategic allocation targets.
  4. Change in Strategy. Organizational missions and investment mandates can change, instigating a shift in portfolio preferences. A major pension fund, for example, could decide to reduce its commitment to private equity and allocate resources elsewhere—due to the political climate or other factors at play in a particular administration. In the wake of such a decision, there may be a flurry of transfers and buyouts relating to interests the pension giant no longer intends to manage. In those instances where capital is sourced externally, LPs may find themselves at the mercy of others, regardless of any intention to stick with a VC investment.
  5. Change in Investment Personnel. The appointment of a new CIO or PE chief can spur a rejigging of relationships and capital deployments, often to the detriment of existing managers. LPs with strict allocation targets, for example, may need to free up capacity through transfers. Even when such constraints are not an issue, other factors can intervene. For one thing, the efforts of now-outgoing personnel may have played a key role in why a commitment was made; without that support, an LP may decide to go in a different direction. That said, LPs aren’t always the catalysts for change. In some situations, discriminating GPs may, from an investor relations perspective, view a transfer as a plus.
  6. Overstretched Teams. LPs may be short-staffed or lack certain resources. This can be the case at large family offices or other organizations where staff must juggle a wide variety of duties, including researching investments, traveling to company meetings, responding to capital calls, and performing due diligence on prospective new managers. Constrained by time, lean LP teams may seek to reduce the number of commitments they have or allocate a larger share of available capital to a smaller number of funds.
  7. Tail-End Funds: A fund’s residual investment exposure may not be enough to warrant an LP’s continued commitment. Some funds may distribute the lion’s share of contributed capital but be left with unrealized investments well into their ninth, tenth or eleventh years—or longer. In essence, what began as a broad-based portfolio commitment becomes a smaller, less impactful collection of long-tail positions. From an administrative standpoint alone, it can make sense to move such interests off the books, perhaps by transferring them to wholesale-oriented buyers.
  8. Regulatory Requirements. Financial institution and foreign LPs may be constrained by issues that don’t affect other investors. For example, new rules or regulatory policies might require the former group to reorient portfolios or dispose of holdings after commitments had been made. This was the case in 2015 and early-2016, when the Volcker Rule forced banks to pare down balance sheet exposure to risky assets. Still, while such changes can happen unexpectedly whenever politics or regulation is involved, those affected are typically given ample time to make the necessary adjustments.
  9. Mergers and Acquisitions. Changes of control or other M&A activity can spur corporate LPs to reduce or eliminate fund commitments. As with the personnel and investment-strategy shifts referred to above, an evolving enterprise may no longer have the desire or justification for maintaining certain interests. Alternatively, prospective targets may preemptively shed such investments to facilitate a merger, or do so after the fact based on the acquirer’s requirements.
  10. Facilitating Recommitment. When GPs are seeking to raise for new funds, existing LP interest-holders may consider exiting from older vintages to free up capital that could theoretically be allocated to the new offering. It’s worth pointing out while LPs may initiate such transfer activity, it could actually be in the GP’s strong interest to proactively offer a liquidity program like this. Such actions may also be taken at the fund level. For example, capital may be returned to LPs on a pro rata basis to boost the ratio of distributions to paid-in capital, or DPI, for marketing or other purposes (which is a topic for another article).
  11. Expiration of Fund Life. Some of a fund’s most exciting investments can take more time than originally planned to reach fruition, which can lead to GP requests for an extension. While LPs may grant extensions once or twice, excessively long harvest periods can create tension between a manager and LPs that have strict time preferences on the lives of the funds they own. In such a scenario, it may be in both parties’ interest to permit a new investor to come on board—alongside those with more accommodating holding-period thresholds—allowing the full value of the investment to be harvested.

In sum, there are any number of reasons why LPs (or GPs) may want—or need—to suddenly change course. While the rationale may well be liquidity issues or a souring relationship, it is quite likely that neither is to blame. Under the circumstances, there is little to be gained by making matters difficult. Indeed, an infusion of fresh blood could turn out to be just what the doctor ordered.

[1] Source: 2017 Secondary Market Overview Report by Campbell Lutyens

The Importance of Specialization

This article was originally published by Top Tier Capital Partners here on September 5, 2017.

We rely on specialists and generalists on a daily basis. When remodeling a bathroom, a general contractor with some experience in carpentry, plumbing and electrical work is important to get the project planned. When it comes to the execution of each task (laying new tile, wiring vanity lights, installing new plumbing) it’s important to call in a specialist to get the job done right. Specialization is an important facet of the world we live in, but one that sometimes gets overlooked by investors. In this blog, we discuss generalists and specialists in the private equity market.

Investment staff within large limited partners, including public pension funds, endowments, and foundations, often have to invest across several market segments. This commonly involves coverage of both the public and private markets. Regardless of each investor’s intelligence, it is nearly impossible for one person (or even a small team) to be an expert in the intricacies of each private market segment. As a result, staff typically are generalists, but they invest in specialists in order obtain the exposure dictated by their asset allocation models.

We see the specialization of the market in this way:

Importance of Specialization graph 1

Each group increases in specialization from top to bottom of the chart.

Pension funds, endowments, foundations and consultants are all generalists that have the responsibility of investing over multiple asset classes.  It is the responsibility for these generalists to determine the best portfolio allocation for their plan, and then rely on the specialists within each asset class.

Funds of funds are typically specialized in one asset class and as such they should have a higher level of skill in picking the teams and funds with a greater chance for outperformance. For example, Top Tier Capital Partners is a fund of funds focusing on venture. Investment firms are generally specialized within an asset class, and at times focus on a specific stage, sector, or geography.  At the individual venture capitalist or company CEO level, those individuals are experts in their field and are typically highly specialized.

However, we caution against funds being overly specialized such that they lose a broader perspective. Top Tier Capital Partners is a venture capital-focused investment firm, but we still spend time learning about trends in the broader private equity and public markets. These markets affect our industry, especially as they relate to exits of the venture capital-backed companies in our portfolios.

To examine the level of specialization within our portfolio we looked at the 187 primary fund commitments made by Top Tier since 2001. 98% of the managers in our portfolio are ‘venture-only’ (the other 2% may be identified as venture growth, or special opportunities). Within our portfolios we rely on funds that are further specialized, and 88% of the funds are sector experts (i.e. security, enterprise, consumer), 82% are considered stage specialists (i.e. seed, series A, growth), and 73% specialize in stage and sector (i.e. seed-stage enterprise). We believe specialization can be a key differentiator for fund managers.

Importance of Specialization graph 2

There is a place for both generalist and specialist funds, depending on the size and asset allocation model of the investor. We believe that as one gets closer to the actual investment, the more specialized the investor should be, but all investors should maintain a broad enough perspective so as to maintain realistic views.  As a fund of funds we seek to invest in individuals and funds who are experts in their fields and recognize that we are paying for their expertise as specialists.

The Case Against AI: Are We at “Peak Hype” and What Should Founders + Investors Do?

This article is co-authored by Ryan Janssen and Sanjay Gandhi and was originally published here.

The Sanity Check

Well, we got here. Everyone’s wondering when a robot will take their job. Every pitchdeck shows how the company is “AI Powered” with a vague flow chart of a neural net — whether it’s an AI business or not. And Elon Musk is already beginning the Battle with the Machines.

But all of this discussion is happening with very little regard to the actual state of AI technology. Skynet is not about to become self-aware any time soon.

Yes, the advances in AI and machine learning are groundbreaking – but they’re in very specific domains. Areas such as NLP, image recognition and voice have improved dramatically in recent years.

But the case for a reality check in AI start-ups is three-fold:

  1. The current breakthroughs still aren’t there yet. The formless super intelligence we fear still confuses Chihuahuas and muffins about 10% of the time. And even when image recognition works well, humans can still completely break it by adding a little bit of noise in the picture.
  2. Even the current breakthrough technologies are operating in very specific domains (i.e., supervised learning). Current AI technologies are made possible by massive amounts of training data and very smart supervised learning – basically a well-defined goal to seek for. For example, the ImageNet database provides a clear and pre-defined answer to the question “is this a picture of a cat?”  Unsupervised learning, designed to work with much less structured datasets, has been on the horizon for years now but has had little success to-date.
    For us to move to more generalized applications, we either require a much more sophisticated unsupervised model or many datasets to train supervised models. These will likely both come, but both will take time.
  3. Even if progress exceeds expectations, there’s still no way we can live up to the hype. We’re very optimistic on machine learning’s long-term potential. In the long-term, we’ll see it transform the world in many profound ways. But at this level of attention, near-term expectations are being dramatically inflated.

Have no doubt – we ARE in store for a trough of disillusionment. A hype peak this high will inevitably disappoint some people.

So what should I do – as an AI entrepreneur?

So in short, AI has exactly the kind of transformative power that fuels great tech startups, but you’ll need to make it through years of non-plussed markets.

  1. Raise fast. You won’t have the power of the hype cycle on your side forever. Make hay while the sun shines.
  2. Have a real business. The best thing you can do in any market is build a lasting, commercially viable business. Don’t get pulled into the hype, keep your head down, and keep building.
  3. Utilize proven technologies. Let the smart folks in academia work on the unproven stuff while you build a solid foundation with convolutional neural networks and LSTM.

And what should I do – as an investor?

  1. Focus on the next-gen of AI technologies. Unlike the entrepreneurs, your job is to see what’s on the horizon. Right now, the states-of-the-art are in reinforcement learning and adversarial networks.
  2. Avoid AI solutions looking for a problem. Generally, be skeptical of products that are only compelling if an embedded AI delivers superhuman results. Every great investment starts with a commercial case first.
  3. Stay domain specific. Some of the biggest successes to-date in AI are those who have either mastered a single domain (Clarif.ai, x.ai), or are building tools to support specific data science techniques (RapidMiner, Enigma, Trifacta).

Lastly, don’t get us wrong: we strongly believe that AI and machine learning have more potential for disruption than any other technology today, and will enable decades of productive innovation.  We just expect these developments will take longer than the media expects.  But the sooner everyone can get past the hype, the faster we can get down to building lasting AI businesses.

For more information contact Sanjay Gandhi at sgandhi@oxfordvp.com. Look for more funding advice by visiting the knowledge section of the Oxford Valuation Partners website.

© 2005-2017 Square 1 Bank, a division of Pacific Western Bank. Member FDIC. All rights reserved.