Insights Into AI Investing From Alacrity Global

Artificial Intelligence (AI) is a strong trend today in terms of technology innovation, start-up company focus and investment firm interest. Much hype can be found in the typical investment pitch-deck being circulated. But the term “AI” has become so common, even overused, it can often be misleading and unnecessary as a value-add to these materials. It does not always help in assessing the value of an early stage company.

The commercial market for AI is still in its early days, but it will no doubt continue to develop across multiple business sectors, and in some cases do so rapidly. However many companies exaggerate their AI proficiency and product capabilities. A clear understanding of the innovation and disruption potential of the technology is a core element driving investor interest, particularly in an emerging field such as AI. The complexity of the AI landscape continues to grow and the field is sub-dividing into different areas (Advanced AI and more advanced Analytics, for example). Therefore, different approaches are used for determining AI technology and start-up value, and these are blended with traditional valuation criteria of company performance to date, customer adoption, churn rates, leadership and team attributes, etc. The following are some examples.

Class of AI

AI Class is a means of categorizing the focus of business activity for the start-up. It helps simplify the AI landscape somewhat, suggesting the way in which the technology will used by customers. It also points to the way that tech­nology will be acquired by customers, and therefore the way in which it will need to be brought to market. Here are some general classes:

Business Process Class

AI used to enhance business operations — by automat­ing manufacturing processes, for example. This form of AI is used primarily in an organizational support capacity, to lower costs or increase efficiencies.

User Class

AI used to increase business user or team effectiveness — by analyzing customer data in real time to improve an offer to a retail customer, for example. This type of AI is inte­grated into existing applications or offered incrementally, extending user capabilities, speeding processes or improv­ing results generated.

Platform Class

AI offered as tools to build custom applications for or by a customer — to create a reseller portal, for example. This class of AI is marketed as a tool set, including architecture frameworks, best-practices, etc. and may or may not be supported by the vendor with professional services.

Classes of AI are not mutually exclusive. A company may offer a hybrid product or service but each class helps frame the nature of the target customer, the market opportu­nity, the competition to be expected, the cost structures to be planned for and other factors that will influence the value and viability of a new AI company or product. One AI class should not be rated higher or lower than another in isolation, nor should it suggest a strong or a poor invest­ment (i.e., there are market opportunities across all these classes). The final preference, if there is one, will be more a reflection of the investment group objectives.

Market for the AI

The next step in assessing the value of an AI start-up is a deeper dive into the dynamics of the target market, and that all-important competitive landscape. For example, it is important to get a clear understanding of:

  • Whether the AI solves a real business problem, and if so, does it do so more effectively and/or more cost efficiently than existing solutions (i.e., is it a “need to have”, or merely a “nice to have”)?
  • If it does this, how widespread is the business problem that is being solved? How many true customers are there for the new solution, and where are they (local, global, one industry, across many industries) — what is the Total Addressable Market (TAM), over what period of time?
  • If the answers so far are solid, the assumption needs to be that others have seen this opportunity too. Who are they, where are they, how mature are they, how deep are their pockets or those of their backers, etc.? And again, what will be the competitive differentiation for the new solution?
  • What are the barriers to entry for new players looking at this same space? Is there valuable Intellectual Property (IP) in the new solution that will maintain differentiation for the start-up? Is that IP protected, by patents for example? More on this in a moment.

This list could go on easily. And while these questions apply to any technology assessment, they are particularly relevant when talking about an emerging innovation such as AI, to avoid getting caught in the early-market hype and assuming undue risk in the investment.

AI Assets in the Organization

Depending on the degree of technology development, the start-up will have a level of AI ‘assets’ that can be fac­tors in a final valuation process. The deeper the company is embracing AI, such as those in the AI Platform class, the more value these assets may hold. For example:

  • As the AI field has limited specialists or experts to date, is there a team of AI developers and other professionals managers in the company? Are they committed believers in the direction of the company and product? If so, that can represent considerable value. If not, what is the source of the AI expertise in the firm? Are there external subject matter experts and training materials available, and are they committed, repeatable, etc.?
  • To the earlier point, what is the extent of the Intellectual Property under ownership? Are there patents under way or in place? Is the technology/ solution set differentiated and protected/defensible? Do these assets create barriers to entry for competitive solutions? Being able to answer ‘yes’ to as many of these questions as possible plays well into valuation discussions. On the other hand, if the solution relies on outsourced Machine Learning algorithms, for example, it does not create a stronger case for valua­tion when those same AI assets are available to, and under the control of, others in the market.
  • Who owns the data the AI product is leveraging, and under what policies, permissions and rights is it being used? From whom/how is it being sourced? Clean data is a critical element of AI application functionality. By extension, this makes data and data sources valuable assets for a company.

Resulting AI Score Card

Based on the preceding, determining an AI investment valuation can be more subjective than objective. The goals, timelines and priorities of the investor can carry more significance than in other technology due-diligence exercises. But in an effort to take more of the guess work out of the process, the following score card is an example of a tool that can help capture and rank the resulting data during an assessment.

Company, AI Class Company 1 Company 2 Company X
Market Disruption      
Data Resources      

Subjective Ratings: 1 = minimal value; 2-3 = marginal/average value; 4 = clear added value; 5 = best in class

Closing Thoughts

Artificial Intelligence is in the early stages of commer­cial application, and given the technology and human resources needed to become a fully established AI com­pany, it is likely a start-up has several more steps to take on that journey. It is also likely they are using AI technol­ogy either sourced openly or developed elsewhere to augment a software product rather than craft something completely innovative, and there is nothing wrong with that. The assessment process outlined in this article is merely one starting point for investment discussions. And all of these factors will continue to evolve and mature as the technologies and markets inevitably do as well.