TODAY, all companies are tech companies — or at least they need to be. It doesn’t matter what your “real” business is; if you don’t hypercharge your supply chain, marketing channels and internal processes with digital and artificial intelligence (AI), you simply will not succeed in the 21st century. It’s really just that simple.
You will therefore do two important things to become a better tech company. First, you will organically hire more tech-savvy staff, and augment your internal digital teams with external service providers. But second, you will make acquisitions of small, but incredibly critical, tech teams that currently operate a business in the vertical or horizontal digital space that you desperately need to get better at. What follows is a brief guide to some important issues in doing these deals.
Let’s imagine you’re a major Canadian financial institution (but what follows applies equally if you’re a retailer, a manufacturing concern, a telco, a transportation company, or just about any other business operating in the North American economy). You have identified a real need for acquiring some AI expertise. And it just so happens you’ve found a young AI company (let’s call them the Target), with seven staff, who have just launched an AI-based customer relationship management (CRM) software product for the financial institution market. The product is just what you are looking for, but you have very big plans for future follow-on additions to it — and you don’t want to be sharing those with your competitors.
You have discussions with the CEO and the other key team members of the Target, and they are ready to join you, provided their office stays separate in the funky, digerati part of town (no offence, but they do not want to be seen as becoming a part of a big bank). And you’re fine with that. So the lawyers are called, and the acquisition process begins.
Phased Disclosure of Confidential Information
The legal activity will start with signing a non-disclosure agreement (NDA). This one should be mutual, because in addition to receiving non-public information from the AI company, you will be letting them in on a bunch of your sensitive material as well, such as your aggressive plans for AI applications in the CRM space. Equally, however, even though you have signed a mutual NDA, you should both be motivated to disclose your respective crown jewels in a phased manner. In essence, instead of one big data dump of all that is secret and valuable, you each should dribble it out to the other, in smaller pieces.
All the while, as you roll out bits of information and receive it, you are looking for deal breakers; that way, if you discover an issue that blocks the deal from proceeding, you have been wise to have limited the release of your non-public information. Equally, you are just as wise to have limited yourself to receiving as little information as possible from the Target before you broke off discussions. This is because you don’t want to contaminate your company with the secret information of the Target if you don’t end up buying it. Remember, anything of a confidential nature that you learn from the Target cannot be used for your own purposes. This is why you also want to be very careful about the distribution of the Target’s information within your organization.
In short, you must approach the exchange of confidential information in a very thoughtful and strategic manner. Don’t involve your key technical people until you are quite certain that a deal is doable with the Target; there’s no need to infect your R&D people with information that, if the deal falls apart because you can’t agree on price, they may not be able to replicate in the future and will then become an impediment for them in the AI space.
The Science — And Art — Of Tuck-In Pricing
This all brings us to the finicky matter of price: what should you pay for the Target? For deals in which more established tech companies are being acquired, when the company has a solid customer base and profits, valuations often drive off a multiple of sales or earnings. But that’s where you really are buying a balance sheet, or more specifically the profit and loss statement. With very early-stage tech companies (often they are “pre-revenue”), sometimes the pricing metric is based simply on a dollar amount per engineer, plus a premium for the software already developed (or the software may be still in development, as even a first version has not yet been completed).
Whatever the pricing formula will be for the acquisition, I suggest agreeing to it early on. Then, later, no matter what you may find during the due diligence phase, the actual price derived from the formula will fluctuate with the actual data unearthed, but the pre-agreed metric will drive the final price regardless. This can avoid the otherwise awkward process of re-trading on price part way through the diligence exercise, which is not pleasant for either you or the Target; and remember, you have to live with the staff of the Target after the acquisition, so they need to be treated well and with respect.
An Important Milestone: Signing the Letter of Intent
Once you have agreed with the Target on price (or, more specifically, the formula by which price will ultimately be determined), you are ready to sign a letter of intent (generally known as the “LOI”). The LOI is a short document, written in plain English rather than in legalese, which sets out the major terms of the deal such as what you are buying, who is joining you from the Target, and what you’re planning on paying.
The LOI is generally non-binding, except for one very important provision: a clause in your favour that says that the Target and its shareholders, and its staff, cannot solicit any other bidders for the Target for a period of time, usually in the 45- to 90-day range. In other words, during this exclusivity period they may only talk to you about doing a deal.
Uncovering Truth Through Due Diligence
Once the LOI is signed, what ensues is an intensive period of kicking the tires of the Target. In small tuck-in deals a few items are researched very thoroughly by the prospective buyer, and none more so than the Target’s key people. Frankly, you can almost think of a tuck-in deal as a recruiting exercise for a number of software engineers, data scientists, and a range of other “techies” and analysts, but in a group setting (which presumably will allow them to work better together once they join you).
Especially if the Target has been around for only a matter of months (rather than years), one of the issues you will want to research carefully is where each of their key staff came from before they joined the Target. You’ll want to make sure that they are not subject to a problematic non-compete provision of a prior employer that is competitive with your organization. Equally, you’ll want to be sure that whatever intellectual property they worked on at the Target was not first “invented” for a prior employer, which might give that earlier employer ownership of it.
Owning the All-Important IP
This brings us to the critical issue of determining if the Target indeed owns the intellectual property it says it does (including, for example, the key AI software that you plan to use for your new CRM platform). It would be devastating if, a few months after closing the deal to acquire the Target, another company approached you to indicate that they own the Target’s important software code — what an expensive mess that would precipitate!
The challenge in determining ownership of software code (or any other copyrightable material, such as documentation) is that there really isn’t an effective central registry of ownership for these sorts of assets; there is a voluntary register for copyrights, but practically its use is limited. Therefore, as the purchaser of intellectual property you must do a “title search” by tracking actual authorship of the software code, because the first human creator of it (or that person’s employer, if the person was an employee who developed the material in the course of their employment) is the first owner of the copyright under black-letter copyright law.
As you do your due diligence (your rigorous review of the Target’s materials, people and software code), you should therefore be on the lookout if the Target used any third-party consultants to prepare any portion of the software code or other materials. Ideally, in such circumstances the consultant has assigned their intellectual property rights to the Target by means of a written assignment. In the absence of such a signed document, you (and the Target) may have a material issue regarding ownership.
Another area for due diligence review involves privacy law. For example, in building their AI capabilities, did the Target collect and use data sets that included personal information of any type? And, ideally, if they have such data, did the Target anonymize the data in a manner such that the material is no longer identifiable back to specific human beings? That would be good.
Otherwise, you might make it a condition of closing the deal that such an anonymization process is carried out on the data. Again, you don’t want to inherit any problems created by the Target, even if you will receive an indemnity from the sellers in the purchase agreement. And that is our segue into the purchase agreement, which we will discuss next month.