A bespoke solution was key for ECI’s deal origination process according to Suzanne Pike, partner and head of origination at the firm.

RD: How does your firm incorporate technology into the deal origination process?

Suzanne Pike: We use Amplifind, which is a prorietary AI platform for lead sourcing and prioritisation. At the heart of Amplifind is a machine learning tool that takes and processes data that we have entered into Salesforce about the many thousands of businesses with which we are, or have been, in contact with as potential investee companies. 

It looks at areas including company financials, ownership information and keyword metrics, and then produces a prioritised list of opportunities. Each month, one of my team reviews the shortlist on the Amplifind platform and we have a meeting to review each that has been positively selected. This process then flags new lead opportunities in Salesforce for us.

The other way that it helps us is through the automated profiling of companies in which we are interested. We used to have to ask the analysts in our team to produce these profiles, using data from Companies House, LinkedIn, business websites, Google searches and other sources. Now all of that can be done with the click of a button.

 Amplifind uses AI to generate rich descriptions of what a company does and how it is currently performing. This means that our analysts are freed up for more in-depth research because Amplifind has taken care of the simpler tasks.

So while it started out as a lead-sourcing tool, it has grown in terms of remit and has effectively become the backbone of our origination research process.

RD: Was Amplifind bought off the shelf or built in-house?

Pike: Amplifind was created completely in-house and is bespoke to us, as the ML models are trained on proprietary data and the UI is designed with our specific business processes in mind, but we did get some help. We used a firm of data analysts and engineers to productionise our prototype in Microsoft Azure and we have intentionally used standardised code in the UI to make the product relatively low risk and flexible from a technological perspective.

We did not want to use an expensive third-party platform that would not actually give us any edge over the competition. In that case we would be making our business fit the technology, rather than making the technology fit our business.

RD: When did you start implementing Amplifind, and what drove you to do so?

Pike: We created the first Amplifind models in 2020 and then embedded the prioritisiation tool in our core processes from 2021.

One driver was that there is so much data out there now that it is not physically possible for humans to get through it all. Using a system like Amplifind ensures that our human skills are deployed to the right business. Another related driver was that, in a competitive market, we were finding we needed to identify leads and meet management teams of target companies earlier and earlier. 

That meant that although we had gradually moved up in deal size, the potential universe of relevant growth businesses to trawl through had significantly increased. On average, we engage with a business for five and a half years before investing, and sometimes as much as a decade.

RD: What were the biggest challenges you faced when you first implemented Amplifind, and what are your current challenges?

Pike: We are in the process of exploring a possible partnership with a university applied AI department to look at how we might best leverage our proprietary language-based data. An area that has been harder to crack, however, is the availability of third-party language-based data and good quality keywords. 

We have a subsector-driven approach to origination so the better the keyword data we can extract, the better indication we will have of whether a business is a good fit for us.

However, the nature of the challenges has changed radically in just the last few months. Our business goals regarding technology are still the same; to use technology to enhance the origination work we do as humans and to support value creation in our existing portfolio companies. What has changed is that the barriers to accessing data are falling away. There are more and more off-the-shelf AI tools available too. 

Everybody has been talking about ChatGPT since its launch but, even since then, the growth in the number of new technologies providing quick wins for a business like ours has been phenomenal. The main challenge now is not getting carried away with the technology and staying focused on what we are trying to achieve. The danger is that technology just creates a lot of extra complication in our business processes, which is not what we want.

RD: What are some of the successes you have had since starting to use Amplifind? What pain points have you been able to eliminate or mitigate? 

Pike: Ideally, we would be able to point to all the deals we have completed that came to us through Amplifind. But in reality there will always be multiple touch points along the way to an investment. Our goal is to talk to a potential portfolio company before anyone else, but we understand that advisers will get involved later and the process of originating and winning a deal will be multifaceted.

However, I can say that we have reduced the time spent reviewing potential leads by more than 80% and we have seen an eight-fold improvement in terms of converting companies reviewed into leads on Salesforce, when compared with our previous traditional Companies House searches. This is because with Amplifind we can now look at a wider range of businesses in less time. 

More than 40% of our pipeline is now sourced from the Amplifind platform. That ratio will increase because we have had Amplifind in place now for almost three years but we typically identify and meet target businesses more than five years ahead of a deal.

RD: In what ways does your use of Amplifind extend to the value creation process?

Pike: We use Amplifind to assess potential bolt-ons and portfolio companies’ M&A strategy in general. 

Going further, we would like to give the management teams at our portfolio companies direct access to Amplifind so that they can use it to monitor potential targets as well.

RD: Do you have any highlights that stand out in the value creation area when it comes to technology?

Pike: We have found that people are using Amplifind in interesting and unexpected ways. 

If we are working hard on a potential deal and we know who our competition is, then we always want to check how investments they have previously made in the industry are performing. Before Amplifind, doing so was hard and would involve a lot of digging around in company accounts. But now we have the ability to automatically profile any business in the UK with one click of a button in Amplifind, this information is available instantly. It is just a marginal gain but, once added to the many other marginal and major gains we get from Amplifind, the effect is powerful.

RD: Do you think LPs see the use of technology during origination and value creation as a differentiating factor for a private equity firm?

Pike: We ran a number of demonstrations of Amplifind for LPs during our last fundraise and had very positive responses. 

Investors want a track record – nothing trumps that. They also look forward, however, wanting to know that a private equity firm is not just a group of individual deal-doers but a business with a coherent, repeatable model. Effective technology really helps to show that you have good central management and robust processes. 

We are going to continue to invest in technology such as Amplifind because we feel that technology will be an even bigger topic next time we fundraise.