Capital Appreciation banking on AI for the future

With four digitally enabled banks launching in less than a year, SA’s financial services sector has become a lucrative environment for fintech software developers like Capital Appreciation’s (Capprec) Synthesis business.

In Capprec’s results to end-March, Synthesis grew revenue by 48% and contributed a quarter of the group’s operating profits, from 17.6% in 2018.

Since Capprec bought Synthesis in 2017, it’s almost doubled sales, to R137.8m.

Synthesis is divided into three units: cloud services, building digital channels for financial services companies, and regulatory compliance technology.

"Those areas of service are critical to our existing customers and to the success of banks in general, given the environment today where there are newcomers disrupting the market," says MD Michael Shapiro.

The market has given Capprec a wide berth: Capprec shares have drifted since listing at 100c a share in October 2015 to their present level of 83c.

Still, Synthesis is one subsidiary to watch.

A quarter of the revenue it generates is recurring, contrary to Capprec’s payments and payment infrastructure business, where 62% of revenue generated in the 2019 financial year was from once-off sales of payment terminals.

Capprec joint CEO Bradley Sacks wants to see recurring revenue for the group hit 70% from its present 47%.

The payments and payment infrastructure business has plans to increase recurring revenue streams. These include growing its maintenance, support service and transactional fees.

But Synthesis, for now, is the lowest-hanging fruit. Its big blue-chip clients include the big five banks, Discovery Vitality, and asset managers.

It also recently signed up TymeBank — no surprise given that TymeBank’s parent company, Patrice Motsepe’s African Rainbow Capital, is also a shareholder in Capprec.

Its smaller clients include specialised lender Sasfin, Afgri’s recently acquired Grobank and Mauritius’s CIM Financial Services.

Irnest Kaplan, MD of Kaplan Equity Analysts, says Synthesis is in a good position to further grow its market share. "I think banks are increasingly making use of cloud computing and new technologies to try to differentiate themselves and get scalable technology solutions," he says.

In fact, Synthesis has been around for more than 20 years, providing cloud data management and digital platforms.

But artificial intelligence (AI) and machine learning are providing interesting new avenues of growth.

For instance, when Barclays sold off Absa, the local bank had to find all traces of Barclays in its documents, pictures and online archives. Shapiro says there were more than 30,000 artefacts that Synthesis had to trace and erase.

But perhaps the most interesting AI implementation project the company has delivered is sentiment and mood detection on e-mails received by one of its insurance clients.

"One insurance provider we are working with is receiving thousands of e-mails a month," says Shapiro. "The machine learning model now helps them recognise immediately what that e-mail relates to, filter it and reroute to the appropriate team, whether it’s new business, claims or complaints department, which helps with quicker turnaround times."

Discovery Health CEO Jonathan Broomberg waxed lyrical, at the Discovery Health Medical Scheme’s AGM in June, about machine learning that reads mood and sentiment.

The fact that Synthesis is developing solutions that the likes of Discovery think are future-forward may bode well for its ventures into new markets. The company is also starting to market its cloud services, AI and machine learning services to retailers and telecoms companies.

Last year, Synthesis was named advanced consulting partner of cloud services for Amazon Web Services. That means it gets referrals from the US tech giant’s blue-chip clients in the Middle East and Africa region.

Synthesis is branching out into agritech, having established a joint venture with Afgri. The technology it provides helps the agricultural services company assess the feasibility of loans to commercial farmers, using projected cash flows from grains that farmers have already harvested or are busy planting, and using data on weather patterns that Afgri monitors.