Verascient Bets on South African Engineers to Make Enterprise AI Work

Verascient CTO Emile Dos Santos Ferreira, left, and CEO Keagan Stokoe, right.
Verascient co-founders Emile Dos Santos Ferreira, CTO (left), and Keagan Stokoe, CEO (right). Photo courtesy of Verascient.

Verascient is betting that enterprise AI needs engineers working alongside customers as much as it needs software. The South African-founded startup combines infrastructure for organising company knowledge with specialists who help businesses redesign the workflows that use it.

In written responses to AfricaBusiness.com, co-founders Keagan Stokoe and Emile Ferreira explained how that approach shapes their product, recruitment plans and international ambitions. Their answers follow the company’s August announcement of a US$1.2 million pre-seed funding round.

The opportunity they describe is practical: businesses may already have AI tools, but still struggle to connect them to reliable information, appropriate permissions and everyday operations.

From scattered information to usable company knowledge

Verascient’s initial focus includes financial services, insurance and logistics, where information about a customer, transaction or operational decision can be distributed across emails, documents, meetings and internal systems.

The company is building what it calls a temporal knowledge graph. In plain language, this is a connected record designed to preserve information together with its relationships, sources and history, rather than keeping only the latest version.

That historical dimension matters to the founders’ proposition. An AI system answering a question about a customer commitment, for example, needs to distinguish an earlier proposal from a later agreement. It also needs to identify the source and respect who may access it.

Ferreira describes the platform as infrastructure that can work alongside existing AI tools and business software. Its purpose is to give those systems company-specific context, rather than require customers to replace their entire technology stack.

The “operating system” description used by Verascient is therefore a product analogy: a shared foundation for knowledge, permissions and workflows, rather than a replacement for a conventional computer operating system.

Why the business model includes engineers

The software is only part of the offering. Verascient also supplies forward-deployed engineers: technical specialists who work directly with customer teams to identify problems and implement solutions.

The founders argue that successful deployment requires understanding the work itself. That includes deciding which tasks should change, which information is needed and how employees will use the resulting system.

“Our model combines infrastructure with engineers working directly alongside our customers,” Stokoe said in the company’s announcement.

Ferreira acknowledges that some businesses should build internally, particularly those with experienced platform teams and the resources to maintain their own infrastructure. For others, Verascient aims to provide both the technology and implementation capability they would otherwise need to assemble.

The commercial question is how efficiently that combination can scale. Close customer involvement may help solve difficult deployment problems, but it also requires skilled engineering time. The responses supplied to AfricaBusiness.com did not disclose pricing, typical implementation duration or the balance between recurring software income and services revenue.

What the pilots show — and what remains undisclosed

Stokoe outlined potential applications ranging from reconstructing an investment deal’s history to retrieving shipment correspondence and reviewing the reasoning behind an operational decision.

He said an investment team could use the platform to assemble context from emails, notes and data-room materials, while a logistics team could investigate shipment status and disputes without chasing information across multiple colleagues.

These examples explain the intended use of the platform. They are not independently verified performance benchmarks. Verascient said its pilots were commercially sensitive and did not provide named customer case studies or supporting measurements for the claimed time savings.

The outcomes the company intends to improve include shorter turnaround times, less rework, faster responses and reduced dependence on a few experienced employees. Assessing those outcomes would require comparing deployment and review costs with the previous process, alongside the quality of the resulting work.

Stokoe also cautioned that technology alone would not deliver the employee benefits he envisages.

“This only holds if the company redesigns the work,” he told AfricaBusiness.com. “If you layer AI on top of an unchanged process, you get the same busywork slightly faster, and people notice.”

Permissions and human responsibility

According to Ferreira, Verascient is designed to preserve source references, inherit access permissions from connected systems and retain historical information, including contradictions. He said the company also supports on-premises deployment for customers with particular security or data-ownership requirements.

Those are descriptions of the product’s intended controls, rather than independent assurance of security or accuracy. The interview responses did not include audit findings or reliability benchmarks.

The founders place responsibility for consequential decisions with people. Ferreira said such decisions should have a named human owner who can explain and stand behind the outcome, with human approval required for sensitive actions.

For prospective customers, that makes implementation details important: which sources an answer relies on, whose permissions apply, what actions an agent can take and when a person must intervene.

What Verascient means by South Africa’s “top 1%”

The company’s recruitment campaign describes its target as the country’s “top 1%” of AI talent. Asked to explain, Stokoe said the phrase referred to qualities such as learning speed, judgement, curiosity and evidence of work, rather than a formal ranking or particular university qualification.

Verascient plans to keep its team small and selective. Its engineering roles require technical depth, while customer-facing deployment work also demands an understanding of business operations and the ability to work through ambiguous problems.

The founders want to offer South African engineers internationally relevant work without requiring them to leave the country. However, the supplied responses did not specify a recruitment target, salary range or number of new jobs.

Ferreira said the company was supervising a student research project and discussing a potential partnership with another local university. Possible hackathons and involvement in initiatives such as the Deep Learning Indaba remain ambitions, rather than announced partnerships.

Funding and international expansion

Verascient announced its US$1.2 million pre-seed round on 24 August 2026, describing it at the time as approximately R19.5 million.

According to the announcement, investors included Founder Collective, Andrena Ventures, Cambridge Enterprise and Summit Ventures, alongside angel investors including Alan Knott-Craig and Shayne Mann. The round was also reported by Africa Private Equity News on 25 August.

The founders say Verascient is incorporated in the UK, with its first office in Cape Town and South Africa as its initial market. Planned expansion would take it into the UK, Europe and subsequently the US.

The funding is intended to expand the engineering team and develop the technology for further enterprise deployments. Alongside its initial sectors, the company sees opportunities in legal and professional services, where historical knowledge and controlled access to information are important.

Verascient’s proposition brings together two challenges: building software that can work with company knowledge and finding people capable of embedding it in daily operations. Its next stage will be judged by repeatable customer results and whether that implementation-intensive model can grow sustainably.

Sources and information

This article draws on written responses supplied to AfricaBusiness.com. Product capabilities, pilot outcomes and future plans are attributed to the company; customer performance claims have not been independently verified.