Valenta says AI must start with process not hype
Automation firm says companies should fix inefficient workflows before handing them to digital assistants
AS companies rush to adopt artificial intelligence (AI), Valenta co-founder and CEO Jayesh Kasim is warning that many businesses may be starting in the wrong place — focusing on technology before fixing the processes they want to automate.
Kasim said successful automation begins with identifying repetitive work, documenting existing workflows, and redesigning weak processes before introducing AI.
“For us, we solve business problems using technology,” he told the Jamaica Observer during a recent visit to Jamaica. “Our methodology is to always optimise before we automate.”
Valenta, founded in Brisbane, Australia, in 2014 and now headquartered in Toronto, provides AI-powered automation, data, and advisory services to small and medium-sized businesses.
Valenta first maps what it calls the “as-is” process, then designs a more efficient “to-be” workflow before introducing automation.
The best starting points, Kasim said, are repetitive, high-volume and rules-based activities, particularly in finance.
“It always starts in finance. Finance is ripe in any organisation,” Kasim said, citing accounts receivable, accounts payable, payroll and other activities involving data extraction, document matching and approvals.
He said automating those tasks can reduce processing time significantly and lower the risk of manual data entry errors.
“You first want to have processes that are repetitive and they have a logic,” he said. “You want processes that have good volume, because that’s how a business is going to justify a return on investment.”
Kasim advised companies not to automate everything at once.
“Think big, start small,” he said. “Get those first few processes going. Once you’ve got the ROI, let them fund future implementations.”
He said the business case is also being driven by staff shortages and the amount of time employees spend on routine administrative work.
Jayesh Kasim, co-founder and CEO of Valenta, says businesses should fix weak workflows before introducing artificial intelligence.
“What we’re really solving for businesses is giving them their people’s time back so they can do things that are more important in the organisation,” Kasim said.
Although finance accounts for about 80 per cent of Valenta’s initial automation engagements, the company also works in supply chain, advertising analytics, and other operational areas.
Its digital assistants—software tools configured to perform routine business tasks—can operate across a client’s existing applications using client-issued credentials. Depending on the process, they may combine robotic process automation, which follows fixed rules to enter or transfer data, with agentic AI, which can determine and carry out a sequence of steps with limited human direction.
Kasim stressed, however, that automation does not always remove people from the workflow.
“A decision-making process is one that you typically want to leave with a human in the loop,” he said.
In a multi-step process, a digital assistant may complete most of the work, a person may review or approve one stage, and the system may then finish the remaining tasks.
Less complex projects can go live within three to five weeks, while larger implementations may be divided into phases.
The main measure of success, Kasim said, is the amount of staff time saved.
“If a person was spending eight hours a day doing an activity, as soon as the digital assistant goes live it should reduce to one to two hours,” he said.
Roger Grant, Valenta’s managing partner for Jamaica and the Caribbean, said digital assistants are tested with subject matter experts before being placed into production.
“We do rigorous testing, and we work very closely with the subject matter experts,” Grant said.
Valenta’s pricing may be based on hours, transactions, or a dedicated digital assistant model, depending on the process.
Roger Grant, Valenta’s managing partner for Jamaica and the Caribbean, makes a point during a recent interview as Erica Anderson, founder and director of Aurora Technologies Limited, looks on.
Kasim added that return on investment should be measured not only by lower costs but also by whether staff can be reassigned to work that generates more value.
Erica Anderson, founder and director of Aurora Technologies Limited — which holds the Valenta franchise licence for the Caribbean — said automation can also improve working conditions, particularly in accounting departments where employees often work long hours around reporting cycles.
“You’re giving them time to go home,” Anderson said. “You’re giving them time to be with their kids, and reducing the stress and panic at the end of each reporting cycle.”
On data security, Kasim said Valenta holds ISO 9001 and ISO 27001 certifications and does not use clients’ data to train AI models.
“We’re only processing the activity; we’re not using that to train any other models,” he said.
Kasim also warned companies against confusing individual productivity tools with end-to-end business process automation.
He said tools such as Copilot or Claude may help individual employees, but companies often fail to measure whether they improve an entire workflow.
“What we tell clients to focus on is business process,” he said.
Aurora Technologies represents Valenta in the Caribbean, giving regional clients access to local implementation and support.
Kasim said companies also face competitive pressure as rivals begin using AI to redesign how work gets done.
“If you’re not doing this, your competitors are,” he said.
Still, he argued that the first test of any automation project should be practical rather than promotional: whether it reduces wasted time, improves accuracy and allows employees to focus on work that creates more value.