SHARJAH, 14th October, 2026 (WAM) -- Artificial intelligence (AI) is being used to test investment proposals at an Abu Dhabi sovereign wealth fund, providing alternative arguments for investment committees while human reviewers check its analysis, delegates at Sharjah Investment Forum 2026 (SIF 2026) heard on Wednesday.

Maria Sanchez, Senior Vice President of AI Product Delivery at Inception (G42), outlined the deployment at the fund - which she did not name - during the "Powering Adaptive Economies with Agentic AI" panel discussion. The panel also featured Rima Semaan, Chief Solutions Officer at Microsoft UAE, and was moderated by Kunal Devani, Partner, EY.

Describing the investment application, Sanchez said: "For every investment, there are months of work from a big team preparing documentation to defend why they should be investing in a deal. Then you get AI presenting why they shouldn't invest."

Human reviewers check the recommendations for any inaccuracies or inappropriate use of data before they inform investment decisions, she pointed out.

Sanchez also described procurement applications in which AI supports workflows from purchase requests through to contract awards, including checking budgets, preparing scopes of work and proposing evaluation criteria. People retain the ability to review and override the system's proposals.

"You start automating low-risk workflows, things that can be reversed, and then you keep getting confidence and keep moving in the journey," Sanchez said.

She stressed that organisations must be able to trace agents' actions and identify where human review has taken place. "You understand what's going on at every point in time. You understand why agents took this decision and not the other," she said. "As long as those controls are put in place, it is also up to every leader to undertake some experimentation."

Building a human-led, agent-operated workforce

Semaan urged businesses to build the infrastructure needed for their current use cases and begin implementation, rather than continually waiting for the latest model. Scaling adoption, she said, requires connected organisational data, appropriate access controls and the ability to trace agents' actions.

Workforce preparation is equally important. As employees increasingly supervise AI agents, they will need judgement, delegation and leadership skills alongside continuous learning.

"You need to know how to delegate. You need to make sure that they take the right actions and hold those agents accountable," Semaan said. "Every employee, at the end of the day, becomes kind of a manager."

She described the emerging operating model as "human-led, agent-operated", arguing that successful businesses will redesign workflows and pursue new revenue streams alongside productivity improvements, supported by clear safeguards and leadership commitment.

Sanchez similarly stressed that leaders' own use of AI helps drive adoption throughout an organisation.

"Whenever I see a leader of an organisation who has the mindset that they want to embrace AI for the most valuable activity they do in their day-to-day, it always cascades down in a massive way," she said.