Automation has become commonplace in subsurface workflows, helping geoscientists process larger datasets and run increasingly complex analyses. Yet most automated workflows still follow a predefined sequence of steps, regardless of what the results reveal.
At the Dig X Subsurface 2026 conference in Lillestrøm, Umedzhon Kakhkhorov, Senior Software Analyst Geoscience at Cegal, will present how structural uncertainty analysis in Petrel could evolve from workflow automation towards more agentic geoscience.
The distinction is important. Traditional workflows follow predefined steps, while agents can dynamically select tools and calculations based on intermediate results, within boundaries set by the geoscientist.
Kakhkhorov’s presentation uses the Volve field as an example of how structural uncertainty can be assessed through ensemble modelling. By generating multiple depth-map realizations from different velocity scenarios, geoscientists can better understand how uncertainty affects closure size and volume estimates.
Instead of relying on a single depth interpretation, the resulting ensemble can be used to calculate closure probability and estimate P10, P50 and P90 volumes, helping teams assess prospect uncertainty before drilling.
Using Python integration, calculations can be run directly against Petrel project data without manual export and import steps.

The next step, according to Cegal, is to allow language models to use these tools more autonomously. Rather than simply executing a predefined workflow, an agent could compare velocity scenarios, determine whether additional realizations are needed and recommend the next analytical step.
The geoscientist would still define the geological assumptions, quality controls and stopping criteria, while the agent assists with organizing calculations and synthesizing the results.
Although the current implementation focuses on automated ensemble calculations, the work offers a glimpse of how agent-based systems could eventually become part of everyday subsurface interpretation workflows.
Join us in Lillestrøm on December 2-3 to learn more about how Cegal is exploring the transition from workflow-based automation to agent-guided geoscience. The presentation is part of Dig X Subsurface 2026, a conference dedicated to digital workflows, data and AI in subsurface geoscience. The program can be found on the conference website.
![var rnd = window.rnd || Math.floor(Math.random()*10e6);
var pid980600 = window.pid980600 || rnd;
var plc980600 = window.plc980600 || 0;
var abkw = window.abkw || 'gpbanner';
var absrc = 'https://servedbyadbutler.com/adserve/;ID=188736;size=0x0;setID=980600;type=js;sw='+screen.width+';sh='+screen.height+';spr='+window.devicePixelRatio+';kw='+abkw+';pid='+pid980600+';place='+(plc980600++)+';rnd='+rnd+'';
var _absrc = absrc.split("type=js"); absrc = _absrc[0] + 'type=js;referrer=' + encodeURIComponent(document.location.href) + _absrc[1];
document.write('');
From workflow to agent](https://geo365.no/wp-content/uploads/2026/10/Cegal_Kakhkhorov_Figure-1_cut-1024x684.jpg)