Success Stories
From Zero to 100K In 24 Months: How We Enabled Useful AI in Offshore Drilling
The Client: A large offshore drilling company operating a diverse rig fleet of ultra-deepwater drillships, versatile semisubmersibles and modern shallow-water jackups across the world.
Slide 1 of 7 — use arrow keys, swipe, or click the arrows to navigate
Seeing the Opportunity
In early 2024, leadership at the company and DTS knew AI had the potential to provide value to crews using their asset management system. DTS had delivered prototype demonstrations of an early query-writing conversational model, so the potential was obvious.
But obstacles remained:
- Even frontier models at the time had real limitations
- Hallucinations
- Short context windows
- A tendency to go "off script" during long conversations
Crews working on the rigs are busy. They're only going to use AI solutions that are fast, reliable and — crucially — help them get their work done faster.
Key Insight
The breakthrough was identifying use cases where AI could deliver value in a single interaction — no long conversations required.
Now You're Speaking My Language
A new functional requirement emerged: users needed to read detailed, technical maintenance content in languages other than English. DTS quickly delivered a prototype solution, adding translation functionality to the Work Order form.
Translation turned out to be a perfect use case for LLMs:
- More colloquial, natural-sounding text compared to dedicated translation tools
- Better at translating technical maintenance language
- Simple ease-of-use — just pick a language and tap Translate
Quick Wins
DTS included three additional LLM integrations in the same release, each solving a specific, tightly-scoped problem.
Reverse Translation
Takes crew notes in any language and translates them to English.
Significantly improved English-language comments from international rig crews.
Failure Coding Assistant
Reads completion comments and selects correct failure cause, mechanism, effect and mode codes.
~5 minutes saved per corrective Work Order. More consistent failure coding for better analysis.
Completion Comment Assistant
Takes first-draft notes and helps write cleaner, easier-to-read completion comments.
Time savings for users, clearer completion comments, improved analysis.
What Made It Work
"One Shot"
Users don't need to get into a long conversation. Just one click, and their work gets done faster. These problems are tightly scoped, with the AI delivering on very specific objectives.
Letting Crews Focus
Rig crew members are highly compensated specialists. A tool that quickly helps code failures and write comprehensive notes gives them more time to focus on applying their skills to the mission.
Opt-In
Every AI integration is optional. Users on the rig use AI hundreds of times every day because it speeds up their work and makes life easier — not because they're required to.
Human In The Loop
The AI in these implementations is an assistant, not the decision-maker. Users have a helping hand, but are responsible for making decisions.
Talk Me Through It
DTS then moved onto conversational LLM integrations to guide users through complex, detailed change management processes: Deferrals and Operational Risk Assessments.
Deferral / ORA AI Helper
Guides users through multi-step deferral and risk assessment processes. Interviews the user to document reasons, impact and mitigation. Helps write hazard/mitigation statements. Fills out the form automatically.
Better deferrals, significant time savings for senior rig personnel, and reduced "failure to launch" scenarios.
Custom Chat Builder
AI helps users build sophisticated queries and schedule LLM analysis — all without writing code.
Users can now perform qualitative analysis at scale — auditing Work Order completion comments for quality, consistency and adherence to standards.
84,000+ AI Interactions and Counting
Two years later, users have leveraged AI inside of VAMS more than 84,000 times.
Ready to Transform Your Enterprise Operations?
Let's discuss how DTS can help you get the most out of your existing systems — and prepare for the AI-powered future.