Workflow pt 3/3: The industry should really be talking about decision intelligence more

Practically....what is decision intelligence? The word "actionable insights" is too broad. Actionable insights really means, we (technology company) think there COULD be value in this number or this metric, and that is where it stops. If the software was confident enough, it would actually tell the end user what to do....and that is where the current technology landscape falls short within industry operation.

Our view of decision intelligence is an actual user recommendation. Automating the analysis to suggest a specific action. Here is the practical view of it. Change this setpoint from x to y to get 7% production uplift. Replace the filter for this well, this should improve reliability by 25%. Upgrade to this part to increase production by 5%.

If a system is describing itself as actionable insights and isn't being as practical as that, then its just lazy marketing built on poor assumptions. We think this metric could be insightful (assumption), but we have never actually engaged with the end user to get feedback if it is or not.

🔥

...sorry not sorry to all the other technology companies who advertise "actionable insights"

All of this relates to workflow and how a system can be best poised to understand what work is effective & make more effective recommendations than humans.

Main points:

• Decision intelligence is a system making decisions based on the known data...this presents a problem...if the data collected in the system is limited, then decision intelligence will be shit
• Having a connected system that understands actions, users, assets, outcomes, set points, targets, baselines, equipment....all together can be pretty powerful...thus enabling specific user recommendations which enhances work effectiveness.
• Current gap in technology is nobody is measuring work effectiveness and enhancing user recommendations through that process
• Field input & knowledge is the KEY to making better decisions

WHY decision intelligence is so important stems from the wells that are shown below

The question is (first image), what is being done day 1 - 30 of this well deferring and what is done differently to get the well back up to what it should be producing.

Second image: What is being done each time the well defers to get it back up...doesn't seem like it is attacking the root cause as this seems to be heavily unreliable.

Current tech landscape:

• Field data capture systems were never designed around "decision intelligence" as the end state...and pivoting that direction for FDC companies will be limited
• Current work management systems validate what work was done, not what work generated the was the highest rate of return
• Others measure what is complete, we measure what is effective
• Bolt on systems disrupt intelligence. Fragmented systems limit decision intelligence throughought the whole chain because they don't talk to each other and are not structured in a way to understand each other
• Daily prioritization should be based on effectiveness...(What is the most positive action to take...what gives me the best outcome)

What everyone really wants:

• Tell me where to go & tell me what to do. Daily tasks are aggregated into ROI trends that understand impact of certain decisions. Projects are suggested based off a system and marketplace.
• What everyone is really after is, how can we make the most intelligent decisions in the shortest amount of time managing the most production. But no systems have linked work management to time series data before, so this is all separated and not known well
• Everyone wants knowledge to be shared instead of silod
• If work identification can link to work effectiveness, then you have a full circle system learning from previous separated or non existent functions
• Better projected outcomes before work is distributed. Work prioritization is not just about what the well is currently deferring. It is about what is currently deferring and what is the action that will yield the best rate of return.
• Example: A field operator might want to prioritize a 100bbl deferral over a 200bbl deferral. If the action on the 200bbl doesn't fix the issue and another person has to visit it tomorrow all while the action on the 100bbl fixes the issue for 6 months, then it is better to prioritize known rate of returns. Prioritization of work should be based on better projected outcomes (work effectiveness), while also taking into account value/cost.

Lasting thought:

User recommendations & more explainable UX is still so early within industry. The majority of companies still advertise "actionable insights", meaning they haven't taken the next step. This means there is still soo much opportunity to apply tech in a better way to get higher ROI.

The main crux of getting user recommendations in everyone hands means the platform has to incorporate most of the work that user has to do, which involves knowing all the work that a field tech, optimizer, engineer and more do on a daily basis. That is why workflow and ML are so important to be utilized together, and platforms that don't connect the two will have a limited value ceiling ROI over time...which people are seeing from FDC and setpoint optimization platforms...they are now old tech and don't provide the ROI across the organization, & do not have a great future outlook over the next few years.

‍