INSIGHT · Project Timeline · Nov 2024 → Present
A record of what we've shipped, when we shipped it, and what's coming next. Covers the Android field apps (Insight NxtGen and its successor, Insight NxtGen Neo), the Windows administration app, and cross-project operations work.
Released the mobile app under new ownership. The previous vendor's proprietary component was removed, so from this point we could update the app on our own schedule instead of waiting on someone else's release cycle.
Took ownership of the servers and codebase. The first weeks went into reading through everything and building an internal understanding of how the system works. Where documentation did not exist, we wrote it, so the knowledge now stays with us.
Ran the SAP import into INSIGHT during Martin's absence, so there was no gap in the data. We also documented the steps, so the process is no longer tied to one person.
This release focused on cutting keystrokes for field teams and keeping the app working on our newer devices.
Built for the days when operators are working through a lot of seals at once. Fewer taps between scans, and no risk of accidentally logging the same seal twice.
Updated the Delivery Accuracy dashboard on Insight Web from "Plattformveien 1" to "OEG Dusavik", as Bent requested.
Released the new Windows app in time to cover a scheduled server update that took the legacy INSIGHT desktop offline. The new app came up cleanly and carried the workload while the old system was down, so our operations were not interrupted.
Set up the access ConocoPhillips needed for their annual integration testing, which is tied to their SAP upgrade cycle. Their test window ran to plan.
Second Windows release, bringing the new app to the point where it replaces the old INSIGHT desktop. Focus was on removing manual work that had built up around the old app over the years.
Built a working understanding of the NLP system end to end, then delivered the specific changes Bent Ove requested.
Insight NxtGen Neo launched on the Play Store. This is a completely new mobile app built from scratch on modern foundations. It was released alongside the existing app, so operators could adopt it at their own pace with no reconfiguration and no risk to their existing work.
Two small things had been costing operators time every day. Sessions were quietly timing out mid-shift, so operators had to sign back in for no visible reason. And when something did go wrong, the message on the screen usually said nothing useful. Both fixed in this release.
Lise flagged that the Halliburton report had not gone through Visma the previous evening, and it was needed the next morning before invoicing could go out. We investigated and worked through the resolution the same day.
When an operator taps an object that has attachments (inspection PDFs, photos, and so on), the object details now appear right away. The attachments continue loading in the background, so a large or slow attachment no longer holds up the whole screen.
Began the work to send container data cleanly from INSIGHT to MagnetJQS, an integration we did not previously have. Active workstream since then, working through the identifier questions (GLN vs GIAI vs Container ID), format alignment (containers whose IDs contain dashes, e.g. B-550 becomes B550), which inspection dates to send, payload validation, and communication issues on their side.
Added a new dashboard to Insight Web and fixed issues on several others, all against Bent Ove's requests.
The single largest quality-of-life release for Neo. Operators can now unlock themselves in a few taps, personalize their view, and refresh objects without losing their place.
Removed the overhead of keeping the mobile fleet on the latest version.
Received the server configuration from IT and moved the API and the Web application onto separate servers. Before this, if one went down it took the other with it, and vice versa. Now a problem on one side stays contained on that side.
The mobile app now talks to the server over an encrypted connection instead of the old unsecured one. Operators did not have to reconfigure anything. The app moved with them, and existing work carried over untouched. This also sets the stage for the new back end going live on Sept 16.
Insight Web went down repeatedly and blocked customer access. Emergency restarts (via PCL) restored service each time. While managing the incidents, we identified the underlying cause: a long-standing weakness in how the applications share database access, which was allowing a single stuck request to bring the whole platform down. That understanding set up the Stage 1 fixes that followed.
First round of fixes released to address the underlying issue behind the recent outages. This ends the class of incident that had been requiring PCL restarts. Further fixes planned for the following week.
Our Cyber Security team reviewed Insight NxtGen Neo and cleared it. The security work built into Neo since launch has now been formally validated by an internal review.
The largest single piece of engineering work of the year, and the main reason the app has felt slow up to now. Everything shown in the app has to be fetched from somewhere behind the scenes, and the piece that did that fetching was built many years ago. It could only handle one request at a time, which meant that when several people were using the app together, everyone was effectively standing in a queue behind each other. It also could not be made secure enough for today's requirements, and it was too fragile to extend, so new features were very difficult to add. If a customer had asked to connect their own system to ours, we had nothing workable to offer them. Rather than patching the old one any further, we spent two months building a replacement from scratch. Once it is in place:
Neo has felt modern on the surface since launch, but has been talking to the old back end behind the scenes, which caps how fast it can be. Pointing it at the new back end is where operators will notice the speed difference.
Fixes on the Desktop side that address the same underlying issue behind the recent Web outages. Together with the Sept 7 Web release, this completes the current round of stability work across the platform.
Ran a full pass through the tasks admins do every day on Desktop (checking object locations, creating loadouts, adding customers, and more) and identified thirteen things that slow people down or catch them out. Seven rated high impact. Work to fix them starts next.
Insight Web is the customer-facing side of the platform, used by our clients every day. Once the Desktop quality work is underway, the next major chapter is a full revamp of Insight Web: a new look, a rethought user experience, and a big step up in speed on top of the new back end.
Today Neo runs on Android only, which means field teams on iPhones and iPads have no way to use it. Bringing Neo to iOS means the same field-team experience is available to everyone who works out in the yard, regardless of what device they carry.