Catch the trend
before it repeats.
Nuclearn Performance Improvement AI codes your condition reports against WANO performance objectives and criteria, shows which codes stand out by unit, department or quarter, and turns the trends you follow into a report. Your performance improvement team decides what they mean.
- Works on the CAP data you already have
- Single-tenant, hosted by Nuclearn
- Signs in through Nuclearn
- Backed by NuclearnAI software built for nuclear plants since 2021
- Over 70 facilitiesNuclearn products serve sites in North America and the U.K.
- ISO/IEC 27001:2022Certified information security management
From thousands of reports to the few trends that matter.
Each condition report is a data point. Performance Improvement AI codes them the same way, adds them up by unit, department and period, and keeps the trends you care about in front of you.
Trend-code each report
A trained model assigns WANO performance objective and criteria codes to each condition report, with a confidence for each, so reports are coded by one standard across units and years.
See trend coding → PI managers and department headsFind what stands out, and where
Slice the codes by unit, department or any field you carry, and see which ones sit furthest from the site's average.
See trend analytics → Trend coordinatorsFollow a trend over time
Save the codes that belong together as a bucket and re-run it on the latest reports, by month, quarter or year.
See buckets → CAP coordinatorsBuild the trend report
Turn your buckets into a report on your own Word template, with each code's description, example condition reports and how it compares.
See reports → PI, oversight and fleetCompare with a benchmark
Choose a benchmark dataset, and each code shows how far your site sits from it, as well as from its own recent history.
See comparisons → Screeners and initiatorsScreen and record each report
Screening, similar past reports and the condition report record are in Nuclearn CAP AI, on the same platform and the same datasets.
Visit cap.nuclearn.ai →Thousands of reports a year, and a trend somewhere in them.
A corrective action program is read one report at a time, and it also has to be read in the aggregate. NRC inspectors assess whether "the organization periodically analyzes information from the corrective action program and other assessments in the aggregate to identify programmatic and common cause issues."2
That aggregate view is hard to keep up. When codes are applied differently from one person to the next, or the trend report is rebuilt by hand each period, a low-level trend can look like noise until it repeats.
Performance Improvement AI codes the reports to one standard and re-runs the counts on the latest data, so your team spends its time on what the trends mean.
INL's conservative estimate, from a survey of condition report datasets from multiple utilities.
Under Inspection Procedure 71152, inspectors review the site's issue tracking, trend reports, metrics and performance indicators for trends that might indicate a more significant safety issue.
These figures describe report volume and the NRC's inspection schedule. They are not savings from Nuclearn Performance Improvement AI. Sources at the foot of the page.
Starts from the reports you already have
No new forms for anyone at the plant. Performance Improvement AI works on condition report data, wherever it was written.
Bring your condition reports
Upload a CAP export as a dataset, or work from the condition report datasets already on your Nuclearn platform.
Trend-code them
The WANO PO&C model reads each report and returns the codes that apply, each with a confidence.
Analyze, follow, report
Compare the codes by unit and period, save the trends you follow as buckets, and build the report from them.
One set of codes, applied the same way.
The pre-installed model reads the text of each report and returns the WANO performance objective and criteria codes that apply, with a confidence for each. It covers 905 codes in 26 areas, from operations and maintenance to performance improvement and operating experience.
Codes are the model's suggestions, each with its confidence. Your team reviews the trends they add up to.
See which codes stand out, and where.
Slice the coded reports by any field you carry: unit, department, organization or site. Choose how to compare the slices, and look across time from a single day to a year.
Use it to see which areas are drifting from the site's average, or how this quarter compares with the last.
Follow a trend, not a spreadsheet.
Cluster Analytics draws each code as a bubble, placed near related codes and sized by how often it's coded. Select the ones that belong together, such as a set of emergency preparedness codes, and save them as a bucket.
The trend report, built from your buckets.
Reports gather related buckets into one printable document, such as a strengths and weaknesses evaluation. Upload your plant's Word template, choose the buckets, and Performance Improvement AI fills it in.
The report gives the numbers and the examples. What they mean, and what to do about them, stays with your team.
Take the trends to where your team works.
Export the analysis, connect Power BI, or pull the results into your own tools with a Nuclearn platform API key.
# An analytic's results, as CSV (Power BI reads the same URL) curl -H "X-API-KEY: $NUCLEARN_API_KEY" \ "https://nuclearn.your-site.example/api/stats/\ analytic/$ANALYTIC/results?download_as_csv=true"
# Each report with its codes and confidences, as CSV curl -H "X-API-KEY: $NUCLEARN_API_KEY" \ "https://nuclearn.your-site.example/api/stats/\ predictions?analytic_id=$ANALYTIC&include_input_data=true&as_csv=true"
Access uses a Nuclearn platform API key, so the platform's sign-in, roles and dataset permissions apply.
Built for real CAP data
905 codes, 26 areas
WANO performance objectives and criteria, from operations, maintenance and engineering to performance improvement and operating experience. Each code carries its description, so it reads the same to everyone.
Search the map
Find a code by its number or by words in its description.
Top codes or a threshold
Keep each report's top N codes, or those above a confidence you choose.
A benchmark to compare against
Set a benchmark dataset scored with the same model, and the analysis shows each code's difference from it. Reports call this "relative to industry".
Count or confidence
Count each code once, or add up the model's confidences.
Shared buckets
Let colleagues edit a bucket, or keep it to yourself.
Permissions that follow the data
Who can read, score and analyze a dataset follows its permissions on the Nuclearn platform, and a report checks that its reader can read the data behind each bucket it uses.
Day to year
Group by day, week, month, quarter or year, with growth from the period before.
Your report template
Upload the Word template your team already uses, and reports fill it in.
Your reports stay in your Nuclearn deployment.
Condition reports, their codes and the analyses built on them live in your plant's datasets on the Nuclearn platform, and access goes through the platform's sign-in and roles.
Single-tenant, hosted by Nuclearn
Each plant gets its own instance, hosted by Nuclearn in our own datacenter on our own hardware. Azure Government or your own servers on request.
Built for 10 CFR Part 810
How Performance Improvement AI is hosted, operated and supported is set up for 10 CFR Part 810 export controls.
Nuclearn platform sign-in
Single sign-on through your identity provider, or the platform's own accounts. There is no separate user store.
Dataset permissions
Reading, scoring and analyzing a dataset follow its permissions, and reports check them for each bucket they use.
Keyed access for tools
Power BI and your own tools read results with a Nuclearn platform API key.
A certified company
Nuclearn is ISO/IEC 27001:2022 certified.
Start with the reports you already have.
Trend coding and analytics work on a CAP export on their own, or alongside the reports CAP AI screens.
Performance Improvement AI
Trend coding, analytics, buckets and reports on a CAP export or the condition reports already on your platform.
Nuclearn CAP AI
Screening, similar past reports and the condition report record. cap.nuclearn.ai
Power BI
Connect the analysis and the coded reports to your own dashboards.
Observation conversations
A guided conversation about an observed activity, graded against your criteria. In development.
What it is not
Performance Improvement AI doesn't decide what counts as an adverse trend, and it doesn't write your corrective actions. It codes, counts and compares; your performance improvement team and your program decide.
Questions plants ask first
How is this different from CAP AI?
CAP AI works on each report as it comes in: screening, similar past reports and the condition report record. Performance Improvement AI works on the reports in aggregate: the trend codes, the analytics that compare them, the buckets you follow and the trend report. Both run on the Nuclearn platform and the same datasets. Screening is described at cap.nuclearn.ai.
Which codes does it use?
WANO performance objectives and criteria: 905 codes in 26 areas, each with its description. The model returns the codes that apply to each report with a confidence, and you choose how many to keep.
What data does it need?
The text of each condition report, as a dataset: upload a CAP export with the description in one column, or use condition report data already on your Nuclearn platform. Other fields, such as unit, department or date, become ways to slice the codes and trend them over time.
Can we compare against the industry?
Yes, if you have a benchmark. Choose a benchmark dataset scored with the same model, and the analysis shows how far each code sits from it; reports label this "relative to industry". Each code in a report also shows how it compares with your own recent history.
Does it decide what counts as an adverse trend?
No. It shows which codes are rising or sit away from the average, with the reports behind them. Whether that is an adverse trend, and what to do about it, is your program's call.
What about observations?
A guided observation conversation, which grades an observed activity against your criteria, is in development. Ask us about it in the demo.
Where does it run, and who can sign in?
On the Nuclearn platform: each plant runs in its own single-tenant instance, hosted by Nuclearn in our own datacenter on our own hardware, with hosting, operations and support set up for 10 CFR Part 810 export controls. Azure Government or your own servers on request. People sign in through the platform, with single sign-on if you use it; Power BI and other tools use a platform API key.
See your reports coded and compared.
We'll show condition reports coded against WANO performance objectives and criteria, the codes that stand out by unit and over time, a bucket followed period to period and the report built from it, then answer your IT team's questions about deployment and sign-in.
- A live walkthrough on public NRC event reports
- Your questions on deployment, sign-in and data
- Which of your trend reports it could build
Nuclearn Performance Improvement AI is backed by Nuclearn, whose products serve over 70 facilities in North America and the U.K.
Book a demo
Tell us a little about your site and we'll set up a time.
Thanks, we've got it.
We'll be in touch to set up a time. If it's urgent, email contact@nuclearn.ai.
Prefer email? contact@nuclearn.ai