Practical guides for teams choosing, building and checking useful software. Explore document workflows and business automation alongside our industrial inspection and exploratory seismic research.
Browse the guidesA shipment checklist is useful only when its requirements come from the agreed workflow and its statuses mean something precise. Follow a fictional account to distinguish missing, received, reviewed, and superseded documents without presenting an administrative checklist as customs advice.
A delegate substitution changes who attends without necessarily changing the booking itself. This guide follows a fictional replacement to preserve the original trail, update the right records, and identify which messages or access arrangements need staff review before anything is sent.
A customer portal is useful only if it supports the tasks customers and staff actually need to complete. This guide helps buyers test existing portal features before commissioning custom development, with a fictional document-collection example, access checks and a pilot that includes support and correction work.
AI evaluation needs an agreed expected answer, including when the evidence does not support a definite answer. This guide explains who should label business examples, how to separate observation from judgement, and how to resolve disagreements without simply treating the model or the oldest record as correct.
Two documents can be readable and still disagree. Cross-document validation makes those disagreements visible, but it needs clear rules about which fields refer to the same thing. Work through a fictional service pack to distinguish safe formatting changes from conflicts that require a person.
Many quotation workflows need cleaner inputs and a reliable template before they need AI. Others involve interpreting varied requests or identifying missing detail. This guide separates those tasks and gives buyers a practical way to choose the smallest system that their staff can review and maintain.
Routing an enquiry means assigning responsibility, not merely sending a notification. This guide shows how a fictional service business can combine service, territory and capacity rules without losing enquiries between teams, then explains which exceptions and reassignment decisions should remain visible to staff.
Combining several service visits into one invoice can hide missing evidence or repeated work. This guide shows a fictional monthly account how to build a visit register, map supporting records, preserve exceptions, and review the exact group of visits proposed for billing.
A different consignee or reference can mean the wrong document was attached, but it can also reflect a legitimate role difference. Use this fictional shipment to build identity checks that preserve context, surface ambiguity, and leave resolution with the responsible logistics operator.
A bespoke training enquiry often mixes a business goal with an assumed course format. This guide develops a fictional programme brief that separates the two, captures practical constraints, and prepares useful clarification questions without inventing learning needs or promising that a programme will solve them.
Low-code workflows can be a practical delivery choice when their logic remains understandable and recoverable. This guide explains how to recognise growing maintenance difficulty, test changes safely and compare a visual workflow with a custom component using the same operational requirements, rather than relying on a simple step-count rule.
A document pilot needs examples that expose the ways the workflow can fail. This guide shows how to choose document families, revisions and awkward cases, separate development from evaluation, and track coverage. A fictional service-pack test set demonstrates why the variety of examples matters alongside their number.
A folder with several attachments can still be incomplete. The practical question is whether the right evidence has arrived for this particular case. This guide explains conditional requirements, document identity and exception handling using a fictional supplier-onboarding pack, without treating file presence as approval.
An RFQ assistant should help a colleague ask the right question, not invent the answer. This guide shows how to turn a conflicting schedule into a short clarification draft, preserve the evidence behind each question and avoid sending unnecessary or commercially sensitive requests automatically.
A won deal is a commercial milestone, not a complete delivery brief. This guide uses a fictional training booking to show what operations needs, how to distinguish approved facts from assumptions and why the receiving team should acknowledge the handoff before the workflow considers it complete.
A signed field ticket can support several charges, and one charge can depend on several tickets. This guide works through a fictional labour-and-parts invoice to show how line-level links, unresolved allocations, and staff decisions can remain visible throughout billing review.
A commercial invoice and packing list can describe the same shipment in different ways. This guide uses a fictional shipment to show how administrative comparison can expose inconsistent references or quantities without treating every difference as an error or making customs decisions.
A group booking and a purchase order may describe the same training differently. This guide uses a fictional cohort to compare the agreed course, customer, dates, and quantities while separating administrative mismatches from commercial decisions that the booking or finance team must resolve.
Custom software is justified by a specific unresolved requirement, not by frustration with a current screen. This guide shows how to test native CRM configuration against real work, identify gaps across import and integration paths, and decide whether a small extension is enough before commissioning a larger build.
A useful baseline describes the whole workflow, including follow-up and correction. This guide explains how to observe work without confusing elapsed time with staff effort, choose consistent counting units and record exceptions. A fictional enquiry illustrates the measurements needed before comparing an automated alternative.
A practical coordination guide to building form, structural continuity, façades, interiors, and recovery goals before detailed seismic analysis.
A shake-table dataset becomes useful when another researcher can identify the tested state, trace every channel, reproduce each processing decision, and compare the measured motion with the intended input.
A closer look at NIST’s Nonstructural Element Database, observation counts, specimen histories, and the evidence needed before modeling.
A table is useful only when each value remains attached to the right row and heading. This guide focuses on purchasing and administrative schedules: how to retain page evidence, expose uncertain matches and prevent a neat spreadsheet from concealing a broken source relationship.
An RFQ does not need every possible detail, but it does need enough information for the kind of price being prepared. This guide uses a fictional display-order request to distinguish missing essentials, acceptable assumptions and questions that should be answered before an estimator starts work.
A repeat message is not always a duplicate sales opportunity, and the same person may represent several projects. This guide separates contact matching from enquiry matching, then shows how to preserve the conversation while preventing repeated intake from creating unnecessary records or hiding new work.
Matching a service report to an invoice takes more than finding a similar address. Use a fictional two-site customer to define the identifiers that matter, distinguish harmless formatting differences from identity conflicts, and give billing staff a traceable comparison they can review.
Completing a service visit does not automatically complete its billing paperwork. This guide uses a fictional maintenance job to separate job completion from invoice readiness, identify missing evidence, and design a pilot that supports billing staff without making payment or technical decisions.
Interface automation can help when a necessary system lacks a usable connection, but it needs a clear operating boundary. This guide compares RPA and API integration through a fictional legacy-system task, including safe stopping conditions, duplicate prevention and the evidence needed before allowing unattended work.
Choosing a document tool starts with the output your team needs. Reading characters, finding fields and deciding whether a record is ready are different jobs. This guide separates them, then shows how to test the smallest useful workflow before committing to a larger build.
A quote assistant can organise a request and explain a calculation without choosing prices itself. This guide uses a fictional service quotation to show how an approved rate sheet, deterministic calculations and a clear exception path can keep commercial decisions with the business.
Fourteen copy-ready prompts for the work around a CUI inspection plan: cleaning a circuit register, exposing missing evidence, building an explainable shortlist, drafting workpacks, backtesting the order and reviewing new findings. Each prompt keeps the source rows attached and makes the model say what it does not know.
Corrosion under insulation is external corrosion of steel that happens where water gets trapped between the metal and its insulation. It is hidden, it is localized, and on the wrong line it can go from sound wall to a leak between two inspection campaigns.
There is no single test for corrosion under insulation. Plants combine a walk-down, a few through-insulation screening techniques and a limited amount of insulation removal. The method choice matters less than the choice of where to apply it first.
Risk-based inspection replaces fixed intervals with a ranking: likelihood of damage times consequence of failure, per item of equipment. For corrosion under insulation, API RP 581 spells out the likelihood side in detail. Here is how that ranking works and how to use it without a full RBI program.
Pulsed eddy current measures the wall of a carbon-steel pipe through its insulation and jacketing, without stripping anything. It is the fastest way to screen an insulated rack for corrosion under insulation, provided the crew knows what the number means and where the method goes blind.
API Recommended Practice 583 is the reference document for corrosion under insulation and fireproofing. This is a working summary of what it covers, how it relates to API 570, API 581 and AMPP SP0198, and which parts an inspection team actually uses.
A corrosion under insulation program is a loop: know what is insulated, rank it, inspect the top of the list, record what you found, repair, and re-rank. This checklist lays out each step with what it needs and what it produces, following the program elements in API RP 583.
The inspection history a CUI program needs usually exists, as PDFs, scanned forms, spreadsheets and photos. AI document extraction can pull the circuit, date, method, readings and findings out of that folder into a table, provided the extraction is reviewable and the exceptions go to a person.
Tell us about the work, the tools, and the bottleneck. We can explore a focused pilot by email.
jamil@clearanvil.com