Responsible AI in the public sector: principles and practice
What responsible use of AI means for agencies, the frameworks that guide it, and concrete practices.
Articles
AI & Automation
Every reference article on hqt.org in AI & Automation, written for agency procurement, technology, and program staff.
Generative AI in agency operations: where it helps today
The agency tasks where large language models are already useful, the ones where they are not yet, and how to tell the difference.
Retrieval-augmented generation: AI assistants that cite agency sources
How an assistant can answer questions from the municipal code, policies, and records without inventing answers.
AI-assisted redaction for public records requests
How machine learning can find personal information, criminal justice data, and other exempt content faster, with a person making the final call.
Intelligent document processing: turning forms and PDFs into data
How agencies extract structured data from applications, invoices, and scanned records with modern document AI.
AI security: prompt injection and other risks agencies should plan for
The new attack surface that language model applications create, and the controls that address it.
Automating agency workflows with integration and AI
Where automation removes manual steps between systems, and where AI extends it to tasks that used to need a person to read.
Virtual agents for 311 and resident services
What a well-designed AI assistant for residents can do, the failure modes to avoid, and the requirements to set.
Building an AI inventory: tracking where AI is used in your agency
What belongs in an agency AI use case inventory, how to find AI already embedded in licensed products, and who keeps the record current.
AI impact assessments: a practical process before a system goes live
What an AI impact assessment contains section by section, who signs it, when it is refreshed, and how much work it takes.
Writing a generative AI use policy for agency staff
What a staff generative AI policy has to decide: approved tools, data limits, review obligations, disclosure, and personal accounts.
Bias testing for agency AI: what to measure and how often
How to test whether an agency AI system produces uneven results: comparison groups, error rates, thresholds, and retesting.
Designing human review for AI outputs in agency workflows
How to design the human review step in an AI workflow so it catches real errors: triage, evidence, workload, and override data.
Machine translation and language access for agency communications
Where machine translation fits in agency language access, which content still needs qualified human translators, and how to say so.
AI coding assistants on agency software projects: policy and review
How agencies and their contractors should govern AI coding assistants: data exposure, licensing, ownership, review, and disclosure.
AI transcription for public meetings: minutes, captions, and the record
How councils and boards use speech recognition for meeting transcripts and live captions, and where it goes wrong.
AI agents that take actions: permissions, logging, and limits
How agencies control AI agents that update records and send notices: service accounts, permitted actions, thresholds, and logs.
Automating accounts payable: invoice capture, routing, and controls
How agency finance teams automate invoice intake, coding, and approval routing while keeping vendor bank change verification human.
Automating reconciliations: bank, cash receipts, and subledger matching
How rule-based and machine-assisted matching can clear routine reconciliations, and how to design exception queues auditors accept.
Onboarding automation: HR as the source of truth for accounts and access
How agencies can drive account, email, badge, device, and training provisioning from HR records, and where approvals belong.
Fault detection and diagnostics: turning building automation data into work orders
How fault detection and diagnostics software turns building automation trend data into work orders, and what it takes to act on them.
Video analytics and license plate readers: accuracy, alerts, and policy limits
What camera analytics and license plate readers actually detect, how they fail, and the use, retention, and audit policy an agency should set first.
Triaging resident messages with AI: classification, routing, confidence
How agencies can classify and route inbound resident messages with AI while keeping staff in charge of every outcome.
AI scribes and ambient documentation in agency health clinics
How public health, behavioral health and correctional clinics should evaluate AI scribes: review, consent, recordings, and contract terms.
Using AI to sort and summarize public comments on environmental documents
How planning staff can use language models to cluster, deduplicate, and route comments on an EIR or general plan update without losing the record.
RPA bots or a real integration: cost, fragility, and when each fits
A practical comparison of screen-level RPA against API, file, and middleware integration, with the conditions that favor each.
Automating document intake: classification, routing, and exception queues
How agencies classify incoming documents, route them to the right queue, set confidence thresholds, and keep exceptions worked.
Entity resolution: matching people and businesses across agency systems
How agencies match resident, business, parcel, and vehicle records across systems with no shared identifier, and what to log.
Forecasting demand and staffing levels with your own agency data
How to turn a few years of your own permit, call, dispatch, and counter history into staffing forecasts you can defend.
Anomaly detection in agency financial transactions and card spend
How agencies can use rules and models to flag duplicate payments, split purchases, and card misuse without creating an unworkable alert queue.
AI-assisted plan review: what it can check and what it cannot
A guide to which parts of building and zoning plan review can be automated today, and which decisions must stay with a human reviewer.
Computer vision for code enforcement: imagery, privacy, and due process
How agencies use vehicle, drone, and aerial imagery to detect code violations, and the privacy, retention, and due process rules that govern it.
Model drift monitoring: watching an AI system after go-live
What to watch after an AI system goes live: weekly and quarterly measures, and the triggers that should force retraining or a pause.
Automate it or staff it: deciding when a workflow is worth automating
How agencies can weigh staff hours saved against build, license, exception, and upkeep costs before automating a workflow.
From AI pilot to production: ownership, runbooks, and funding
What an agency needs before an AI pilot becomes a production system: ownership, runbooks, monitoring, and recurring funding.
Text analytics on 311 comments, complaints, and survey responses
How agencies turn 311 text, survey open ends, and public comment into themes they can track, map, and defend.