Putting a voice model on a real phone number is no longer mostly an AI demo exercise. OpenAI now documents low-latency live audio sessions, SIP call ingress, and server-side sideband control in its Realtime stack. Twilio documents the WebSocket bridge, greeting behavior, language settings, interruptions, and session callbacks through Conversation Relay. Once that stack is live on a published number, the real question shifts. It is not just whether the model can talk. It is whether the call can be answered, directed, handed off, and accounted for without losing the caller along the way.
Articles
By Greg Nowak. Last updated 2026-07-08.
If you inherited a Drupal 8 site and the organisation now wants existing English content to become Swedish source content, treat the job as a controlled legacy-data repair. It is not the same as translating content, and it is not a safe one-click change in the language settings.
By Greg Nowak. Last updated 2026-07-02.
ChatGPT apps are becoming part of the work surface, not just a shortcut for fetching a file or summarising a web page. OpenAI’s current app model includes search, deep research, sync, interactive app experiences, custom apps, and write actions. For a small company, that changes the question.
By Greg Nowak. Last updated 2026-07-04.
AI automation often starts cheaply. One workflow, one model call, one clear job. Then it gets useful. A button begins triggering tool calls, retries, background checks, follow-up prompts, and scheduled work that nobody watches closely every day.
By Greg Nowak. Last updated 2026-07-08.
AI integrations are no longer just helping someone write a first draft. They are being wired into practical jobs: update a CRM record, create a support ticket, prepare invoice data, route a lead, publish metadata, notify a team.
That is where the risk changes. A weak paragraph usually gets noticed by a person. A bad field inside a JSON object can move quietly through an API workflow and only show up later as a messy record, a wrong status, or an automation that did exactly what the software allowed it to do.
OpenAI has made file-based retrieval look almost too easy. Create a vector store, upload files, and the platform handles chunking, embeddings, and indexing for you. That is great for getting a demo running. It is not the same as having a knowledge assistant you would trust inside a real business. The same documentation that makes the setup look simple also makes the production challenge obvious: retrieval stops being mainly an API problem and turns into a governance problem.
By Greg Nowak. Last updated 2026-07-07.
WordPress custom fields are useful because they separate important business information from page copy. Instead of asking editors to paste a client sector, event date, download URL, campaign owner, or product specification into a long content block, you store that information as named data. The result is easier editing, cleaner templates, more reliable filtering, and less manual rework when the same content needs to appear in several places.
By Greg Nowak. Last updated 2026-07-07.
Inline image pasting is a small Drupal feature with a real operational payoff. When editors are publishing screenshots, support notes, QA findings, product callouts, or agency handoff documentation, the difference between paste-and-continue and download-upload-insert can decide whether the CMS feels helpful or hostile.
By Greg Nowak. Last updated 2026-07-05.
Cloudflare Workers are usually introduced for sensible reasons. A webhook needs cleaning up before it reaches the CRM. A redirect needs a little more logic. A form has to call an API. A small endpoint needs to run close to users without standing up more server infrastructure.
By Greg Nowak. Last updated 2026-07-06.
Agentic AI changes the practical risk of automation. A chatbot can give a poor answer and stop there. An agent connected to a CRM, helpdesk, mailbox, file store, or website can give a poor answer and then act on it.
By Greg Nowak. Last updated 2026-07-07.
AI search has made an old SEO bargain feel much less reliable. The old bargain was clear enough: let search engines crawl the site, win visibility, and turn the resulting visits into revenue, leads, subscribers, support deflection, or product discovery.
By Greg Nowak. Last updated 2026-07-06.
Most support bot projects start with a reasonable assumption: the answers are already in the help center. Often, they are. The awkward part comes later, when the help center changes and nobody is quite sure what the bot still has access to.
By Greg Nowak. Last updated 2026-07-07.
Product visibility used to look like a mostly front-end job. Improve the copy. Tighten the category pages. Add reviews. Tune the ads. Make the search result more appealing.
That work still matters, but it is no longer enough. For ecommerce teams, a lot of the commercial risk now sits underneath the page: identifiers, feed attributes, structured data, crawlable product content, return policies, prices, availability, and the checks that keep those things saying the same thing.
A few polished demos can get an internal AI workflow pilot approved. They should not be enough to release it into real operations. A system that summarizes tickets, drafts replies, routes requests, or prepares records can look reliable in a calm test and still fail when the prompt changes, the model is upgraded, a tool returns messy data, or a user asks in an unexpected way. Once the workflow touches customer service, internal operations, or decision support, the standard needs to change.
By Greg Nowak. Last updated 2026-07-07.
Debugging PHP is not a personality test for developers. It is a visibility problem. When a checkout fails, a webhook disappears, or a legacy template throws a blank page, the costly part is usually not the fix itself. It is the time spent guessing what the application actually saw, which branch it took, and where the failure first became visible.
By Greg Nowak. Last updated 2026-07-07.
AI shopping is changing a very practical part of ecommerce: the quality of the product data that customers, search systems, and shopping agents rely on. This is not just an SEO concern. It is an operations problem that touches merchandising, templates, feeds, analytics, support policies, and release management.
By Greg Nowak. Last updated 2026-07-07.
Composer is not just a developer convenience in a Drupal project. It is part of the delivery system. It decides what gets built, which versions the team shares, and whether staging, production, and local environments can be explained without guesswork.
By Greg Nowak. Last updated 2026-07-07.
Drupal AI has moved past the lab-demo stage. Drupal’s AI overview presents it as an enterprise-grade open source AI solution, with AI-human collaboration, governance, model flexibility, and demos for content, accessibility, and marketing work. That is a good sales story. It also means agencies need to treat client demos with more discipline than a polished prompt walkthrough.
Operations teams have wanted software to take care of repetitive browser work for years: supplier portals, admin consoles, claims dashboards, finance back offices. What has changed is not just model quality. OpenAI’s current computer-use guidance lays out practical ways to run browser agents through screenshots, structured UI actions, existing automation harnesses, and code-execution runtimes. That makes narrow pilots realistic. It also raises the cost of getting the rollout wrong.
By Greg Nowak. Last updated 2026-07-07.
If you inherited a Drupal 8 site, inline responsive images are not mainly a design polish task. They are an operations problem. Editors need a predictable way to place images inside body copy without guessing pixel widths, uploading oversized files, or breaking mobile layouts after every content update.