LuminQuestions & answers / Lumin Marketing Group
Answers to common questions about AI integration, business chatbots, voice assistants, paid advertising, project costs and working with Lumin Marketing Group.
Start here for plain-language answers about what Lumin can build, how projects are scoped and what stays with your team. These answers describe our services and approach; your proposal defines the agreed deliverables.
Reviewed 13 September 2026
Lumin builds AI integrations, business assistants, connected workflows and custom applications. We also provide paid advertising support. We are based in Edmonton, Alberta, and begin with your business problem, existing tools and an agreed project scope.
Yes. Paid advertising support can be a standalone engagement. You do not need to buy an AI project. If inquiry handling or follow-up also needs attention, we can discuss that as a separate or connected scope.
Bring one task or campaign you want to improve, the tools you currently use, a budget range and any target date. A sample of the current process is helpful; remove private customer information first. You do not need a technical specification.
Lumin quotes after reviewing the work. The scope, number of systems, access requirements, testing, training and ongoing support affect the price. Ask for service fees, advertising spend, subscriptions, AI usage and hosting to be listed separately. We do not publish a universal package price.
Timing is agreed after scoping. A focused workflow and a custom application have different delivery needs. Account access, data readiness, review time and testing affect the schedule. We confirm milestones before making a delivery commitment.
Original video and local storytelling are offered through Edmonton Spotlight, Lumin’s video and media brand. Production can be scoped separately from an AI build or advertising engagement.
AI integration connects a model or assistant to a defined business process and the tools that process uses. For example, a system could classify an inquiry, draft a response from approved information and prepare a handoff for staff. Each action, data source and approval step needs a clear scope.
Choose a repeated task with a clear input and an output someone can check. Examples include routing inquiries, finding approved information or preparing a document draft. Measure the current time and error rate, then test a small pilot before expanding it.
Yes. Lumin assesses Claude, OpenAI, Gemini, local models, Cursor, AWS and other suitable tools for a project. This is not a fixed tool list. Availability, compatibility, permissions, privacy requirements and operating cost determine the design; a provider name is not a claim of endorsement or certification.
Not necessarily. We first assess your existing website, CRM, email, calendar and other systems. Supported APIs or connectors may make an integration possible. Exact product plans, administrator permissions and allowed data flows must be checked before we promise a connection.
A local model may be an option when the selected software, hardware and workload support it. A local model does not make every connected workflow offline. We assess external connectors, updates, storage, maintenance and performance as part of the design.
Yes. A custom application may be more useful when staff need a structured interface, repeatable steps or a specific tool. Our work includes PKTD, an iPhone receipt and expense app. We choose the interface around the task rather than making every solution a chat window.
Agree on a baseline and a test before the build: completion time, accuracy, rework, response time or another relevant outcome. Include staff review and running costs. A plausible answer is not proof of a correct result, and time saved does not automatically become revenue.
Yes, an assistant can be built around approved information about your services and processes. We define what it may answer, when it should ask a clarifying question and when it must hand over to a person. Knowledge sources and representative questions are part of testing.
Voice can be included when the chosen technology and visitor device support it. Lumi on this website offers optional live voice after the visitor starts a conversation and grants microphone permission. Browser playback and usage limits apply; typing remains an alternative.
Yes. A scoped workflow can collect visitor-approved contact details, save the inquiry and notify the responsible team. On Lumin’s website, the contact form asks for consent and sends the details to our private Slack leads channel. A delivery receipt confirms system delivery, not that a person has read or answered it.
That can be assessed as a separate integration. Lumi currently sends notifications to Lumin’s private Slack channel; staff replies in Slack are internal and are not sent back to visitors. Do not assume a notification connection includes two-way customer support.
Booking links and calendar integrations are different scopes. Lumi can direct visitors to Lumin’s discovery-call booking page. A system that creates, changes or cancels appointments needs separate rules, calendar permissions and end-to-end testing.
We build around approved sources, test difficult questions and provide a human route. The assistant should acknowledge uncertainty rather than invent an answer. Billing decisions, sensitive account changes and other consequential actions need the agreed staff approval process.
No. Question patterns can be recorded for human review under the site’s disclosure. Visitor statements do not automatically become approved business knowledge or train a new model. Changes to the answer guide should be checked before use.
Lumin stores chat history, voice transcripts and contact records in Cloudflare D1. AI conversations are processed by OpenAI; optional voice streams audio to OpenAI. Lumin does not store raw microphone audio. The contact form sends its details to the private Slack leads channel with consent. Read the site’s Disclaimer and Terms for the full disclosure, and avoid sending sensitive information.
We assess Google Ads, Meta, Microsoft, Apple, ChatGPT advertising and other paid channels against your audience, offer and budget. Availability, region, eligibility, formats and account access must be checked before promising a launch. A platform mentioned here is not a guarantee that every business can advertise on it.
An agreed scope can include campaign assessment, planning, account setup, audience and messaging work, creative, landing pages, conversion measurement, testing and ongoing refinement. Confirm the deliverables and who is responsible for approvals, media spend and customer follow-up.
There is no useful one-size-fits-all budget. Start with your offer, location, customer value, margins, available follow-up capacity and a clear test objective. The team can propose a test budget after assessing those inputs. Media spend is separate from service fees unless a quote says otherwise.
No. A click is a visit, and a conversion is a configured action. Neither necessarily represents a qualified inquiry or sale. Agree on what is counted, test the tracking and compare reported actions with actual inquiries, bookings and customers before judging performance.
No. Outcomes depend on the offer, market, budget, website, tracking and follow-up. We can agree on work, tests, reporting and decision points. A performance forecast should be labelled as an estimate, not a guarantee.
Yes, a useful project may connect the landing page, an assistant, an inquiry form and staff notification. The goal is to help a visitor get an answer and choose a next step. Advertising, consent, follow-up and measurement should be designed together; the assistant does not replace staff accountability.
Use the contact page, email info@luminmarketinggroup.com or call (825) 967-9337. You can also book a discovery call. Existing clients should explain the project and issue without including passwords, payment details or private customer records.
Tell us what happens today and what you want to improve. We can discuss the fit, required access and a useful next step.
Talk to the Lumin team ↗