Nine modules.
One system.

Knowledge base + RAG

The brain of the entire system, and the single most important thing we build, because almost every other module draws its intelligence from here.

Right now everything BrightLink knows about its products, customers, installs, and past conversations lives scattered across Shopify, email threads, spreadsheets, spec sheets, and your own head. None of it is searchable by software, and none of it can be put to work automatically. This module converts all of it into a structured, searchable knowledge base. Every product, quote, install story, customer record, and spec sheet becomes a chunk the AI can pull from on demand.

RAG means retrieval augmented generation. In plain terms, the AI does not make things up. Before it produces anything, it retrieves the actual facts from your knowledge base, then writes grounded in those facts. That is the difference between generic AI output and content that reflects BrightLink's real products, real specs, and real customer language.

Customer support is only the most obvious use. The far bigger value is everything else this brain powers. Sister site content generated at scale. Cold email that is dramatically more effective because every message is personalized from real data. Email marketing tailored to what each segment actually bought. Blog posts and programmatic SEO pages produced in volume, each accurate and targeting the long tail of searches buyers make. Every demand generation channel in this proposal gets sharper because it is fed by this one source of truth.

It is not a support feature. It is the engine underneath the revenue generating channels. Build it once, and every other surface inherits the intelligence.

Shopify Admin integration

The plumbing that keeps the brain honest. A knowledge base is only as good as the data inside it, and BrightLink's most important data lives in Shopify.

Live inventory levels, order history, customer records, product details, and the custom meta fields and meta objects that hold spec information. This module keeps the knowledge base in continuous sync with Shopify. When inventory changes, the system knows. When a new order comes in, the system sees it. When specs update, every surface that draws on the brain starts using the new numbers. Without this, the AI would answer from a stale snapshot and tell a customer something is in stock when it sold out last week.

It connects directly to the Shopify Admin API, the same trusted channel Shopify uses internally. It pulls inventory across all warehouse locations, reads order and customer data, and reads the structured spec fields on each product. This module is invisible to customers, which is exactly the point. Good plumbing is never noticed.

The value here is trust. A system that gives wrong answers about stock or pricing destroys customer confidence faster than having no system at all. It also makes the data usable for segmentation, so marketing and outbound can target customers based on what they actually bought, when, and how much they spent.

Customer chatbot

A chat assistant on the public BrightLink website, available every hour of every day, answering the questions that currently go unanswered overnight or eat up your team's time during the day.

A customer lands on a product page with questions: pixel pitch, will this ship to my state, power requirement, outdoor use, warranty. Instead of leaving to find a competitor or filling out a contact form and waiting, they get an immediate answer with the real specs and real pricing, not generic guesses, so they trust it enough to act on it.

It does three things beyond answering. It qualifies leads, understanding whether a visitor is a casual browser or a serious buyer with a real project and budget. It books calls, offering to schedule directly on your Cal.com calendar when a lead is ready, turning a visit into a booked conversation without anyone lifting a finger. And it escalates, flagging high value signals like a large project or a recognizable enterprise name so you can step in personally on the deals that matter most.

The value is conversion and time. Customers who would have bounced instead get answers and book calls. Your team stops answering the same twenty questions and focuses on closing. The website stops being a brochure and starts being a salesperson that never sleeps.

AI voice agent

An AI voice agent on BrightLink's existing phone line, answering calls twenty four hours a day, designed to replace the AnswerConnect service and do far more than take messages.

When someone calls, the agent answers in a natural voice, understands what they are calling about, and answers product questions, confirms shipping options, explains specs, and handles the routine calls that currently go to voicemail or to a paid answering service that just takes a name and number.

Like the chatbot, it qualifies, books appointments, and escalates, but on the phone. A caller mentioning a large multi location project does not get parked in a voicemail queue. The agent recognizes the opportunity and gets it in front of you immediately, day or night.

The cost math alone is compelling. AnswerConnect would be eliminated, saving four thousand two hundred dollars per year. But the real value is what AnswerConnect never did: it never answered a product question, qualified a lead, or booked a call. It just took messages. This turns inbound calls into booked appointments and captured leads, which is what a phone line is supposed to do for a business trying to grow.

This is one of two voice agents. This one faces your customers. A second one faces you, the brain-dump capture agent, which calls you to capture your own knowledge by voice. That one is part of Module 09.

Customer-gated portals

A private, logged in experience for BrightLink's larger accounts. The first target is Forte, the fifty one location account with ten open quotes. The same system then serves every multi location customer that follows.

Today, when anyone at one of Forte's fifty one locations has a question about a quote, an order, or a spec, they wait for you. You are the bottleneck. With a gated portal, Forte logs in and sees their own world: their open quotes, full install history, spec sheets for the products they buy, and the status of anything in progress. The chatbot, in this logged in context, answers questions about their specific account, not just generic product info.

This changes the relationship. Instead of the vendor they have to call and wait on, you become the vendor who gave them a tool their whole team can use, at any location, any time. When you fly to Minnesota to train their team, you could walk in with a system that already knows their quotes and answers their questions live. That is what a fifty one location company talks about internally, which is exactly how you get invited into the other locations.

This is a revenue lever, not a support feature. Making it effortless for a multi location account to see their quotes and reorder is how you turn ten open Forte quotes into ongoing business across fifty one sites. The architecture is built once and reused: church groups with multiple campuses, hockey rink chains, school districts, restaurant and retail groups. Forte is the template, and every account after gets the gated experience with little additional work.

Sister sites

A network of focused landing sites, each targeting a specific vertical or geographic market the main BrightLink site cannot rank for on its own. BrightLink Vancouver, BrightLink Ontario, BrightLink Church LED, BrightLink Hockey Rink.

A single broad website struggles to rank for every narrow search a buyer makes. Someone searching for LED video walls for a church has a very different need than someone searching for a hockey rink scoreboard or a Vancouver based AV installer. A dedicated site speaking directly to one audience, with copy tuned to their pain points, ranks better and converts better than a general page trying to speak to everyone.

The AI writes the content for each site, drawing from the knowledge base. It generates copy for each market's specific concerns: the church that needs easy operation, the hockey rink that needs brightness and durability, the corporate buyer that needs reliability. Each site gets specs, application examples, and messaging written for that exact audience, at a scale impossible to produce by hand. High quality content at volume is exactly what wins this kind of search traffic.

Each site does two jobs. It captures narrow, high intent searches the main site misses. And it provides its own sending domain for outbound, which lands in inboxes far better than blasting from the main domain. Leads from any sister site flow into the same funnel, the same chatbot, the same booking flow. BrightLink already attracts buyers like NASA, the Air Force, churches, and hockey rinks without marketing to any of them. Sister sites turn each into a dedicated front door.

Cold outbound (B2B Rocket replacement)

Replaces B2B Rocket, the cold email tool currently producing nothing, with a system that actually generates replies.

B2B Rocket is at zero percent opens and seventy four percent bounce on the last several thousand emails. That is not a campaign that needs tuning. That is a dead channel. The reason it failed is two things: bad lists and bad sending reputation. Bouncing at seventy four percent means the lists are full of dead addresses, and sending that volume of bouncing mail destroys the domain's reputation, which is why opens collapsed to zero.

This module fixes both. Clean, verified lists so messages reach real people. The sender domains from the sister sites, so outbound for a vertical comes from a domain dedicated to that vertical with its own clean reputation, instead of damaging the main BrightLink domain. And the AI personalizes each message using the knowledge base. Instead of a template blasted to thousands, each prospect gets copy that speaks to their specific situation, industry, and likely needs.

It identifies prospects matching BrightLink's best customer profiles, writes personalized outreach, sends from healthy domains, and routes replies into the same funnel as every other lead. The value is a channel that works versus a channel that is dead. You are currently paying for outbound that produces silence. This turns that spend into actual conversations with real prospects.

Admin dashboard

The single screen where you see everything the system is doing and decide where to push next. Every other module generates activity. This one turns that activity into information you can run the business on.

In one place: how many leads are coming in, how many conversations the chatbot and voice agent are having, how many calls are booked, how many quotes drafted, how many deals closed. Crucially, it shows attribution, which channel each result came from. You will see whether a lead came from the main site, a specific sister site, the voice agent, or outbound, so you know what is actually working.

It surfaces patterns you cannot see today. Which objections the bot hears most, so you know what is stopping sales. Which products are getting interest this week, so you know where demand is moving. Which markets are converting, so you know where to invest next. This is information that currently lives nowhere, because no one has time to track it by hand across phone, email, and the website.

This is the decision surface. When you decide whether to launch another sister site, it tells you which existing ones convert. When you decide what to stock, it tells you what people ask about. When you wonder whether the system is paying for itself, it shows you the deals it sourced. A system doing this much work could become a black box. This makes sure it never does.

CRM + capture loop

What makes the entire system improve on its own over time, without anyone manually feeding it. The loop that turns every interaction into fuel for the next one.

Every conversation the chatbot has, every call the voice agent takes, every quote drafted, every install completed becomes new data captured back into the knowledge base. The objection a customer raised yesterday becomes information the system uses to handle it better tomorrow. The install completed last week becomes a reference story for the next similar prospect. The question no one could answer becomes a gap the system flags and fills.

There is also a part built specifically for you: the brain-dump capture agent. A second voice agent that does the opposite job of the customer-facing one. It calls you with a short agenda of the highest-impact knowledge gaps from recent conversations. You answer by voice for a couple of minutes, a quick brain dump, and it asks follow-up questions, then writes your answers straight into the knowledge base, attached to the right products, customers, and topics. This is how the system captures the knowledge that lives only in your head, the install stories, why customers chose you, the local deals, with zero typing. You said this is your department to stay on top of but not work you want to do by hand. This is the answer: you talk for two minutes, the system does the rest.

This is the difference between a system that is static and one that compounds. A normal chatbot is built once and slowly goes stale. This is designed to do the opposite. The more it runs, the more it learns. The more conversations it has, the better its answers. The more installs BrightLink completes, the richer its library of proof becomes. None of that improvement requires someone sitting down to write content.

The CRM side keeps every lead, conversation, and customer in one place with full history, so nothing falls through the cracks and no lead gets the same pitch twice. The value is compounding return. Most software depreciates the moment it ships. This is designed so the system appreciates instead. Six months in it knows more than at launch. A year in it is materially better, because every interaction made it so.