AI Consulting FAQ: Questions to Ask Before You Hire

AI can take repetitive work off your plate, like drafting reports, proposals and email, answering common customer questions, and moving data between apps that don't talk to each other. We help small and mid-sized businesses put it to work through strategy, automation, custom AI builds and training, all in plain English. The goal is AI your team can actually use, instead of another tool that gets bought and never adopted.

Start small. Pick one or two tasks that eat up your week, then test AI on just those with a small pilot before you spend more. If you know AI should be helping but you don't know where to start, that's normal, and it's what our free 20-minute consult is for. We'll talk through how your business runs and point you to a sensible first step.

No. We work with owners and operators who wear ten hats, and with office managers, admins, HR staff and agents who don't write code. We explain everything in plain English, without the hype or the jargon. Businesses shouldn't need a PhD in machine learning to benefit from AI, and you won't need one to work with us.

That's not how we approach it. Our goal is to add capacity to the team you already have, not to cut people. In practice, that means using AI for the repetitive parts of a job, like first drafts or entering the same data twice, so your people have more time for the work that needs them. We'll also tell you plainly when a task isn't a good fit for AI.

It depends on the scope, so we don't list prices. After a free 20-minute consult and a scoping conversation, you get a written proposal with a fixed price or a range. Cost mostly comes down to scope: how many workflows, apps or systems are involved, and whether you want training or ongoing support afterward. Starting with a small pilot keeps the first step small too.

Payments are tied to milestones, starting with a deposit, so you pay as the work moves forward instead of all at once. If the scope changes partway through, we put it in a written change order for you to approve before anything changes. That way you'll never get a surprise invoice.

Sometimes, and we'll tell you about them up front. Hosting and domain costs are passed through to you. Ongoing support is optional and available as a monthly plan. For websites and apps, the optional maintenance plan covers security fixes, updates and monitoring after launch. Being honest about cost is part of how we work, so nothing about ongoing costs should catch you off guard.

Next comes a scoping conversation, where we dig into the workflow or problem you want to solve. Then you get a written proposal with a fixed price or a range. Once you approve it and pay the deposit, the work starts, and you get progress updates every week or two until it's done.

We follow four steps: Discover, Design, Build and Deploy. First we learn how your business runs today, then we design the fix, build it, and put it to work. We ship a first version fast and improve it from there, and you stay in the loop the whole way with weekly or biweekly progress updates.

It depends on how much you want built, but it shouldn't take six months to see results. We start with a small pilot on one or two workflows, ship a first version fast, and improve it from there. You'll get progress updates every week or two, so you always know where things stand and what's next.

Not on their business plans by default. Anthropic says data from Claude Team, Claude Enterprise and the Claude API (the connection other software uses to talk to Claude) isn't used to train its models by default, and OpenAI says the same of ChatGPT Business, Enterprise and Edu and its API. Claude's consumer plans (Free, Pro and Max) can be used for training unless the user opts out. For company work, a business plan or the API is the safer place to start.

For automations, yes. n8n, the workflow automation tool we build with, can be self-hosted, which means it runs on your own server and your workflow data stays there. One thing to know: any AI service a workflow calls, such as Claude's API, still receives the information you send it. If keeping data in-house matters to you, tell us early and we'll plan around it.

That depends on which tool you use, which plan you're on and how it's set up, so it's worth checking before anyone pastes in customer details. As part of our AI consulting, we help you choose tools and plans that fit the information you handle. We're upfront about the risks, too, including when a task shouldn't go through AI at all.

An AI consultant helps you figure out where AI will actually help your business and how to put it in place. Our consulting covers AI readiness assessments, choosing tools and models, implementation roadmaps and ROI analysis. We also review the AI tools you already pay for, so you can decide what to keep and what to cut.

An AI pilot is a small test of AI on one or two of your workflows before you roll it out further. We build a first version fast, keep you in the loop with regular updates, and improve it from there. If it pays off, you scale up. If it doesn't, you've found out on a small project instead of a big one.

Measuring AI's return starts with the work you want it to help with: how long that work takes today and what it costs you. ROI analysis is part of our consulting, and we pair it with a small pilot so you can see real results on one or two workflows before you invest more. We'll also review the AI tools you already pay for, since cutting unused ones counts too.

Our Claude setup is a paid implementation for teams that want to skip the learning curve. We set up Claude for your business in Claude Desktop and Cowork, connect it to the tools your team already uses, such as Google Workspace or Slack, build reusable projects and workflows, and train your staff to use them day to day.

It depends on the work you need done. Choosing the right tool and model is part of our AI consulting, and we work with tools from both Anthropic, which makes Claude, and OpenAI, which makes ChatGPT. Our founder has used Claude in his own businesses since early 2025, and it's the AI assistant we set up and train teams on.

Good candidates are repetitive, rules-based tasks that move information from one place to another. For example, we built a nightly pipeline for a real estate investment company that pulls new foreclosure filings from public county court records, uses Claude to read the filing PDFs, and delivers outreach-ready leads to a Google Sheet with a morning summary. We also connect apps that don't talk to each other, so nobody retypes the same data.

n8n is a workflow automation tool that connects your apps and runs multi-step tasks for you, either on a schedule or when something happens, like a new form entry. We design and deploy n8n workflows for businesses. It can run in the cloud or be self-hosted on your own server, which keeps your workflow data in-house.

Yes. An n8n workflow can send a task to Claude through Claude's API, the connection other software uses to talk to Claude, and use the answer in its next step. Our nightly lead pipeline works this way: it pulls new court filings, Claude reads the PDFs and pulls out the property details, and the results land in a Google Sheet.

An MCP server is a connector that lets an AI assistant such as Claude work with one of your business systems. MCP stands for Model Context Protocol, a standard way for AI assistants to connect to other software. With an MCP server in place, the assistant can look things up in that system or take actions there. We build custom MCP servers for a business's own tools and data.

An AI agent works through a multi-step job on its own, while a chatbot answers questions in a conversation. For example, we run a scheduled Claude agent every morning that scans AI news, docs and community sources, then writes a short brief on what changed and why it matters, with sources. A chatbot waits for someone to ask. An agent can do the work on a schedule.

That's exactly what an MCP server is for. We build MCP servers that let AI assistants such as Claude work with a business's own systems, like a CRM or a database. Whether it's a good fit for your setup depends on the systems you use, so we'd start by looking at those together on a free consult.

We build a knowledge base from your business's own information, then train the chatbot on it, so it answers questions about your business using what you've given it. Our custom chatbots are built on Botpress, a chatbot platform, and can answer customer questions and qualify leads. Getting the knowledge base right is where the work starts, because it shapes every answer the chatbot gives.

Yes. Along with answering questions, the custom chatbots we build can qualify leads and guide visitors to the right next step. One chatbot we built on Botpress, trained on a knowledge base we created from the ground up, routes each conversation, asks follow-up questions that fit what the visitor said, and follows an engagement workflow tailored to that business.

A chatbot has two kinds of cost: building it and running it. The build is priced in a written proposal with a fixed price or a range, paid in milestones. Running it can include hosting, which is passed through to you, plus an optional monthly support plan if you want help keeping it up to date. We'll lay out the monthly side before you commit.

We build web apps with Next.js and React, mobile apps for iPhone and Android, and a secure database and user management with Supabase, plus connections to the software you already use. For small businesses, we also offer a brand kit and a mobile-friendly, SEO-ready website that can take bookings and payments. After launch, an optional monthly maintenance plan covers security fixes, updates and monitoring.

Yes. We build mobile apps for iPhone and Android from one codebase, meaning a single set of code runs on both instead of two separate apps. That gives you one app to plan, build and update rather than two. If you'd rather start with a web app, we build those too, with Next.js and React.

It depends on the scope. A focused app can come together quickly: our own ACT prep web app, Cram ACT, was built in about 36 hours. Projects that connect to other systems, like a booking site that creates jobs and invoices in field-service software, need more time for testing. Either way, we ship a first version fast and improve it from there.

Yes. We're based in Tulsa, Oklahoma, and work on-site with businesses across the Tulsa metro, including Broken Arrow. For teams elsewhere in the US, we work remotely. Our workshops and training are hands-on and in plain English, built for business owners and their teams rather than developers.

Pick one task you do every week that eats up time, like a report, a proposal or a batch of emails, and practice using an AI assistant on just that. Once it saves you time, add the next task. Hands-on training helps because you learn by doing, with someone there to answer your questions in plain English.

Our AI training is for business owners and their teams, including non-technical staff like office managers, admins, HR and agents. You don't need any coding background. If you're a developer or AI researcher looking for deep technical content, our workshops probably aren't the right fit. They're built for people who want AI to help with everyday work.

A Claude skill is a saved set of instructions, sometimes with files, that teaches Claude how to do one specific task your way. Instead of explaining the task from scratch each time, your team uses the skill and Claude follows the same steps. Skills can be bundled into a plugin, like the brand-voice plugin we built with four skills and six commands.

A skill teaches Claude one task. A plugin bundles several skills and commands so a team can install them together. A subagent is a specialized helper that Claude Code, Anthropic's coding tool, hands a specific job to. For an app launch, we built subagents for beta-feedback triage, growth outreach and release QA, including one that runs a pre-release checklist and makes a go/no-go call.

A custom Claude plugin turns the work you repeat most into commands your team can run in Claude. Our brand-voice plugin, for example, drafts social posts, content calendars, how-to guides and proposals, repurposes content and researches topics, while keeping every draft in one brand voice. A plugin built for your business would follow your own process and style.

We're based in Tulsa, Oklahoma, in the Broken Arrow area. We work on-site with businesses across the Tulsa metro and remotely with businesses anywhere in the US, so you don't need to be local to work with us. The easiest way to reach us is by email or through the contact form.

Email admin@legacyaisolutions.ai or send a message through the contact form, and tell us a little about your business and what you'd like AI to help with. From there, we'll set up your free 20-minute consult.

Yes. The first consultation is a free 20-minute call. We'll talk about your business, where your time goes, and whether AI is a good fit. If it makes sense to keep going, the next step is a scoping conversation and then a written proposal. There's no cost to find out whether we can help.

LegacyAI Solutions is founder-led. Nate Araskog, the founder and owner, is based in Tulsa and has used Claude in his own businesses since early 2025. He completed the AI Generalist Accelerator Program by Outskill. We teach and build, and everything we teach has been tested somewhere it actually mattered.