cat ai-services --verbose
> AI that actually works for your business.
Not hype. Not a chatbot on your website. I help small businesses figure out where AI fits, train their people to use it well, and build automation that multiplies what your team can do — securely, ethically, and without the buzzwords.
01 // reality_check
Where most businesses are right now
Most small businesses are in one of two places with AI: they're ignoring it entirely, or a few employees are using ChatGPT on their phones and pasting company data into free tools with zero oversight. Both are problems.
The first group is falling behind. The second group is creating security and compliance risks they don't even know about. Your employees are going to use AI whether you have a policy or not — the question is whether you're going to get ahead of it and make it work for you, or let it happen in the shadows.
That's where I come in. I take businesses from "we should probably do something about AI" to having a real strategy, trained employees, secure tooling, and automation that gives your team capacity they didn't have before.
02 // strategic_assessment
AI Strategy & Workflow Assessment
Before anything gets built or deployed, I sit down with your team and figure out where AI actually makes sense in your business. Not every process needs AI. Some do. The trick is knowing which ones — and being honest about it.
Workflow Mapping
I walk through your day-to-day operations with the people who actually do the work. Where are people spending hours on repetitive tasks? Where are bottlenecks? What's getting done manually that could be automated? I document your current workflows and identify the spots where AI and automation can make a measurable difference — not theoretical, measurable.
Opportunity Scoring
Not everything that can be automated should be. I score each opportunity based on time savings, implementation complexity, risk, and how much it actually affects your output. A task that takes 20 minutes a week isn't worth building a custom agent for. A process that eats 10 hours a week across your team? That's where we start.
Tool Selection
The AI market is a mess of overlapping products, overpriced subscriptions, and tools that do the same thing with different branding. I cut through that. I'll recommend the right tools for your workflows — whether that's Microsoft Copilot because you're already on M365, Claude for technical and analytical work, or purpose-built vertical tools for your industry. No vendor kickbacks, no commissions. I recommend what works.
Roadmap & Budget
You get a concrete plan with phases, costs, and expected outcomes. Phase one might be getting your team trained and setting up secure access. Phase two might be deploying Copilot across your M365 environment. Phase three might be building custom automation for your specific workflows. You know what's coming, what it costs, and what you'll get from it.
03 // workforce_enablement
Training & AI Enablement
Buying AI tools without training people to use them is like buying a CNC machine and handing someone the manual. The tools are powerful, but only if your people know how to use them properly — and understand what they should and shouldn't be doing with them.
Getting Past the Chatbot
Most people's experience with AI is typing a question into ChatGPT and getting a paragraph back. That's using a jet engine as a desk fan. I train your team on how to actually work with AI — structured prompting, providing context, iterating on outputs, using AI as a thinking partner instead of a search engine. The difference between "write me an email" and a well-structured prompt that produces exactly what you need is enormous, and most people have never been shown the difference.
Role-Specific Training
Your operations manager and your marketing person don't need the same training. I build training around what each role actually does. Accounting gets trained on using AI for financial analysis, report generation, and data reconciliation. Marketing learns content drafting, audience research, and campaign planning. Operations learns process documentation, vendor communication, and project tracking. Everyone gets trained on what matters to their job, not a generic overview.
Ethical AI Usage
AI makes things up. It's called hallucination and it's not optional — it's how the technology works. Your team needs to understand that AI output is a draft, not a source of truth. I train people on verification, fact-checking outputs, understanding model limitations, recognizing when AI is confidently wrong, and knowing when human judgment is non-negotiable. If your business makes decisions based on AI output, those decisions need to be verified by a human who understands the domain. Period.
Security & Data Protection
What can go into an AI tool and what can't. This is non-negotiable. I establish clear policies about what data is allowed in which tools, how to use enterprise-grade AI services that keep your data private versus consumer tools that train on everything you type, and how to work with AI without accidentally sharing client data, financial information, or protected records. If you're in a regulated industry, this part is especially critical.
Ongoing Coaching
One training session isn't enough. AI tools change fast and people need reinforcement. I provide ongoing check-ins, answer questions as they come up, share new techniques and use cases, and help employees level up over time. The goal is self-sufficiency — your team should be discovering new ways to use AI on their own, not waiting for me to tell them.
04 // platforms_and_tools
AI Platforms & Tools
There is no single AI tool that's best at everything. Anyone who tells you otherwise is selling you something. The major platforms — Microsoft, Anthropic, OpenAI, Google — each have real strengths and real weaknesses. I work across all of them and I'll recommend whatever actually fits the job.
Sometimes the right answer is Microsoft Copilot because you're already in M365 and it works inside the tools your team uses every day. Sometimes the right answer is Claude because the task requires deep reasoning, long-form analysis, or working with complex documents. Sometimes it's OpenAI's Codex for code generation and developer tooling. Sometimes it's Google Gemini because it handles multimodal work or integrates with your Google Workspace environment. I don't have a favorite. I have experience with all of them, and I pick the one that makes sense for what you're trying to do.
Right Tool, Right Job
Each platform does different things. Microsoft Copilot is embedded directly in Outlook, Word, Excel, and Teams — your team uses it without leaving the apps they already work in. Anthropic's Claude handles long-form writing, document analysis, and complex multi-step reasoning — I use it for contract review, document drafting, research, and building agentic workflows. OpenAI's models power a massive ecosystem of integrations and third-party tools. Google Gemini processes text, images, video, and audio in a single pass and plugs into Google Workspace. I evaluate each use case on its own and match it to the right platform. Sometimes a single business uses two or three for different functions. That's normal. Better than forcing everything through one tool that's mediocre at half the tasks.
Agentic AI & Developer Platforms
Beyond chat interfaces, these companies offer developer-grade platforms for building real automation. Anthropic's Claude Code and Agent SDK let me build purpose-built agents that connect to your systems and execute multi-step workflows autonomously. OpenAI's Codex does the same for code-heavy automation. Microsoft's Copilot Studio lets you build custom copilots that work inside your M365 environment with your business data. Google's Vertex AI provides enterprise-grade model access with your own data governance. I work across all of these depending on what your workflow needs. The important thing is that the tool serves the process — not the other way around.
Microsoft Copilot
M365 integration · Copilot Studio
Embedded directly in Outlook, Word, Excel, Teams, and PowerPoint. Works with your existing M365 data — emails, documents, calendars, SharePoint. Copilot Studio extends it with custom copilots tied to your business processes. I handle readiness audits, tenant cleanup, licensing, and deployment. If your SharePoint permissions are a mess, Copilot will surface things people shouldn't see — I fix that first.
Anthropic Claude
Claude Code · Agent SDK · API
Long-form writing, document analysis, contract review, research synthesis, and complex multi-step reasoning. Claude Code and the Agent SDK are what I use to build custom agents and automated workflows — systems that connect to your tools and execute real work autonomously with human checkpoints.
OpenAI
GPT · Codex · API
Broad general-purpose models with the largest ecosystem of third-party integrations and plugins. Codex translates natural language into working code and powers developer automation workflows. The API connects to thousands of existing business tools through pre-built integrations.
Google Gemini
Gemini · Vertex AI · Workspace
Processes text, images, video, and audio together in a single pass. Integrates with Google Workspace for organizations on that platform. Vertex AI provides enterprise-grade model access with data governance and access controls for organizations with regulatory requirements.
This landscape changes fast. New models drop every few months. Capabilities shift. Pricing changes. I stay current so you don't have to. When something better comes along for your use case, I'll tell you — and I'll help you switch.
05 // automation_engineering
Custom Automation & AI Agents
This is where it gets interesting. Beyond chatbots and Copilot, there's a whole layer of AI that most small businesses don't know exists — agents, skills, and automated workflows that can take complex, multi-step processes and run them with minimal human input. This isn't science fiction. I build these right now.
What AI Agents Actually Are
An AI agent isn't a chatbot. It's a purpose-built system that can reason through problems, use tools, access your data, and execute multi-step tasks autonomously. Think of it as a specialist employee that works 24/7, doesn't make typos, and follows your procedures exactly. A chatbot answers questions. An agent does work.
Skills & Workflows
Agents are powered by skills — reusable modules that encode specific expertise and processes. A data analysis skill knows how to import, clean, validate, and visualize data. A document creation skill produces formatted Word docs, PDFs, and spreadsheets. A research skill can review literature, build evidence tables, and draft structured reports. I build custom skills that match your business processes, then wire them together into workflows that handle entire pipelines — intake to output — with human checkpoints where they matter.
Power Automate & M365 Integrations
Not everything needs a custom AI agent. A lot of business automation runs perfectly well on Power Automate — Microsoft's workflow engine that's included in your licensing. New form submission triggers an email, creates a SharePoint entry, and notifies a Teams channel. Invoice comes in, gets routed to the right approver based on amount and department. Employee submits PTO, calendar gets blocked, manager gets notified. I build these flows and maintain them.
Custom Agent Development
For more complex needs, I design and build custom AI agents tailored to your business. These connect to your systems — M365, your CRM, your project management tools, your databases — and perform real work. An agent that processes incoming support requests, categorizes them, pulls relevant customer history, and drafts a response. An agent that reviews contracts against your standard terms and flags deviations. An agent that takes raw data and produces a formatted report with analysis and recommendations. I build the agent, test it against your real workflows, and deploy it with appropriate guardrails.
Multiplying Your Workforce
The point of all of this isn't to replace people. It's to take the 10–15 hours a week your team spends on repetitive, structured tasks and give that time back. Your operations person spends less time formatting reports and more time on strategy. Your admin spends less time on data entry and more time on the work that actually requires human judgment. One employee with well-built AI tooling produces what used to take two or three. That's the value proposition — not fewer people, more output per person.
06 // governance
AI Policy & Governance
You need a policy before you need a tool. If your employees are using AI without guidelines, you already have a problem — you just don't know the shape of it yet.
Acceptable Use Policy
I draft your organization's AI acceptable use policy. Which tools are approved. What data can and can't be used. Which roles have access to which AI capabilities. How outputs should be reviewed and attributed. What happens when someone uses an unapproved tool or puts protected data into a consumer AI service. Clear, enforceable, and written in language your team can actually understand.
Data Classification for AI
Not all data is equal when it comes to AI. Public marketing copy? Feed it into any tool you want. Client financial records? That goes nowhere near a consumer AI service. I help you classify your data and map it to approved tools — what's safe for general AI, what requires enterprise-grade tools with data protection guarantees, and what stays out of AI entirely.
Compliance Considerations
If you're subject to HIPAA, if you handle PII, if you work with government contracts, if your industry has specific data handling requirements — AI adds a layer of complexity to your compliance posture. I don't replace your compliance officer or your attorney, but I make sure the technology side is set up correctly. Enterprise AI tools with the right data processing agreements, audit logging, and access controls.
Vendor Evaluation
Every AI vendor says they're secure and they don't train on your data. Some of them are telling the truth. I evaluate AI tools and vendors for data handling practices, enterprise agreements, SOC 2 compliance, and data residency. If a tool doesn't meet your requirements, I'll tell you — and I'll find one that does.
07 // engagement_model
How It Works
AI services are available to all managed services clients and as standalone engagements. Here's what the process looks like.
01
Discovery
I spend time with your team — leadership and the people doing the work. I learn your processes, your pain points, and your goals. No questionnaire. Actual conversations.
02
Assessment & Roadmap
You get a written report: here's where AI fits, here's where it doesn't, here's what it'll cost, and here's the order we should do things. No jargon.
03
Policy & Foundation
AI acceptable use policy drafted and approved. Enterprise tools set up with proper security and data handling. Governance in place before anyone touches a new tool.
04
Training
Role-specific training across your team. Effective prompting, ethical usage, security boundaries, and hands-on practice with the tools they'll use daily.
05
Build & Deploy
Copilot deployment, Power Automate flows, custom agents, integrations — built, tested against your real workflows, and deployed with guardrails.
06
Iterate & Expand
Ongoing coaching, new use cases as your team gets comfortable, expanding automation to more workflows. AI isn't a one-time project — it's a capability that grows with your business.
08 // boundaries
What I don't do
I'm upfront about scope. AI services from ATS are focused on practical business automation and workforce enablement. Here's what falls outside the line:
Custom model training or fine-tuning. I don't train LLMs from scratch or fine-tune foundation models. I use existing models through their APIs and enterprise platforms. If you need custom model development, you need a machine learning engineering firm.
AI-generated customer-facing content without review. I won't deploy AI that talks to your customers unsupervised. Chatbots that hallucinate to your clients are a liability. Any customer-facing AI output goes through human review.
Replacing human judgment in critical decisions. AI assists. Humans decide. Especially in legal, financial, medical, or personnel matters. I build tools that give your people better information faster — not tools that make decisions for them.
Promising magic. AI is a tool. A powerful one. But it's not going to fix broken processes, compensate for understaffing, or replace expertise you don't have. I'll tell you what it can realistically do for your business — and I'll tell you when the answer is "not much."
// lets_go
Ready to stop guessing about AI?
Let's figure out where it fits in your business. No pitch deck. Just a conversation.