AI advisory
AI is changing your costs and your risks, whether you use it or not.
A formal advisory practice: a written policy your staff will actually follow, an honest look at what is already being used in your business, training on one tool done properly, and security built for a threat landscape that AI has already changed.
The one thing worth understanding
The tools that help a business write better emails, clone a voice and automate outreach are the same tools criminals are now using against your employees. Phishing is grammatically perfect and personalised. Voice cloning can impersonate your CEO on a call. None of it takes much technical skill any more, because AI lowered that barrier for everybody at once.
That is the part most people have absorbed. The part they have not is the cost side, which arrives regardless. A business that never opens an AI tool is still paying for the energy, hardware and infrastructure being built to run everyone else's. Not engaging with AI is not opting out. It means carrying the risk and the cost without the benefit.
"We are not selling fear and we are not manufacturing urgency. We are sharing what we are living ourselves, and what we see across every client environment we manage."
Eric Grimes, PartnerYou are almost certainly using it already
AI is switched on inside Teams, Gmail, Zoom, Adobe and Salesforce, usually by default, before anyone in the business made a decision about it. The same is true of industry software: practice management, accounting, insurance and service platforms have been adding AI features to products you already pay for. Most businesses we walk into are paying for AI they have never thought about, and their staff are using more of it than leadership knows, on personal phones and personal accounts.
That is the actual risk, and it is worth saying plainly: the danger is not AI. It is unmanaged AI.
What to do, in this order
The instinct is to solve all of it at once. That produces sprawl, confused staff and nobody accountable. This sequence is the one we walk every client through, and most of the value is in the first two steps.
Why the policy comes first
It is one page and it starts from default deny: the tools leadership has approved are permitted for company and client information, and everything else is not. That sounds restrictive until you notice it is also permission. It tells the people already using AI to do good work that they may carry on, on accounts the business controls.
We have a template and we will fit it to your business. Have your attorney review it; we are not your attorney, and we will tell you where that line is. Then get it acknowledged in writing, because a policy nobody signed is a policy you cannot enforce.
"Draven was able to restore the functionality of the Excel macro and resolve the issue with the Project Tracker file. He communicated clearly throughout the process and explained the remaining issue. Although one template issue was not fully resolved, he provided recommendations for next steps and offered additional assistance. I appreciate his help and responsiveness."
Vanya, engineering firmMeet Draven Gold Star survey"Matt Wilson is the best! He is so knowledgeable, kind, and the best assistance when running into any issues. He's always willing to go above and beyond to ensure that our office needs are met. Today I had an issue he hadn't run into before, and without hesitation he figured out how to address the problem, and in less than five minutes had a solution. So grateful to have him on our team!"
Lillian, insurance brokerageMeet MattYour costs are going up, and it is not us
Four separate pressures are converging on every small business right now, and not one of them requires you to use AI.
The security floor moved
AI-assisted attacks are more targeted and more frequent than what businesses defended against two years ago. Endpoint detection, advanced email filtering and real staff training are the baseline now, not upgrades.
AI infrastructure is raising prices
Hardware, cloud services and software subscriptions are all absorbing the energy and infrastructure demands of AI data centres. It is already showing up in renewal pricing across the industry.
AI subscriptions will cost more
Today's enterprise tools are priced to win adoption. As those platforms mature and move to value pricing, the cost rises for early adopters and latecomers alike.
Compliance now has real overhead
State AI privacy laws, federal requirements and licensing board rules create documentation and audit obligations that cost time, even for a business doing everything right.
We are not raising costs because of AI. The world is, and we would rather explain that honestly than let it arrive on a renewal with no explanation. The job is to help you navigate it efficiently, not to profit from it.
Every answer says where it came from
Clients on the program get a login to our advisory library. The centrepiece compares the major AI platforms on the questions that actually decide whether you can use one: where your data physically goes, whether your prompts train the model, how long they are retained, what the vendor's compliance posture really is rather than what its marketing page implies, and what it costs at your size.
Every answer carries a note saying where we got it and how sure we are. A vendor's own trust centre is not the same kind of fact as a term in a signed enterprise agreement, and we do not present them as though it were. That is the difference between advice you can hand to your attorney and advice you have to take on faith.
What to watch for
We are a small business too
Everything above applies to us, not just to our clients. We are navigating the same threat landscape, the same rising costs and the same governance questions we are asking you to address, and we apply the same policies to ourselves first. When we recommend a control it is because we already run it here. That is the only basis on which this kind of advice is worth anything.
What it is actually good for
Once the policy is in place and people know the rules, the honest first-year value is unglamorous and real. Summarising a long contract before it goes to your attorney. Turning rough meeting notes into decisions and owners. A first draft of a policy or a job posting. Proofreading a client letter before it goes out. Comparing three vendor quotes side by side. Cleaning up messy notes into a usable table.
None of that is a transformation and anybody promising you one should be asked to put the number in the proposal. All of it assumes a person reads the output before it goes anywhere, because AI is confidently wrong in ways that are hard to catch. It drafts; your people stay accountable for what leaves the building.
"Which AI tools are people in your organisation using right now, officially or unofficially?"
The question we open withWhen the answer is software that does not exist yet
Sometimes the right answer is not a subscription. It is the spreadsheet that outgrew itself three years ago, the form three people retype into three systems, the report somebody rebuilds by hand every Monday, the job done on paper in a field and re-entered at a desk.
We build those. Affordable custom automation and software, using AI where it earns its place and ordinary engineering where it does not, scoped and quoted in writing before anyone starts. It is not an AI pitch with a software project attached; it is software, built by people who already know your network.
Why your IT company is the right one to ask
You need this for the same reason you need managed IT: the expertise is real, it moves quickly, and it is not cost effective to build it inside a business your size. An AI consultancy hands you a deck. We are already inside the systems the questions are about. We run your Microsoft 365 tenant, so when the answer is to enable a tool for twelve people with certain data excluded, we are the ones who do it. We hold your backups, so we know what your data actually is. We answer your insurance questionnaire every year, so we know which of these answers your carrier will want in writing.
Every engineer and account manager here completed formal AI training this year, assigned as a ticket with a due date, before we sold a single hour of advisory.
What we will tell you not to do
Do not buy an AI product because it was bundled into a renewal. Do not put client records into a free tool on a personal account, whatever the box says about privacy. Do not deploy three products at once. Do not build something custom before a two-week trial of a boring existing tool. And if your backups have never been tested or your staff are not on multi-factor authentication, do those first. AI is not the thing standing between you and a bad year.
If a vendor tells you your setup is fine, ask them to put it in writing. The answer changes once it has to be written.
The ERP under a desk Small businessSunk cost is real. A backup server you bought last year deserves a plan before it goes in a dumpster because a move made it convenient.
Server or cloud, honestlyWhat is included
Questions we get asked
Which AI tools are our employees using right now?
Almost certainly more than leadership thinks, and that is the question we open with. It is not fully auditable, but between a staff survey, what is enabled by default in your Microsoft 365 or Google tenant, and what your industry software has quietly added, we can establish a real starting point rather than an assumed clean one.
The full answerWhy are our technology costs going up if we are not using AI?
Because the world you operate in changed. The security baseline rose permanently, AI data centre demand is pushing up hardware, cloud and software pricing across the industry, and compliance now carries real documentation overhead. None of those require you to use AI. We would rather explain that than let it turn up on a renewal unexplained.
The full answerWhat is the first thing we should actually do?
A written acceptable use policy, on a default-deny basis, with signed staff acknowledgments. It is one page, it costs almost nothing, it is enforceable, and it solves the largest real risk most offices have. Everything else is easier once it exists.
Will our data be used to train the AI?
On a consumer account, often yes, and you frequently cannot tell. On a business or enterprise agreement, usually no, and the vendor commits to it in the contract rather than a help article. That difference is the single most important thing to get right.
The full answerIs AI actually worth it for a business our size?
For most thirty to two-hundred person businesses the first year of value is drafting, summarising and finding things: real, but unglamorous. Sometimes it is not worth it yet, and we will say so. If someone has promised you a transformation, ask them to put the number in the proposal.
The full answer
Not sure what your people are already using?
That is the first question. A partner will help you answer it.
Talk to a partnerDraven was able to restore the functionality of the Excel macro and resolve the issue with the Project Tracker file. He communicated clearly throughout the process and explained the remaining issue. Although one template issue was not fully resolved, he provided recommendations for next steps and offered additional assistance. I appreciate his help and responsiveness.Vanya, engineering firmFrom a ticket survey, 2026
Talk to a partner, not a sales script.
Tell us what is going on. A partner will call you back the same business day, look at what you have, and tell you honestly whether we are the right fit. No pricing pressure, no pitch deck.