Frequently asked questions

Straight answers before you spend time or money.

If your question is not here, email me directly at jennifer@jennirayconsulting.com.

Straight talk

What I will and won’t do.

I use AI to cut repetitive work. Not to replace judgment, relationships, or responsibility.

I will

  • Start with one specific business problem
  • Measure the result on your own work
  • Set information-handling rules before implementation
  • Keep human review where it matters
  • Build from your approved language
  • Teach your team to operate the finished workflow
  • Tell you if AI isn’t the right fix

I will not

  • Promise that AI will run your business
  • Sell you five tools when one will do
  • Automate a broken process
  • Send client-facing work without review
  • Leave you with a pile of prompts and call it a system
  • Lock you into something your team cannot understand
What happens during the assessment?

I look at how the work actually moves through your business: where a job starts, who touches it, what gets retyped, where it sits waiting. Then I write up where your time is going, what AI could handle now, what isn’t ready, and what would have to change first.

It ends when there’s a real answer, not at a fixed time. Sometimes that’s a couple of hours. Sometimes it’s most of a day.

What does it cost?

The assessment is a fixed fee, quoted on our call before anything is scheduled.

If the assessment finds a workflow worth building, that gets scoped and priced separately, in writing, before any work starts. And if the scope changes along the way, we agree the change before the work does.

We already pay for Copilot. Is this a waste?

Probably not. Most unused AI licenses aren’t unused because the tool is bad. They’re unused because nobody connected them to a specific job. Part of what the assessment does is work out whether what you already have is enough.

Do we have to switch tools?

No. I start with what you already use and already pay for. If something new is genuinely needed, I’ll say so, but that’s rarely the first move.

What if we’re early on?

Then I’ll tell you what to put in place first, and that groundwork is work I can do with you. AI works when your information is organized enough to use, and most small businesses aren’t there yet. That’s normal, it’s fixable, and it comes first.

Knowing it going in costs a lot less than finding out after you’ve spent money building something. Some of that groundwork you can start on your own: the free guide walks through it.

Can you just teach us instead?

Yes. Some of what I do is done for you, and some of it is done with you.

If you’d rather build it yourselves, I work alongside your team on your real files: setting up the context AI needs and showing you how to keep adding to it. You end up with the thing built and the ability to build the next one.

I also teach the basics to teams that have never used AI properly. What it’s good at, what it isn’t, what should never go into it, and how to check what comes back.

We work out which of those makes sense during the assessment. Either way it’s scoped and priced in writing before anything starts.

Why do you work mainly in Claude?

It’s a practical starting point I know deeply, and depth in one tool beats shallow familiarity with a dozen. But the right tool depends on the workflow, your existing systems, your information-handling rules, and your needs. I won’t recommend adding tools without a clear reason.

Do you build chatbots or phone systems?

It’s not where I start. Most of the time the hours are going into the paperwork behind your business, not the phone that answers it. Internal operational workflows (intake, proposals, follow-up, documentation, reporting) are where the time usually is.

If you get to a point where a chatbot or a phone system is genuinely the right answer for a specific problem, we can talk about it then.

Where do you work onsite, and do you work virtually?

In person across the Salt Lake Valley, and virtually anywhere the workflow and information can be handled safely online. If you’re outside the valley, ask and we’ll sort out travel before anything is scheduled.

In person is usually better for the assessment, because seeing how work moves tells me things a call doesn’t.

What applications can you connect?

Microsoft 365, Google Workspace, email, files, calendars, forms, accounting software, CRMs, and industry-specific applications are all possible. A connection is recommended only when the application supports it, access is authorized, the risk is understood, and it actually improves the workflow.

What do we actually get at the end?

A workflow your team can run without me. The work isn’t finished when a prompt looks good. It’s finished when the process is tested, documented, taught, and handed to someone who owns it.

The written playbook covers:

  • Purpose and approved use
  • Required inputs and where they come from
  • Step-by-step operating instructions
  • The review checklist and escalation points
  • Information that must not be used
  • How to spot and report a bad result
  • Who maintains the workflow
How much do you actually automate?

Every workflow starts with a person reviewing the output. It only moves up when it’s working and the risk is understood.

  1. 1AssistAI drafts and organizes. A person reviews everything.
  2. 2StandardizeYour approved examples and templates make the output consistent.
  3. 3SystematizeIt becomes a written process someone else on your team can run.
  4. 4IntegrateIt connects to email, files, or your CRM when that’s safe and worth doing.
  5. 5AutomateOnly proven, low-risk steps run on their own. Client-facing work keeps human review.

No skipping levels. Automation isn’t the goal. A workflow that saves time and protects trust is the goal.

Do we own the finished work?

Yes. You own the assessment, the documentation, the templates, the instructions, and the workflow materials created for your engagement, subject to the service agreement and any third-party software terms. If you never work with me again, you keep all of it.

Are there continuing software costs?

Possibly. Your existing subscriptions may already support the first version. If a paid AI plan, connector, or application is useful, I’ll identify the recurring cost before it becomes part of the design.

Is client data safe?

Safety depends on the tool, account configuration, data, contracts, and professional obligations. We set information-handling rules before implementation, limit access, keep human review, and exclude sensitive data when the available setup isn’t appropriate. You remain responsible for applicable legal, regulatory, ethical, and contractual requirements.

What information should never be entered into AI?

There’s no universal list that fits every business and tool. Passwords, secret keys, payment-card data, medical information, government identifiers, privileged material, and other regulated or contract-restricted information need especially careful handling and often shouldn’t be entered at all. We create rules specific to your business and your exact tool configuration.

Will AI replace employees?

The goal is to cut repetitive administrative work, not to remove judgment, relationships, or responsibility. The best first workflows give your team a better starting point while keeping them accountable for the result.

How is success measured?

We choose a practical measure before implementation: minutes per item, weekly backlog, rework, turnaround time, or consistency against an agreed checklist. Then we compare the current process with the tested workflow on your own work.

Testing covers normal examples, edge cases, and deliberately bad inputs. Outputs get checked against the agreed quality list before anyone depends on them.

What should we prepare beforehand?

Nothing, really. Bring someone who understands how the work actually gets done and real examples: actual proposals, actual intake notes, actual reports. Not a tidied-up version. The awkward handoffs and exceptions are the useful part.

If we get as far as building something, I’ll also need your business rules and exceptions, any approvals that have to happen, access to tools only where it’s needed and authorized, and someone available to review during testing. None of that is needed for the assessment itself.

Please don’t send sensitive files through the website form.

What if AI isn’t the right answer?

I’ll say so. You’ll still leave with a clearer map of how the work moves and a written recommendation for a better next step.

Have a different question?

Tell me about the work, the team, and what is getting stuck. Please do not include confidential client information.

Request a fit call