TL;DR Most "I left my job to consult" stories stop at the exciting part. This one starts in month four, when you have three clients, a sick day, and no one else in the building. AI lets a single hospitality professional run what used to be a 5-person firm. It also creates new failure modes nobody warns you about.
Every story about leaving a W-2 to go independent ends at the airport. The hero boards the plane, posts a LinkedIn update, and the story cuts to black right before the hard part.
The hard part is a Wednesday in month four. You have three active clients. One is mid-rebrand. One just emailed at 7am about renegotiating scope. The third needs a training deck by Friday. You also have a wedding this weekend and a fever that has been sitting at 99 since Tuesday.
That is the story nobody tells. That is the one I am living.
I run a hospitality consulting firm for bars, restaurants, hotel outlets, and food and beverage operators who need outside help with operations, training, and brand positioning. I do this work without a full-time team. What I have instead is AI, a short list of human contractors I call for specialized work, and systems built to carry most of the operational weight of a small firm.
I am not selling you on AI. I am telling you what it looks like from inside the actual work, on a Wednesday when three things need to happen and there is only one person in the building who can make them happen.
The honest version of this is a ledger.
What AI handles now: Cold email drafting. Proposal writing. Meeting notes and follow-up summaries. Social posts. Contract first drafts. Transcription of every call. Weekly reporting templates. FAQ replies. Podcast show notes. Blog post drafts. KPI dashboard templates.
That is a full-time administrative job. It runs in the background while I am doing the actual work.
What AI does not replace: Discovery calls. The actual strategy work. In-room training with a real crew. Hard conversations about what is and is not working. The relationship maintenance that keeps clients renewing. Signing contracts. The moments where empathy is the deliverable.
I want to be precise about that second list. These are not the things AI will eventually replace. These are the things where the human IS the work. A client paying me to assess their service program is not paying for a document. They are paying for someone who stood behind a bar for ten years and knows what a Tuesday night crew looks like when they are lost. What a bar manager sounds like when they are covering for a no-show and pretending everything is fine. What a service breakdown actually costs in long-term guest loyalty. That is not a prompt.
The replaced list freed up 20 or more hours a week. Those hours go into the second list.
First: Drafting things you already know how to do but hate doing. Proposals. Follow-ups. Scope documents. You know what these need to say. AI handles the first draft, you cut what is wrong and add what only you know, and you send it. That cycle takes 20 minutes instead of 90.
Second: Pattern-matching from a lot of inputs. Feed it fourteen client conversations from a single month and ask what themes keep surfacing. What used to take an afternoon of manual review now takes 20 minutes. I use this monthly to sharpen my service offering.
Third: Producing structured output from unstructured input. A voice memo at 11pm becomes Monday's first draft of a deck. The thinking is mine. AI organizes what I already thought.
First: Real-time judgment. A client just asked to renegotiate scope mid-call. They are tense. You do not know yet if this is financial stress, a leadership change, or a genuine miscommunication about what they thought they were buying. AI cannot read the temperature of that room. It cannot tell you whether to hold the line or flex. That call takes operator experience and two ears paying close attention. You have to be there.
Second: Anything where being wrong destroys the relationship. Apologies. Fee adjustments. Partner conversations about a deliverable that did not land. Any communication that could change a relationship gets written by me, reviewed by me, sent by me.
Third: Sourcing facts you cannot verify. AI will confidently produce a statistic, a citation, or a name that does not exist. It will phrase it with the same confidence it uses for everything else. A consultant who passes that output directly to a client is one conversation away from being fired. Every number, every attribution, every external reference that lands in a client deliverable gets verified against a real source before it ships. No exceptions.
This last one is not a minor caveat. It is the category most likely to end your practice.
I do not name vendors. Here is what each tool does.
A general-purpose chat AI for drafting, pattern recognition, summarization, and first-pass analysis. A calendar with booking links so I am not losing hours to scheduling email chains. A CRM that tracks every prospect, client, and active conversation. A documentation system holding every SOP, brand voice file, and client-facing template. Transcription running on every call, always. Recurring scheduled jobs for weekly reporting, monthly invoicing, and content distribution.
Plus a short list of human contractors: a video editor, a graphic designer for big projects, an accountant.
Total monthly software cost is under $400. Total monthly contractor cost is under $1,500 most months.
One hour every morning before any AI tool gets opened. Coffee, walk, gym if the schedule allows. A paper notebook with three things that actually matter today.
AI is a force multiplier on whatever you point it at. Point it at the wrong thing and you produce excellent output for a project that should have been killed. The notebook is the rudder. The AI is the engine.
The lonely lunches. No team, no debrief after a difficult client call. You close the laptop and it is quiet.
The vacation that is not really a vacation. When you stop working, the firm stops working.
The single point of failure when you get the flu in week three of an engagement. No backup. No handoff.
The taxes you did not know you owed in April. Self-employment tax, quarterly estimates, the stuff your W-2 handled invisibly.
The constant, low-grade anxiety about the pipeline. Not acute panic most of the time. A background hum. Is there enough work next month? Is there enough work in three months? The hum does not fully stop.
This is not a complaint. It is the actual price tag. Anyone selling this lifestyle without these line items is selling you something.
Every client gets a real human on the phone within 24 hours of their first email. AI does not pick up the phone.
Every deliverable gets my eyes on it before it leaves. AI drafts. I ship. My name is on the work.
Every Friday I look at next week's pipeline. If the pipeline is empty, Monday is a sales day. Not an execution day. Not a content day. The instinct is to keep executing on current work because it feels productive. The math says current work ends and the next client needs to already be in motion.
You hate writing prompts and refuse to learn. Tools take practice. If that sounds like a chore, this is not the right path.
You need the structure of a team to do your best work. Solo consulting is lonely and the loneliness compounds. If you know this about yourself, honor it.
You think "AI will do everything" and you bring no expertise. Clients are not paying for the AI. They are paying for your judgment about their specific operation.
You are three months from rent being a problem. Solo consulting takes runway. If you cannot fund six months of slow ramp, take the W-2 a little longer and make the jump when you have something to land on.
How much do you actually make doing this?
I am not giving you a number. The right comparison is: what would you make at a 5-person consulting firm, minus the overhead of four other people, the office, the HR burden? Solo with AI reframes the margins. The ceiling is different when the cost structure is different.
Won't AI eventually replace consultants entirely?
No. The pattern is the same as every prior tool. Spreadsheets did not replace accountants. They changed what accountants do and raised the floor on what those accountants needed to know. The human who can hold a room, read a client, and take responsibility for the outcome is not replaceable by a model. The consultant providing commodity output is a different story.
What's the first AI tool a hospitality consultant should buy?
A general-purpose chat assistant and a transcription service. That is it for month one. The stack grows from actual needs, not from a list of tools you think you should have.
Do clients know you use AI?
Yes. They do not ask about the tools. They ask about the outcome. If the outcome is good, the tools are invisible. If the outcome is bad, no toolkit saves you.
Could a bartender or bar manager do this?
Yes. Most of the operators I work with came up behind the bar. Operator-to-consultant is the most natural transition in the industry. The missing piece is usually confidence that what you know is worth paying for, and a system for finding the people willing to pay for it.
Step 1: Write down your last 30 hours of work. Mark every line item as either "I had to do this myself" or "an assistant could have drafted this and I would have edited it." That second list is your AI runway. Most people find it is longer than they expected.
Step 2: This week, pick the smallest item from the assistant column and run it through one AI tool. One. See how the output compares to what you would have written. Iterate on the prompt. Run it again. The goal is not to replace yourself. It is to find out where the tool earns its cost and where it does not.
Step 3: If you want to see what running a hospitality consulting practice on AI actually looks like up close, the consulting work and KMS platform are at jlittrell.com and kmsops.com.
More playbooks like this go out in The Ops Wire newsletter at theopswire.substack.com
Jason Littrell spent 10 years behind the bar in NYC (including Death & Co) and served as USBG NYC president. He now runs his hospitality consulting firm entirely on AI. He hosts the Hospitality Strategy Lab podcast and writes The Ops Wire newsletter.
Jason