HireAlfie runs a managed system for small and mid-market businesses. Alfie is that system, built and run for you. Not a person. He takes the repetitive work off your week — review replies, Google posts, inbox drafts, the weekly numbers — inside the software you already use, with AI doing the drafting. Alfie prepares, you decide.
Jobs we can build to order. Four have run in production; the rest start from your audit.
Four workflows ran in production for a three-location restaurant group from July to September 2026. That deployment is paused as of September 2026. The review-and-post workflow left a one-line log of every run it completed, and those logs — with the client's details removed — are what we walk you through.
Google review replies were drafted overnight in the owner's voice and checked against the facts; Yelp replies were drafted from Yelp's notification emails.
A post per location, written with that location's real hours, address and offer. Posts the engine handled were checked and corrected before publishing, and the original text was logged.
Customer and vendor emails were answered in draft and saved in the inbox. Nothing was sent by the system; a person pressed send.
Weekly server-payout summaries built from the sheets and boards the managers already filled in: the week's totals computed, archived and snapshotted, and only then the board cleared for the next week (two locations, June–September 2026).
Designed, not yet run for a client. Each starts from your Operations Audit, and each would run in draft-and-approve mode first.
We would build: welcome notes, check-ins, a review request when a job closes, and the call nobody remembers to make — prepared on the right day and routed to the right location.
We would build: blog posts, newsletters and social updates drafted from what actually happened in your business, waiting for your approval.
We would build: website and email inquiries acknowledged, logged in your CRM or sheet and queued for your follow-up — including the ones that arrive at 9pm on a Friday.
We would build: one short note a week — what went out, what was held, what needs you — instead of a dashboard nobody opens.
Four industries, each with the workflows we would set up first. One has run in production; the others are built to order from the same engines, and each panel says so. The panels change on their own. Hover or tap to hold one.
Restaurants and hospitality groups
Every location in a restaurant group creates the same small jobs each week: review replies, Google posts, reservation emails and the servers' payout totals. Each reply and post has to carry that location's facts, not a sister location's patio line or last season's offer.
This is the one industry where the system has run in production: a three-location restaurant group, July to September 2026. That deployment is paused as of September 2026.
Allergen and dietary wording in any reply, post or email draft would come only from a phrase list you approve. No draft would call a dish gluten-free or nut-free unless you have documented it.
For the three-location restaurant group, each post the engine handled was checked against the verified menu and banned-word list before publishing. That step existed because the posting tool's own AI broke the rules even with them written into its prompt. For your group, the verified menu and the banned words would come from you. Posts outside your written approval would wait for you.
Not if it is clean. We would first check whether it sends only happy guests to the public review link. If it asks every guest, you keep it. If it sends only happy guests there, we would show you how to ask everyone before anything else is built. The rule we work by: Ask everyone. Respond to everything. Never condition the ask on the score.
No. Your menu, word lists and the instructions written for your group are never reused for another restaurant. Only the working method carries over. Your records stay in your own systems; Alfie reads only what a task needs.
Law firms
Each office's Google profile needs its own posts and its own reviews. A web inquiry sent on a Saturday can sit unread until Monday, and nobody chases the ones that go quiet.
We have not yet run this for a law firm. These workflows are built to order from the engines proven on a restaurant group.
Advice on a client's matter, the final word on any deadline, conflict decisions and anything that moves trust-account money are never automated. Attorney review of any legal content is non-negotiable. Alfie prepares, you decide.
Yes. We would suggest review replies first, because that is where a single reply can expose a client. An attorney at your firm would approve every reply before it is posted. We would write the confidentiality rules with you. The rules are built around common bar-advertising cautions, and your firm confirms which ones apply to you.
Yes, as an internal early warning, with a person calling anyone who is unhappy. The Google review request would go to every client, whatever their score. Ask everyone. Respond to everything. Never condition the ask on the score.
No. Review replies and office posts would work from your Google Business Profiles. Inquiry drafts would need only the delegated intake inbox your inquiries arrive in. Review requests would need each client's name, email, office and the date the matter closed. That could come from a shared sheet or read access to your practice-management tool. Your records stay in your own systems; Alfie reads only what a task needs.
Dental and medical practices
Small public jobs slip at a busy practice, like a website nobody has touched in a while. Review replies and Google posts sit on the same list, and none of it needs a patient chart.
We have not yet run this for a dental or medical practice. These workflows are built to order from the engines proven on a restaurant group.
Any draft that touches a treatment or a clinical question would wait for a person at the practice. No automation would be the only check on anything medical.
No. This work would stay on public reviews, your Google profile and your website, and would not touch patient records. Your records stay in your own systems; Alfie reads only what a task needs. Appointment reminders would need patient details, so we would not build them until the data-protection question has a clear answer.
Yes. We would hand the fix pack to your developer to apply, or use editor access to your CMS if you prefer. You or your developer would approve every change before it reaches the site. It would never invent a rating or an award.
What any patient can already see. We would check your website for an old COVID notice, placeholder text, a stale copyright year and a blog that went quiet. The Operations Audit would then choose the first workflow with you.
Home services and trades
What tends to slip is the work after each visit: the quote nobody chased and the finished job with no review ask. Many trades websites also have a blog whose last post is more than a year old.
We have not yet run this for a home-services company. These workflows are built to order from the engines proven on a restaurant group.
We would build the review ask for every finished job, with no survey deciding who gets the link and nothing offered in exchange. A customer who replies unhappy would wait for a person.
Alfie prepares, you decide. Your part is approving drafts one at a time, or signing off a standing rule in writing for the routine ones. Anything sensitive, like an upset customer or a question about money, waits for a person. The time depends on how many jobs and quotes you handle a week. We work that out in the Operations Audit rather than guess here.
It depends on the app, and we check that in the Operations Audit before anything is built. The review ask and quote follow-up would need to see when a job is marked complete and which quotes are still open. If your app cannot show that, a shared sheet your office updates would do. Your records stay in your own systems; Alfie reads only what a task needs.
We would build that as drafts for your approval: posts about the jobs you do in the towns you serve. Nothing would go on your site until you approved it. We do not promise rankings or a number of new jobs.
Alfie was not designed on a whiteboard. He was built to run one company's repetitive work, and we will walk you through how it ran — without naming them.
From July to September 2026, a three-location restaurant group in the United States ran its review replies, per-location Google posting and inbox drafts on this system, plus a weekly payout close at two of its locations. That deployment is paused as of September 2026. We keep client identities confidential as a rule — the same discretion you would get from us.
Replies to three- to five-star reviews were drafted overnight in the group's own voice, checked against its menu and facts, and published under the owner's standing approval. One- and two-star reviews were never answered automatically: the engine drafted a reply that owned the problem and left it unpublished until the owner decided.
Each location got its own post with its own hours, address and offer. The posting tool's own AI kept producing things that weren't true — invented claims, details from the wrong location, a reservation offer in a season with no reservations. Posts the engine handled were corrected before publishing, and the original text was logged.
Weekly server-payout summaries were built from the restaurant's operational boards: the week's totals computed, archived and snapshotted, and only then the board cleared for the next week — with a safe stop if anything failed before that last step. Two locations, June to September 2026.
Ask for a walkthrough of the logs, with the client's details redacted. We would rather show the record than describe it.
No big setup fee, no long contract. Four steps, in order.
You tell us what keeps landing on your desk. We tell you honestly whether it is a fit. Free, and the notes are yours to keep either way.
A fixed-fee, one-week look at one area of your business: the current process mapped, the repetitive work ranked by what it costs you, the safest first workflow to automate, and a fixed-scope plan with a go-live date and your exact monthly fee. The fee is $750, quoted in writing before you commit to anything, and it comes off your first monthly invoice if you go ahead. The plan lists every tool and AI service that would touch your data.
We build the first workflow around the systems you already use and run it in draft-and-approve mode for two weeks. You watch it work on real tasks before any monthly fee starts. If it does not do what the plan says, you stop there.
One fee per workload — one recurring job, such as review replies — billed monthly from go-live, with no setup fee, a three-month minimum and 30 days' notice after that. Your exact monthly fee is in the audit plan, in writing, before you commit. Monitoring, fixes when a platform changes, and a plain-English record of what went out and what waits for you are included. Everything is handed over documented if we part ways. See how pricing works.
HireAlfie is run by Rigers Qoraj (LinkedIn), who builds and runs Alfie from Tirana, Albania. He built the system for a three-location restaurant group first, then turned it into a service. There are no account managers in between: the person who maps your workflows, builds them and watches the logs is the person who answers your email, within one business day.
Everything Alfie runs is documented as it is built. Nothing sensitive goes out on its own, and if he is ever unreachable the whole setup can be handed to anyone you choose.
The things owners ask on the first call.
HireAlfie runs a managed system for small and mid-market businesses. Alfie is that system, software we build and run for you: AI drafts the work, a person decides. We take one piece of repetitive work off your week — review replies, Google posts, inbox drafts, the weekly payout close — build it around the software you already use, run it, watch it, and fix it when a platform changes.
No. You describe the work in plain words. We map it, build the workflow around your existing systems, and run it. You approve drafts where they already live — your inbox, your Google profile, a shared sheet — and you get a plain-English record of what went out and what is waiting for you.
Never on his own. Low-star reviews, complaints, refunds, and anything touching money or legal judgment are held for a person. Alfie drafts; a human decides whether it goes out. That is the rule the whole system is built around.
No. Replies never confirm that a person is a customer, patient or client. For practices and firms that rule is written into the setup before anything is drafted.
The Operations Audit takes about a week and ends with a fixed-scope plan and a go-live date. The first workflow then runs in draft-and-approve mode for two weeks before the monthly fee starts. We give you the date in the plan, not a promise on this page.
One person: Rigers Qoraj, in Tirana, Albania. He built the system for a three-location restaurant group and then turned it into a service. The person who builds your workflows is the person who answers your email.
The system was proven on one business: a three-location restaurant group. For law firms, dental practices, home-service companies and e-commerce brands we have mapped the workflows but not yet run them for a client in those industries, and we say so up front. If your business is not a fit, we will say so on the first call.
Practical notes on the work nobody owns.
Google gives businesses no way to remove a review just because it is negative. But many businesses unknowingly run review gating — asking only happy customers — which is against Google's review policy, which the FTC has said may be unfair or deceptive, and which quietly caps their review growth.
How AI search engines and answer engines evaluate local businesses, and what operational steps ensure your company is recommended.
Why generic automated replies damage local business reputation, and how a human-in-the-loop review workflow builds authentic trust.
Tell us which task eats your week. We'll show you exactly how Alfie would carry it.
Tell us the task