Client Portal + AI Concierge / Micro Portal Builder
a service business that emails the same status update and then answers the same five questions by reply, over and over.
Fill in one config, the client’s name, your brand color, the project steps, the documents you still need, and a short FAQ, and it builds a single branded page with a progress bar, a checklist, and an “ask the concierge” box that answers questions by matching them against your own FAQ.
$39 a month. It buys back the hours you spend re-answering “where are we” and “what do you still need from me.” Cancel any day and the portals you built keep opening.
The five questions you answer every week
A client hires you and then wants to know, on a loop, where things stand and what you’re still waiting on from them. Fair questions, all of them, and answering each one by email is a small tax you pay per client per week. Multiply that across your active jobs and it’s a real slice of the week spent typing the same replies.
A proper client portal solves it, but the portal products assume you want logins, permissions, notifications, and a monthly per-client price. For a small shop with a handful of active jobs that’s a heavy, expensive answer to “show the client a status page.” So you keep emailing.
Validate, render, hand over
- It validates your config, a JSON object with the client name, a hex brand color, the steps, the documents, the FAQ, and your contact info, failing fast with a clear message if something’s malformed before it writes anything.
- It renders the portal as a single page: a header in your color, a progress section computed from which steps are done, a document checklist, and the concierge box.
- It wires the concierge inline, a search box backed by a keyword-overlap matcher embedded right in the page, so a client’s question gets answered from your FAQ with no server and no API key.
What the portal does
- Opens as one self-contained page you can email as an attachment, drop on a USB drive, or put on any static host.
- Shows the client exactly where the project stands, so “where are we” answers itself.
- Lists the documents you’re still waiting on, so the client can see their own homework without you chasing.
- Answers common questions on the spot through the concierge, which matches their wording against your FAQ and shows the best fit, never inventing an answer you didn’t write.
- Carries your brand color and contact details, so it reads as yours, not as generic software.
Who it suits, and who it doesn’t
Small service firms with a handful of active clients who want a clean status page per job without standing up a login system or paying per client.
Not for you if you need clients to log in, upload files, and see status that updates itself live. This is a snapshot you regenerate when things change, not a two-way system of record. If real accounts and file uploads are the point, a full portal platform is what you want.
What it replaces, and what those do better
Status emails and repeated Q-and-A, a client-portal platform like SuiteDash or a HoneyBook portal, and a shared Google Drive folder. The platforms are genuinely better where they charge for it: real logins, file uploads, live updates, and notifications. A shared drive is free and lets the client drop files back.
What none of them do is hand you a branded, self-serve status page from one config, offline, that you can send as a file. And the concierge here is honest about being a keyword matcher over your FAQ, not a chatbot that might make something up. That’s the trade: no live accounts, but nothing to log into and nothing that hallucinates.
Where your client data stays
It runs on your machine. The config is read from your own file and never uploaded, and the only writes are the portal page and its data echo, under your output folder, at 0600 permissions. Your brand color is validated as real hex, and every value that lands in the page is escaped so nothing can be injected.
Because the concierge runs entirely inside the page with no network call, the portal you send a client isn’t quietly reporting their questions anywhere. Cancel and every portal you made is still a file you own.
$39 a month, and the first portal
Write the config for one client, set the output folder, and run it. The portal is built in seconds and opens in a browser. Update the steps or documents and regenerate to send a fresh status.
Security and privacy
Handles client-facing status and document information. The generated portal is opened by the client themselves, so it’s treated as untrusted-facing output. Security is designed in.
Principles
– Local-first. PORTAL_CONFIG is read from .env and never transmitted anywhere; nothing in this tool touches the network. Nothing is uploaded, no telemetry, no cloud dependency. No live LLM call — the “AI concierge” is a deterministic, dependency-free keyword-overlap FAQ matcher.
– No client-side network or storage. The rendered index.html makes no network requests and uses no localStorage/cookies — it’s a static file that works offline, from a USB drive, or attached to an email.
– Least privilege. Reads PORTAL_CONFIG from environment, writes only under OUTPUT_FOLDER/Portal/.
Protections implemented (v0.1.0)
– Path-traversal defense — every output filename (and the Portal/ subfolder itself) is sanitized (safeSegment) and every write is checked with assertWithinRoot(), so output can’t land outside OUTPUT_FOLDER.
– Fail-fast config validation — clear errors for missing PORTAL_CONFIG/OUTPUT_FOLDER, invalid JSON, a non-object payload, or a missing/invalid clientName/steps/documents/faq shape, all before any file is touched.
– Restrictive permissions — the generated index.html and portal.json are written 0600.
– HTML-escaping — every value that flows into index.html (client name, status, step names, document names, contact) is HTML-escaped, so a value like <script> in the config renders inert rather than executing.
– Strict brand color validation — brandColor is only accepted if it matches a strict 3/6-digit hex pattern (sanitizeBrandColor); anything else (including javascript:-style values) silently falls back to a safe default color rather than being inlined into CSS/HTML.
– Safe script embedding — the FAQ data embedded in the concierge widget’s inline <script> is JSON-serialized and then escaped (</> → </>, plus U+2028/U+2029 line separators) so a crafted FAQ answer can’t break out of the <script> block or exploit JS string line-terminator quirks.
– Defensive parsing — parsePortalConfig never throws; it returns a list of validation errors instead, and drops malformed array entries (e.g. a steps entry missing name) rather than crashing the run.
– Secrets discipline — .env and generated Portal/ output are git-ignored; no secrets or client data in the repo.
Distribution / AV trust
Known limitations / roadmap
– The concierge matcher is intentionally simple (keyword-overlap, no synonyms/stemming); it’s designed to be predictable and auditable rather than “smart” — richer matching is on the roadmap but will stay local/deterministic.
– brandColor only supports hex values in v0.1.0; named CSS colors (e.g. "navy") are rejected in favor of the default rather than validated against a color-name allowlist.
– No authentication on the generated index.html — it’s a static file, so anyone with the link/file can view it. Access control (e.g. per-client unique links, expiry) is on the roadmap.
Reporting
Report suspected vulnerabilities privately to the maintainer before public disclosure.
Frequently asked questions about Client Portal + AI Concierge / Micro Portal Builder
What is the client portal + AI concierge, and how does it work?
The client portal + AI concierge from SortShop is a white-label portal you can brand as your own, with an AI concierge that answers a client’s status questions straight from your project data. Clients log in, ask what is happening with their job, and get an answer pulled from the information you already track, instead of waiting for you to write another update email.
Setup is straightforward. You brand the portal, connect your client and project data, and let clients in. When someone asks the common question, which is usually some version of what is my status, the concierge responds directly from your data. You are not staffing a live chat or rewriting the same reply ten times a week. The portal is the front door and the concierge is the part that actually saves you time.
The design goal is narrow on purpose. The portal piece is familiar and crowded, so the real value sits in the concierge that answers from your own records rather than generic canned text. Priced at thirty-five dollars a month, the client portal + AI concierge is aimed below full portal suites, at owners who mostly need to cut the volume of status questions coming at them.
Because a portal like this holds names, contact details, and project specifics, how it handles that information matters. The FTC guide the FTC guide Protecting Personal Information tells businesses to take stock of the personal information they collect and to lock it down. A client portal is exactly the kind of place where those steps apply, since it centralizes details clients trust you to protect.
The tool keeps that responsibility in view rather than hiding it. You control what data the concierge can see and answer from, which means you are scaling down what is exposed to the least you need for good answers. You can read how the setup works on the Client Portal + AI Concierge page and see who it suits among small-business owners and operators.
The honest pitch is modest and useful. A branded portal on its own is not new, but a concierge that answers status questions from your real data cuts a genuine chunk of repetitive work. That is the specific job the client portal + AI concierge is built to do, and it is why the concierge, not the portal shell, is the part worth paying for.
Think about the week it is meant to save. Without a portal, a client emails to ask where things stand, you stop what you are doing, look it up, and write a reply, then repeat that for the next client and the next. With the concierge answering from your project data, most of those questions never reach your inbox at all. The client gets a straight answer at the moment they wonder, and you keep working. That is the entire promise, and it is a narrow one on purpose. The portal shell is ordinary, and the real value is in taking a repetitive task off your plate. For an owner who is also the salesperson and the account manager, reclaiming that scattered time across a busy week is worth more than another dashboard nobody checks.
How does a client portal + AI concierge help me handle client personal information responsibly?
Any portal that lets clients check their status has to hold personal information to do it. Names, email addresses, phone numbers, and the details of each project all sit behind the login. That makes responsible handling part of the job, not an optional extra, and the client portal + AI concierge is built so you stay in control of what the concierge can reach.
The clearest guide here is the FTC’s, which small operators can actually follow. Its guide Protecting Personal Information lays out five steps: take stock, scale down, lock it, pitch it, and plan ahead. A portal supports the first two directly. You take stock because client data is organized in one place you can see, and you scale down by connecting only the project information the concierge needs to answer status questions, rather than dumping everything into it.
Locking it down is the login itself. Clients reach their own information through the portal, and the concierge answers from the data you allowed, not from a free-for-all of everything you have ever stored. That boundary is the point. A concierge is only as safe as the data you let it touch, so the tool keeps that scope in your hands.
Planning ahead, the last FTC step, is about knowing what you would do if something went wrong. Because you decide what the portal holds and what the concierge can see, you already have a clear inventory of the personal information in play. That inventory is the starting point for any sensible response plan, and it is far easier to keep current when the data lives in one branded portal instead of scattered across email threads and spreadsheets.
The client portal + AI concierge does not turn security into someone else’s problem, and it should not pretend to. You still choose strong access rules and decide who on your side can see what. What the tool gives you is a structured place for client data and a concierge whose reach you define, which is a better starting point than answering status questions out of a shared inbox.
You can see how access and data scope are handled on the product page and read about the broader fit for small-business operators. Used the way the FTC guide describes, a portal like this makes the careful handling of client information the normal way you work, rather than a thing you have to bolt on later.
It is worth being concrete about scaling down. You do not have to connect your entire business to the concierge for it to be useful. If clients mainly ask about project status, then project status is what you connect, and financial details or internal notes can stay out of scope entirely. That restraint is not a limitation, it is the FTC’s own advice put into practice, since holding less exposed information means less to protect and less to lose. The portal gives you a dial for exactly how much the concierge can see, and turning that dial down to what the job actually requires is the responsible default. Most owners find the concierge answers the common questions perfectly well on a fairly narrow slice of data.
Is my client data safe, and where does it live in the client portal + AI concierge?
Client data in the portal lives behind the login you control, and the concierge answers only from the project information you connect to it. That scoping is the heart of the safety story. The concierge cannot reveal what it was never given access to, so you decide the boundary of what any client, or the assistant answering them, can see.
Thinking about safety is easier with a framework, and the NIST Cybersecurity Framework gives a plain one. It organizes the work around identifying what you hold, protecting it, detecting problems, and being able to recover. A client portal + AI concierge fits that model because it centralizes client data in a known place. You can identify exactly what is in the portal, protect it with access rules, and keep a clear picture of who can reach what.
Centralizing data this way is usually safer than the alternative most small operators live with, which is client details spread across email, texts, and a pile of spreadsheets. Scattered data is hard to guard because you can never quite say where all of it is. A single branded portal gives you one place to reason about, which is the identify step the NIST framework starts with.
The concierge does not change that boundary. It answers questions from the data you connected, so its reach is exactly as wide as you made it and no wider. If a category of information should not be exposed to clients or to the assistant, you do not connect it. That keeps the protect step firmly in your control rather than in a black box.
None of this removes your own responsibilities. You choose who on your team has access, you set sensible passwords, and you decide what belongs in the portal at all. The client portal + AI concierge gives you the structure to make those choices deliberately, which is more than an inbox and a spreadsheet ever will. Safety comes from that structure plus your own good habits, not from the tool alone.
You can review how data connection and access work on the Client Portal + AI Concierge page and see the wider context for small-business operators. The straightforward answer to the safety question is that your client data stays where you place it, the concierge sees only what you allow, and you keep a clear, single view of it all, which is exactly what the NIST identify-and-protect steps ask for.
Consider the contrast with how a lot of small businesses run today. Client details end up copied into a booking email, a quote in one app, an invoice in another, and a note in a text thread. Each copy is a place you now have to worry about, and most owners could not list them all if asked. Pulling the client-facing view into one portal shrinks that to a single system you can actually describe and secure. The concierge does not widen the boundary, because it only answers from the slice you connected. That combination of one known place plus a defined reach is what turns a vague worry about scattered data into a manageable thing you can point to and protect.
How is this different from Copilot, SuiteDash, or Moxo?
The incumbents are broad client-portal suites. Copilot, SuiteDash, and Moxo each try to be a full workspace, with many modules and pricing to match. The client portal + AI concierge from SortShop is deliberately narrower. It is a white-label portal plus one thing the others do not center, an AI concierge that answers status questions straight from your data, priced at thirty-five dollars a month.
That focus is the point of difference. Plenty of tools can show a client a project dashboard. The value here is cutting the repetitive status-update work by letting the concierge field the common what-is-happening question, so you are not writing the same email over and over. If your real problem is the volume of status questions, a lighter tool built around that is a better fit than a suite you half use.
Price and scope follow from that. Full suites often charge for capabilities a small operator never touches, and the client portal + AI concierge stays below them by not trying to be everything. You brand the portal, connect your data, and get the concierge, without paying for modules you would leave switched off.
There is also a data-handling angle worth naming. Any of these tools holds client personal information, and the FTC guide the FTC guide Protecting Personal Information tells businesses to take stock of what they collect and to scale it down to what they actually need. A narrower portal makes scaling down easier, because you are connecting only the project data the concierge needs to answer questions, rather than filling a sprawling suite with client details across many features.
The honest limitation is the flip side of the focus. If you genuinely need a broad workspace with billing, file management, and project tools all in one, a full suite may suit you better. The client portal + AI concierge is not trying to win that comparison. It is trying to solve the status-question problem well and cheaply for owners who do not want the rest.
You can weigh the fit on the product page and see how it lands for small-business operators. For small operators whose main frustration is fielding the same client status questions all week, the client portal + AI concierge offers the concierge that actually answers them, without the weight or the bill of a full portal suite. That trade, focus over breadth, is the whole reason it sits where it does.
It helps to be honest about who each option serves. The broad suites are genuinely good if you want to run projects, files, billing, and messaging all inside one branded workspace, and some businesses do. The gap they leave is the owner who does not want all of that and is simply drowning in status questions. For that person, paying for a full workspace to solve one problem is a poor trade. The tool answers the narrow question well and stops there, which keeps both the price and the learning curve low. Choosing between them really comes down to whether your problem is running an entire client operation or just answering the same question all week, and this is built squarely for the second case.
Can the AI concierge really answer client questions from my own data, and what does that need from me?
Yes, answering from your own data is the core capability, and it is what separates this from a portal that just shows a static page. The concierge draws on the client and project information you connect, so when a client asks about their status, the answer reflects your actual records rather than a generic template. The client portal + AI concierge is built around that specific job.
What it needs from you is straightforward. You connect the project data you want the concierge to answer from, and you decide the scope of that connection. The quality of the answers tracks the quality of what you feed it, so keeping your project records current is what makes the concierge genuinely useful. In practice the tool is only as good as the data behind it, which is true of any assistant like this.
That dependence on your records is also why record-keeping habits matter around a portal like this. Client and project information often overlaps with the business records you are expected to retain anyway. The IRS guidance on how long to keep records explains that businesses should hold supporting records for set periods, generally at least three years and longer in some cases. Keeping that information organized helps your books and gives the concierge better material to answer from.
The concierge does not invent facts about a job. It answers from what you connected, which means its reach is exactly as wide as you set it. If a piece of information is not in the connected data, the concierge will not surface it, and that is by design. You control the boundary, so clients get accurate status answers without the assistant wandering into records you did not intend to expose.
Setting expectations honestly, this is a status-answering tool, not a replacement for real conversations when a job gets complicated. The concierge handles the routine questions well, which frees you to spend your own time on the ones that actually need judgment. That division of labor is the practical benefit for a busy operator.
You can see exactly how data connection works on the Client Portal + AI Concierge page and read the broader context for small-business owners and operators. The bottom line is that the client portal + AI concierge answers from your data, respects the scope you set, and rewards good record-keeping with better answers, which is a fair and useful trade for cutting down the flood of status questions.
A short word on the limits keeps expectations fair. The concierge is excellent at the routine, repeated questions and is not meant to negotiate a contract or handle a delicate complaint. Those still deserve you. What it removes is the low-value churn of status checks that interrupt real work without needing your judgment. Feed it current, accurate project data and it earns its keep every day. Let your records go stale and its answers will drift with them, which is simply the nature of any assistant that reflects what you give it. Treat the connected data as something worth keeping current, and the concierge stays genuinely useful rather than becoming another thing clients learn to distrust over time.
| Attributes | Value |
|---|---|
| Deployment |
Cloud |
| Platform |
Web |
| License Type |
Subscription |
| Billing Period |
Monthly |
| Target Customer |
Small Business & Retail |
| Free Trial |
Yes |
Frequently bought together
-
AI Grant / Proposal Writer (local RAG)
$39.00 -
AI QA / Citation Layer for AI outputs
$39.00 -
AI Receptionist That Acts
$95.00 -
AI Support Inbox for Microteams
$39.00 -
Batch Image / Brand Asset Generator
$65.00 -
Batch Video/Photo Culler for Creators
$119.00 -
Bulk Document AI Processor / Batch PDF Splitter+Binder
$249.00 -
Construction Takeoff Lite / Drawing Annotator
$65.00







