DoLessWork · For professional services

The front desk for consultancies, agencies & B2B services

Every project enquiry in one queue with the reply already drafted from your rules. Contract, legal and compliance questions go to a person, never to a draft.

Send us your last three enquiries and we'll send back the replies we would have drafted. No signup, no card. If they earn it, engagements start with a $5,000 pilot — 7–14 days on your real enquiries — then a $15–25k full system if the pilot proves itself.

Sound familiar?

What actually lands in your inbox

You tell us your rules once. Every message that arrives gets read, scored and answered, then waits for you.

Your standing ruleNever quote a project price or timeline in writing from a description alone — offer a discovery call. Never state a certification, compliance status or security posture in writing; a person answers those with the evidence attached.

New message from Marcus D.example

We need our data platform rebuilt before the Q1 board meeting. Ballpark, what does something like that run? And are you SOC 2 certified — procurement will ask.

Score 88Qualified enquiry with a real deadline, asking for a price and a compliance attestation. Neither is answerable in writing.

AI draft reply

Hi Marcus,

A Q1 deadline is workable, and it is the right thing to design the scope around.

I would rather not put a number on a rebuild before I understand what is there now — a ballpark given blind is usually wrong in whichever direction is least useful to you. On SOC 2, I am not going to characterise our position in an email; someone here will come back to you with the current documentation, so procurement has the actual evidence rather than my summary of it.

Could we find half an hour this week to go through the platform as it stands, and what has to be true by the board meeting?

Kestrel & Co

No drafted reply is sent until you approve it. There is no auto-send for written replies, on any plan. The AI does the writing; you stay the one who decides.

Marcus D. is invented for this example. We launched recently and don't publish real customer messages. The draft is the genuine output format, refusals included. See a full sample workspace →

Standing rules

Never quote a project price from a description alone — offer the discovery call. Written once, applied to every draft.

Commitment routing

Contract, security and compliance questions are flagged and handed to a person. A draft never signs you up to anything. This is a rule, not a setting.

One approved queue

Enquiries scored and drafted in your voice. You approve, edit, or bin. No drafted reply is sent until you say so.

A draft never makes a commitment on your behalf

Contracts, NDAs, MSAs and data-processing agreements. Liability, indemnity and who owns the work. Security questionnaires, SOC 2, ISO 27001, GDPR. SLAs, guaranteed delivery dates, and anything asking for legal advice. All of it is flagged and routed to a person, every time. It is not a setting you can loosen and not a confidence threshold. If the system is unsure whether a message carries contractual weight, it treats it as though it does.

What your standing rules actually look like

This is not an illustration. These five are the starter rules a professional-services workspace is provisioned with on day one (lib/blueprints.ts) — written in your words during the pilot, applied to every draft, and yours to keep whether you continue or not.

When we are not the right fit

We do not take on law practices. Drafting client communications for a law firm runs into professional-conduct rules (ABA Model Rules 5.4(a) and 7.2(b)), client confidentiality, and unauthorized-practice exposure that we are not set up to carry — and we would rather tell you now than sign you and work it out later. Consultancies, agencies, studios, accountants and other B2B services firms are who this is built for.

Four things this system cannot do

Most of what the AI does is judged work: it drafts, your rules guide it, and a person approves it. These four are different. They are enforced by code, and no busy day and no AI mistake can change them.

We will not tell you the AI never makes a mistake. We built the system so a written mistake cannot reach your customer on its own.

Replies are drafted by AI and reviewed by a person at your firm before anything is sent. We will tell your clients that when they ask, and you can require it be said up front.

Common questions

Can it answer a contract or security question?

No, and that is enforced rather than configured. Contracts, NDAs and data-processing agreements, liability and who owns the work, security questionnaires, SOC 2, ISO 27001 and GDPR, SLAs and guaranteed delivery dates are all flagged and routed to a person, every time. It is not a setting you can loosen and not a confidence threshold. If the system is unsure whether a message carries contractual weight, it treats it as though it does.

Do you work with law firms?

We do not take on law practices. Drafting client communications for a law firm runs into professional-conduct rules (ABA Model Rules 5.4(a) and 7.2(b)), client confidentiality, and unauthorized-practice exposure that we are not set up to carry — and we would rather tell you now than sign you and work it out later. Consultancies, agencies, studios, accountants and other B2B services firms are who this is built for.

Will it quote a project price to a prospect?

Only if your rules tell it to. The starter rules do the opposite: never quote a project price or timeline in writing from a description alone, offer a discovery call instead, and give a published starting range only if your rules include one. Your standing rules are written in your words during the pilot, applied to every draft, and you can change them any time.

What does it cost?

Engagements start with a $5,000 pilot: 7–14 days on your real enquiries — workspace provisioned, standing rules written with you from your actual messages, three working workflows, and nothing ever sent without your approval. If the pilot proves itself, the full system is $15–25k, scoped in writing from what the pilot showed. Billed by invoice — no card on file. Ongoing operation afterwards is optional. Run plans start at $300 a month. That is what carries anything recurring, such as voice answering.

We are a Philippine business — what does it cost?

We are based in the Philippines and bill in pesos, with a BIR-registered invoice for your books. A ₱25,000 pilot proves it on your real enquiries in 7–14 days. If it earns the ongoing partnership, the full system is ₱75,000–150,000, scoped in writing. An optional run plan starts at ₱6,000 a month if you want us operating it after it is yours. Voice answering is not included in these prices. If you want it, we measure what your calls actually cost during the pilot and pass that through at cost as a single monthly amount — no markup, and still never billed per minute. The figure is agreed at the end of the pilot, when there is real usage to size it from. The full peso price list is at dolesswork.com/VA/philippines.

Are there per-minute or usage charges?

No. We do not bill per minute for calls or for anything else, and you are never charged for hang-ups, dead air, or spam. The price is fixed in writing before we start.

Is there a limit on how many calls it answers?

Where AI voice answering is switched on, an active run plan includes up to 1,000 answered minutes a month, roughly 45 minutes of calls every business day. Voice answering is an optional add-on: it is not included in the pilot or in the full-system build price, and it is switched on only when you ask for it. Run plans start at $300 a month. Your usage is visible in your workspace at all times. If you go past it we talk to you about more capacity first: no surprise charge, and we do not silently switch your calls off.

Do I have to buy human hours?

No. Human Partner time is billed by the hour with no minimum and no retainer, on top of the software price — the rate is quoted with your scope. You can use it for the work that should not be automated, or never use it at all.

Does anything get sent without me seeing it?

No. Every reply arrives in your queue as a draft. You approve, edit, or bin it. No drafted reply is sent until you approve it.

Will my clients know they are talking to AI?

Replies are drafted by AI and reviewed by a person at your firm before anything is sent. We will tell your clients that when they ask, and you can require it be said up front.

Do you need my passwords?

Never. We do not ask for passwords or logins. System access happens through your own accounts. Our privacy policy sets out who can read your messages, how long we keep them, and how to export or delete everything in one click.

Can I talk to an existing customer?

Not yet. We launched recently and have no customers to point you at and no case studies. What we can do instead is take your last three real enquiries and send back the replies we would have drafted, before you pay anything.

What happens if we stop working together?

Engagements are fixed-price and billed by invoice, and each one’s refund terms are set out in its own written scope — there is no rolling contract behind them to give notice on. Whichever way you go, you keep the standing-rules document we build with you, and you can export everything we hold in one click.

The front desk is where an engagement starts, not where it ends. See the 5 departments the same engine runs →

Curious what your week is costing you? Calculate how many hours you'd get back →

See it running for a business like yours.

The sample workspace shows two weeks of realistic professional-services leads, calls, orders, and invoices: exactly what yours looks like once it knows your business.

Open the sample workspace →

Not ready to sign up?

Paste in a message a customer actually sent you. We'll send back the reply we'd have drafted. Yours to keep either way, no account needed.

A person reads it. Drafted replies come back within one business day. No signup, no card, and nothing is ever sent to your customers.

← DoLessWork home