AI Agents and Assistants
What is an AI agent? A plain-English guide for business owners
What an AI agent is, how it works, where it helps a small business, the risks, and a 7-point check to see if you are ready for one.
Business Agents
AI Compliance Assistant
Policy questions answered instantly and work checked against your compliance rules, sourced from your own documents.
Policy questions slow work down. Staff either interrupt the one person who knows the rules or guess and hope. This assistant answers internal policy questions instantly and checks submitted work against your compliance rules, using only the documents you approve. It flags gaps for a person to review rather than blocking anything, so your compliance team stays in charge.
- Instant policy answers sourced only from your approved documents
- Submitted work checked against your rules before it goes out
- Gaps flagged early for a person to review
$3,400
20–30 working days
Optional maintenance $249/mo
What it is
AI Compliance Assistant is a chatbot for your own business rules, not for AI governance. It answers internal policy questions and checks submitted work against your compliance rules. It uses only the documents and policies you approve, so nothing it says goes beyond what you have written. Gaps are flagged for a person to review.
Why businesses buy it
Compliance problems are cheapest to fix before work leaves the building. Yet policy knowledge usually sits with a few busy people, so questions queue up and checks happen late or not at all. This assistant gives everyone the same accurate answer at any hour, quotes your own documents, and catches gaps early. Your compliance team spends less time repeating itself and more time on genuine judgment calls. The channel, policy material and topic areas are written on the Scope Card before you order, and answers are tested before launch.
Problems this solves
Staff interrupt compliance leads with the same policy questions
The chatbot answers internal policy questions using only the documents and policies you approve. It runs in a website widget, Slack or Microsoft Teams, where staff already ask questions.
Submitted work is checked against policy inconsistently
It checks submitted work against your compliance rules and flags gaps for a person to review. The final decision always stays with your own reviewer.
People worry an AI tool will invent policy answers
Answers never go beyond what you have put in writing. If the approved material does not cover a question, it is handed to a person instead of being guessed.
A good fit if
Staff ask the same policy questions repeatedly, and pointing them to the right document each time takes real hours.
You have compliance rules or policies already written down, even if they are spread across several documents.
You want work checked against those rules before it goes out, catching gaps early rather than after the fact.
One channel, such as a website widget, Slack or Microsoft Teams, is where your team already looks for policy answers.
Not the right fit if
You need role-based training content delivered and tracked, not policy Q&A and checks. See AI Training Agent.
You need account-level monitoring or reporting against business metrics, not policy compliance specifically. See AI Business Analyst Agent.
You need it to take corrective action automatically, such as blocking a submission. This build flags issues for a person; automated blocking is a custom workflow, quoted separately.
You need more than 100 documents or 200 pages of policy material, or more than five topic areas. Add a note at checkout for a Change Quote, or use the Custom AI enquiry form for larger compliance programs.
You want staff questions answered about how your company may use AI and data. See AI Policy Assistant, built for AI and data policies.
You need your AI workflows monitored against internal compliance rules. See AI Compliance Monitoring, part of our AI Governance & Compliance range.
How it works
- Your policy documents are loaded, up to 100 documents or 200 web pages, organized against the up to five topic areas you define, such as data handling or expense rules.
- A person asks a policy question in the channel you set up, or submits work to be checked, in plain language rather than searching through documents.
- The answer or check comes only from your approved material. If a policy does not clearly cover the case, the agent says so rather than guessing at an answer.
- Work is checked against the rules you set, flagging anything that looks out of line with your policies so it can be corrected before it goes further.
- Anything outside scope, or judgment calls the material does not resolve, is handed off to the email or helpdesk you connected.
- Results are checked against 40 documented test questions before launch, so you can see accuracy before it goes live to your team.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- The policy documents or web pages the assistant should learn from, organized by topic where possible.
- Your top 20 policy questions with the answers you want given, as a starting point for the wider material.
- Access to the email address or helpdesk that should receive questions or cases it cannot resolve.
- Access to your website, or your developer's help, to install the widget if that is your chosen channel.
- Your logo and brand colors if the widget should match your site.
- Admin access and API keys for every connected system, shared by adding our team as a user or creating a key you control, never by email.
- Your business information (company basics, hours and brand voice), shared during intake.
- One decision-maker who can approve the policy material and topic areas during the build.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
Intake verified, topic areas agreed.
Weeks 1–2
Policy material loaded and checking logic built.
Week 3
Channel and handoff system connected.
Weeks 3–4
40 test questions run, results sent to you.
Week 4
Your review and two rounds of changes.
Go-live
Live, recorded handoff session, support starts.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $3,400 |
| Express | 15–25 working days | $4,000 |
| Rush | 10–20 working days | $4,600 |
What you receive at handover
- The AI Compliance Assistant, live in one channel: a website widget, Slack or Microsoft Teams
- Your approved policies and documents loaded, within the agreed limit
- Up to five intents, including work checks that flag gaps for human review, plus a handoff route
- A test report with documented test questions and pass/fail results
- An admin guide and recorded handoff session
- 30 days of post-launch support
Expected benefits and how to measure them
- Can free compliance leads from routine policy questions
- Helps gaps in submitted work get caught before sign-off
- Makes policy answers consistent with your written rules
- Policy questions reaching the compliance team
- Tally questions sent to compliance by email or chat for a month before launch and a month after.
- Gaps found before sign-off
- Count submissions flagged by the assistant and confirmed by a reviewer. Compare with issues found later in past reviews.
- Test question pass rate after policy updates
- Rerun the documented test questions whenever a policy changes, and compare results with the launch report.
Before you start: Record how many policy questions reach compliance each month and how long answers take. Note recent issues that were caught late.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
The build price is one-time. Running it costs roughly $30 to $150 a month in your own accounts, covering AI model usage and hosting. Cost mostly tracks how many questions or checks are run each month.
AI usage during build and testing is billed the same way, in accounts you control.
Ask about running costsCustom requirements
Add a note at checkout if you need something specific. If it fits within the Scope Card it is simply included. If it goes beyond it, you get a written Change Quote within two working days, and that work starts only once you accept it, with its time added to your delivery date.
Request a custom versionKeep it running
Monthly maintenance covers monitoring, fixes, model and platform updates, a monthly health report and enough monthly hours for updated policies, new FAQs or an added topic area as rules change. It starts after the 30 days of included support, bills monthly and can be cancelled any month.
$249/mo (the minimum fee, includes 2 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A policy management platform charges a per-seat monthly fee whether staff use it or not | A compliance consultant bills hourly to answer questions and review work | Ours is one fixed price, shown before you buy |
| Answers | A policy platform's search returns whole documents to read through | A consultant's answers depend on who is available | Ours answers the specific question, sourced only from your approved policies |
| Checking | A policy platform does not check submitted work against the rules | Manual review speed depends on the consultant's availability | Ours checks submitted work against your rules and flags gaps early |
| Delivery | A policy platform is live once your team finishes populating it | A compliance consulting engagement usually runs several weeks | Ours has a fixed delivery time shown before you buy, with faster options |
Questions about this product
Can it approve or block a submission on its own?
No. It flags anything that looks out of line with your policies clearly, so a person can review it and make the final call. Automatically blocking a submission outright is a custom workflow, available separately if that is genuinely what you need.
What does it typically cost to run per month?
Roughly $30 to $150 a month in your own accounts, covering AI model usage and hosting. Cost mostly tracks how many questions or checks are actually run each month, and there is no separate fee to us at all for keeping it running.
What if a policy question is not clearly answered in our documents?
It says so plainly rather than guessing at an answer, and hands off to the email or helpdesk you connected instead. That kind of gap is also genuinely useful to know about, since it often means the underlying policy itself needs clarifying.
Where can staff go to ask it policy questions?
A website widget, Slack or Microsoft Teams, whichever channel your team already checks for policy answers day to day. The build covers one channel; a second is available as an add-on and quoted separately once the first one is proven and working well.
How much time will onboarding actually take my team?
Roughly one to two hours for intake, mostly gathering your top 20 questions and existing policy documents, spread across the 72 hours after you order. Then about an hour reviewing the 40 test questions together before the assistant goes live to your whole team.
Can I update the policies later?
Yes. Updated material fits inside the monthly change hours in maintenance at no extra charge, since it is a configuration change rather than new work. A genuinely new topic area beyond the five in scope comes through a written Change Quote first.
Is this a substitute for legal or compliance advice?
No. It answers only from the policies and rules you provide and authorize it to state, and it never creates legal or regulatory advice of its own accord. Your own compliance function stays fully responsible for writing and maintaining the underlying policies themselves.
Is this for AI governance or for our business rules?
It is for your business rules. It answers staff questions on your own policies, such as expenses, client procedures or data handling. It also checks submitted work against those rules. For rules about how your company uses AI, see AI Policy Assistant for staff questions. AI Compliance Monitoring checks your AI workflows themselves.
Guides on this topic
All guides Stop letting policy questions slow work down.
Order the AI Compliance Assistant and give your team instant, sourced policy answers and early checks on their work. The policies and topics in scope are agreed in writing before you pay.
Works well with
Business Agents
AI Project Manager Agent for Linear
Tracks tasks, deadlines and blockers and posts status updates to the team, connected natively to Linear so it reads and updates issues,…
Business Agents
AI Project Manager Agent for Monday.com
Tracks tasks, deadlines and blockers and posts status updates to the team, connected natively to Monday.com so it reads and updates boards,…
Business Agents
AI Project Manager Agent for Asana
Tracks tasks, deadlines and blockers and posts status updates to the team, connected natively to Asana so it reads and updates tasks,…