What kind of work suits AI, and where should we start?

AI suits work that repeats often, crosses two or more systems and follows rules you can write down. The starting point is the AI Audit. It maps your teams, processes and apps, sizes each opportunity by impact and effort, and shows what to do first.

Do our teams have to learn new software?

Your teams keep the apps they use today, and AI agents work inside them. The company assistant rolls out team by team, with training. AI apps are designed with the people who will use them.

How much of our team's time does it take?

Time with the people who run the work, and a view of the apps they use. In each short cycle, they review a demo on your own data and say what to change. The build itself sits with us.

Where does our data go? Is it used to train AI models?

Your data stays in your systems and your accounts, on your servers or your cloud. The AI models run through your company's own business accounts, with providers you approve. Business accounts with the major providers keep your data out of model training by default, as a term of the contract. Each person sees only what their role allows.

What if our data is messy?

Most companies keep their data across apps, spreadsheets and email, and the AI Audit starts there. It maps where the data for each process sits. The AI works from your systems as they are, and where a process needs cleaner data, that work is part of the plan.

Which AI models do you use, and can we choose?

Comet Lab works with GPT, Claude, Gemini, Llama, Mistral and others, independent of any vendor. Your company approves the providers, and each task gets the model that suits it. The company assistant offers a choice of model for each job, on one bill.

What happens when the AI gets something wrong?

People approve the decisions. Every question, source, action and approval is recorded, so a mistake can be traced to the step that caused it. At first the AI prepares the work for your team to check, and takes on more once it is proven on real cases.

What does it cost to run once it's live?

Three things decide the running cost: model usage, hosting on your servers or your cloud, and the scope of what is built. In the company assistant, usage limits are set by person and team. The AI Audit sizes the running cost before anything is built.

Who owns what you build, and can our team take it over?

You do: the code, the data and the settings. The code and settings live in your repository, and everything runs on your servers or your cloud. Every part is documented at handover, so your team can run it, extend it or move it.

AI consulting and implementation partner

AI in every part of
how your company works.

Built around your systems, your rules and your people.

Book an intro

AI working inside your company.

AI agent

People

AI agents

Run your routine processes, with people on the decisions.
Explore AI agents
Team spaces

Sales
Finance
Operations

Skills

Automations

Running
Sales

CRMAccountsPolicy
Company assistant

One AI workspace for every team, connected to your systems.
Explore the company assistant
Send Sent
AI apps

Software built around your work, with AI inside.
Explore AI apps

Start with the AI Audit.

Find out where AI pays off in your company, and leave with a roadmap to make it happen.

Intensive audit Teams, processes, stack and apps.
Audit report Where AI pays off, and what to do first.
AI roadmap A clear plan for AI implementation.

Book an intro See how the audit works

What should AI do first at your company?

AI models keep improving.
They can handle more complex work.
But routine work does not need the newest model.
AI handles it today.

The operating benefits of company AI

More operating capacity without adding headcount
Faster turnaround on everyday work
More consistent quality across operations

AI use cases

AI at work in every department.

Everyday tasks, in the apps your teams already use.

Explore the library

Your teams stay in their apps. The AI comes to them.

It works inside the CRM, the ERP, the accounts app, email, chat and documents, beside the people who use them.

It stays in your control.

Runs on your servers or your cloud.
Each person sees only what their role allows.
Every action is recorded.
You own the code, data and settings.

Built on your company's context.

Every agent, assistant and app works from your company's facts, rules and live data. Built with your team, and kept current as the business changes.

Your team

Directs the work, makes the judgement calls, does the high-value work.

Company context

Durable facts

The slow-changing truth. Changes rarely, shapes every piece of work.

Company facts
SOPs
Quality specs
Compliance rules
Org chart
Customer history
Brand guidelines
Pricing rules
Escalation matrix
Approval chain
Past lessons
Templates

Live context

The state of the business today. Changes daily, captured as work happens.

Current priorities and projects
Production schedule
Service tickets
Quarterly plans
Payments due
Machine downtime
Customer escalations
Lead pipeline
Raw material prices
Stock in hand
Dispatches this week
Pending approvals

ReadsDraftsCross-checksRoutesUpdates records

Under your team's supervision.

Your systems

CRMERPAccountsEmail and messagesSheetsDocuments

Free resources

Learn how AI works, and a map of tools to build it.

Free to read, and kept current as AI changes.

How AI works, where it fails, what it costs and where it fits.

Read Horizon

Atlas

Where each AI tool fits, and what it is good for.

Open Atlas

Before you choose an AI partner.

ChatGPT and Copilot help each person with drafts, summaries and ideas. AI built for your company is grounded in its facts, apps and history, and completes work across your systems. What it learns stays with the company.

Put AI to work in your company.

Talk to us about your teams, your processes and your apps. We'll show you where AI fits first.

Book an intro

AI in every part ofhow your company works.

AI working inside your company.

AI agents

Company assistant

AI apps

Start with the AI Audit.

What should AI do first at your company?

The operating benefits of company AI

AI at work in every department.

Account managers get any customer's deals, dues, tickets and emails in one answer

Made-to-order quotes are built right first time, priced to your rules and signed online

Account managers get a one-page brief before every customer meeting

Sales heads ask why the forecast moved and get the reasons, deal by deal

Most RFP questions are answered from your past bids, and experts fill the gaps

Discount requests reach the right approver, with the margin impact shown

Quotes for repeat customers go out the same day, approved by the owner in Slack

Every campaign is tied to the pipeline and revenue it brings, in one monthly report

Ask which campaign worked and get one answer from your ads, website and CRM

Ad spend on Google, LinkedIn and Meta is checked against the plan every morning

News from every site and team becomes a LinkedIn post, ready for approval

Case study drafts come from project notes and the customer interview, ready for review

Every brochure and post is checked against brand rules and approved claims before it goes out

Competitor price changes, launches and hiring arrive in one email every Monday

Solved tickets become help articles, so each question gets answered once

Every unhappy customer gets a call back within 48 hours, from someone who knows their history

Website chat answers from manuals, orders and warranty records, and hard cases reach an agent

Every support email is read, tagged and sent to the right team within minutes

Thousands of survey comments and reviews become a monthly list of themes, with examples

Angry customers and repeat complaints reach a senior person the same day

Agents get a drafted reply to every email, with the order details filled in

Customer POs become sales orders in SAP, checked against price and stock

Technicians get answers at the machine, from the manual, the readings and past repairs

Stock is counted on a phone, and every gap reaches the books and the auditor

Supplier quotes are compared line by line, with the best value marked

Engineers ask for a drawing and get the released revision, with what changed and why

Open issues at shift change become work orders for the next shift

Evidence for every audit point is ready before the auditor asks

Any budget variance explained in minutes, down to the records behind it

A 13-week cash forecast that updates itself every Monday

Every bank account is reconciled each morning, with only the exceptions left for the team

Customer payments are matched to the right invoices the day they land

Overdue invoices get the right reminder, and risky accounts reach a person first

Once finance defines a number, every answer across the company uses the same definition

Month-end accruals are ready on day 1, drafted from open POs and goods receipts

Every application is screened against the role, with a reason for each shortlist

Staff get policy answers on Teams, with the clause each answer comes from

New joiners get their first-month questions answered on Slack, with the source

Interviews are booked across busy panels the day the shortlist is approved

Every joiner gets a first-week plan, a buddy and trainings booked before day one

Managers get a one-page summary of each person's year before the review

Every goal has a measure and a target before the manager approves it

Every leaver's access ends on the last day, with a record for audit

Password resets and software access are handled on Teams in minutes

Every joiner's accounts and laptop are ready before day one

Suspicious emails reported by staff are checked and answered within minutes

Unused software licences are found and reclaimed every month

During an outage, staff get a clear update every hour until it is fixed

Every IT ticket lands with the right team, with the details they need

Incoming contracts are checked against your standard positions, with each difference marked

Salespeople get contract answers in a minute, and new questions reach a lawyer

Every contract's renewal dates and obligations are tracked, with reminders in time to act

Every policy is checked when a rule changes, with each clause to update marked

Lawyers ask in plain words and get a contract's current terms, checked against its invoices

Every new circular is summarised and sent to the team that must act on it

Annual policy acknowledgements are chased until everyone has signed

Leaders ask how yesterday went and get plants, sales and service in one answer

The board pack comes together from every department's papers, with late ones chased

Every meeting's decisions are on record, and anyone can ask what was agreed

Every action from a leadership meeting gets an owner, a date and a follow-up

Plant heads set up their own morning brief in plain words, with no request to IT

Quarterly business reviews arrive with every unit's KPIs explained

Each priority project gets a shared space where the team and the assistant work from one record