Skip to content
Custom AI Agents Service

You pick the tasks. We’ll put a proven AI agent on them.

We’re past pilots that go nowhere. Choose from our selection of done-for-you AI agents you can customize and we can help roll out in your business. Start transforming your business in 3 to 6 weeks, and it’s yours to keep.

Let’s chat Browse the catalog
Jobs move through one AI agent one at a time: an invoice, a new lead, a signed offer and the Friday report. For each one the agent reads the right system (finance, the CRM or HR) and the job comes out finished. The invoice is $212 over its purchase order, so it stops for the controller, who approves it, and then it is posted.
  • Try your version before you commit
  • Humans in the loop by default
  • Your choice of AI model
  • Support if you want it
Your week, back

The same jobs come back every week. Hand them over.

At a company of 50 to 500 people, a lot of the week goes to work that follows the same steps every time: matching, chasing, copying, reporting. Each job is too small to justify a software project and too frequent to ignore, so it lands on your best people. It is also exactly the work an agent can take.

One operations team’s week, twice. Now: twelve recurring jobs take up most of it, including chasing open orders twice, reconciling and matching invoices, researching leads, updating the CRM, onboarding paperwork, answering the same leave questions, reviewing refund tickets, collecting audit evidence, and pulling and building the Friday report. With agents: each of those jobs is run by an agent and most of the time comes back to the team. A person still approves three invoice variances and two refunds, welcomes the new hire, chases one missing audit file, and reads the Friday report before it goes out.

The agents do the steps. Your people keep the approvals and the judgment calls, and get the rest of the week back.

What it’s worth

Worth far more than it costs.

Four agents from the catalog, before and after. The cost of the work is your team’s time on it. With the agent, it also includes what the agent costs to run.

Invoice matching

For an AP team of six matching about 6,000 invoices a month

$360,000saved a year6,700+ hours backEach month it waits costs about $30,000.
Hours a week on matching
150 hrs
20 hrs
Days to close the month
10 days
4 days
Cost of the work, per year
$429,000
$69,000

New hire onboarding

For a company hiring about 250 people a year

$159,000saved a year3,000 hours backEach month it waits costs about $13,000.
Days from offer to ready
8 days
1 day
HR hours per new hire
14 hrs
2 hrs
Cost of the work, per year
$193,000
$34,000

Ticket triage and refunds

For a support team handling about 12,000 tickets a month

$391,000saved a year9,100 hours backEach month it waits costs about $33,000.
Time to first response
9 hrs
5 min
Hours a week on triage
220 hrs
45 hrs
Cost of the work, per year
$515,000
$124,000

Weekly reporting

For ops and finance leads reporting on eight teams every week

$155,000saved a year2,200+ hours backEach month it waits costs about $13,000.
Hours a week building reports
48 hrs
4 hrs
Numbers copied by hand
1,200
0
Cost of the work, per year
$175,000
$20,000
Battle-tested agents for you

Already working in businesses like yours.

These jobs started inside bigger builds for mid-market teams. They kept coming back, so they became agents in the catalog.

More from client work
  • 4×faster loan closeMortgage brokerage
  • $450Koperating costs cutInsurance brokerage
  • 3×more accounts per CSMFinTech
  • 90%of renewal touchpoints handled automaticallySaaS customer success
The catalog

See how each agent does the job.

Four for each team. Pick one to see the screen your team would use, and watch it run until it stops to ask a person. They’re sorted by team because finance is finance anywhere. Your industry is part of the setup.

  • Invoice matchingMatches invoices to POs and receipts.
  • Collections follow-upChases overdue invoices every morning.
  • Three-way reconciliationTies sub-ledger, ledger and bank daily.
  • Expense report reviewChecks receipts against your policy.
  • Inbound lead briefResearches and routes every new lead.
  • Quote creatorDrafts quotes from your price book.
  • Proposal generatorTurns a discovery call into a proposal.
  • CRM clean-up after callsUpdates each deal from the call.
  • Request intakeGets internal requests approved and done.
  • The Friday reportHas the weekly report ready by Friday.
  • Cross-system handoffsSets up a signed order in every system.
  • Software vendor approvalRoutes new tools past IT, legal and finance.
  • New hire onboardingTurns a signed offer into day one.
  • Policy and leave questionsAnswers HR questions from the handbook.
  • OffboardingCloses every account on the last day.
  • Leave requestsApproves time off, updates calendars.
  • Ticket triage and refundsResolves what is within your limits.
  • Conversation reviewReviews every call, chat and email.
  • Escalation handoffHands tickets over with the full story.
  • Churn risk alertsSpots quiet accounts before they leave.
  • Control evidenceCollects evidence for every control.
  • Auditor requestsAnswers request lists from the record.
  • Filing calendarKeeps every filing on time, with an owner.
  • Vendor risk reviewsRuns security checks on every new vendor.

The full catalog

60 agents across six teams. The ones marked “See it run” are demoed above. If your job isn’t listed, we start from the closest one.

Finance10
  • Close checklistTracks every close task and its evidence
  • Accrual and treatment checkApplies the same test to every entry
  • Payroll report categorizerCodes every payroll line to the right account
  • Investor reporting packBuilds the quarterly pack from live numbers
  • Budget vs actualExplains each month’s variances to budget
  • Vendor inquiriesAnswers vendor payment questions from the ledger
Sales & Marketing10
  • Lead scoring and routingScores fit and sends leads to the right rep
  • Lead nurture sequencesTailors follow-ups to what each lead does
  • Recordings into clipsTurns podcasts and webinars into shareable cuts
  • Competitor and market digestReads the news and sends what matters
  • Commission calculationsWorks out commissions from closed deals
  • Ask the pipelineAnswers pipeline questions from your CRM
Operations10
  • KPI roll-upPulls every team’s numbers into one view
  • Internal questionsAnswers how-do-I questions from your docs
  • Approval routingSends each request to the right approver
  • Spend limit checksFlags purchases over the limit before they go
  • Vendor onboardingSets up new vendors with tax and bank checks
  • Contract renewalsFlags vendor contracts before they auto-renew
HR10
  • Headcount reconciliationKeeps HR, payroll and the org chart agreeing
  • Offer letter checksChecks every offer against the pay band
  • Benefits enrollmentChases missing elections before the deadline
  • Compliance paperworkFiles and tracks the forms every hire needs
  • Interview schedulingBooks interviews across every calendar
  • Training trackerChases required training before it lapses
Customer Support10
  • Help articles from ticketsDrafts help articles from repeat questions
  • Renewal prepBuilds each renewal brief ahead of the date
  • Promise check on repliesCatches replies that promise too much
  • Call transfer with contextPasses the call with the account attached
  • Customer onboardingRuns each new account’s setup checklist
  • Feedback themesGroups survey and ticket feedback into themes
Compliance10
  • Reportability checkApplies one written test to every incident
  • Policy attestationsCollects and chases policy sign-offs
  • Remediation trackingKeeps every finding moving to closed
  • Evidence packs on demandAssembles an audit pack in minutes
  • Access reviewsRuns quarterly access reviews across your systems
  • Contract obligationsPulls every obligation and date out of your contracts
Let’s chat
What you get

Everything it takes to go live, whichever agent you start with.

  1. 01Before you commit

    Your version of the agent, clickable

    Click through it with your data, your rules and your approvers in it. You decide on the build after you’ve tried it.

  2. 02

    The agent and its screen, set up your way

    Your triggers, your naming, your exceptions, your approval limits.

  3. 03

    Connected to your systems

    Its own login, with only the access its role needs.

  4. 04

    A side-by-side run

    It works next to your team, and people check every output first.

  5. 05

    A scorecard

    One number, today’s baseline and a target, agreed up front.

  6. 06

    A trained team and a runbook

    What it does, who owns it, and how to pause it.

  7. 07

    30 days of tuning after launch

    We keep tuning it on your live work for the first month.

Anatomy of your agent

Not a chat window. A tool your team works in.

Say “AI agent” and most people picture a chat window you type questions into. The useful ones mostly work in the background instead. They do the repetitive steps on their own and put the result on a simple screen for a person to check. Each one we set up has three parts, and we build all three.

An exploded drawing of three stacked layers. On top, the interface: an app window with a queue of work on the left, each row carrying its own status, and a detail card on the right where a finished result has landed. In the middle, the logic layer: six steps wired in two branches that both run forward into a final step. Three are marked with a star and run on AI, two are rules, and one carries a person and stays a human decision. Underneath, the data layer: one wide record of fields that every link from the layers above lands on, ringed by eight of the tools the team already works in, two set into each edge of the layer, drawn as an inbox, a chat window, a calendar, a set of files, a spreadsheet, an invoice, a contact record and a database, each wired into the record. A job drops from the interface into the logic layer, down to the data, and comes back up as a finished result.

01What your team sees

The screen your team works in

A simple screen: a list of work to review, a checklist, a dashboard. Your team opens it, sees what the agent did, and approves it with a click.

02What the agent does

AI does the steps. People make the calls.

The steps someone used to do by hand: reading, matching, checking, drafting. It follows your rules, and when something falls outside them it stops and asks a person.

03What it connects to

Integrated with your systems

The tools you already pay for, like your accounting software, CRM, HR system and inbox. It works inside them, so nobody copies data into yet another tool.

Our process

Start from ours. Make it yours.

The agent works before we meet you, so the weeks go on your process, your rules and your team.

Done for you, fitted to you. We run it on a process proven across dozens of deployments.

  1. Already built
    Before we start

    Pick an agent from the catalog

  2. Shaped to you
    Before kickoff

    See your version, clickable

  3. Filled in
    Kickoff

    Short interviews fill the gaps

  4. Connected
    The build

    Revise it and connect your systems

  5. Live
    In 3 to 6 weeks

    Run it alongside your team, then go live

It’s yours to keep

It runs in your accounts, on your rules.

We set the agent up inside your cloud and your tools, on a model account in your name, and you decide what it can touch and what waits for a person. You pay the model provider directly, and it keeps running whether or not we stay involved.

Everything sits inside the client’s own accounts: the agent and its interface, their systems and data (such as a CRM, a project tool or accounting software), a model account billed to them by the provider (a frontier model from OpenAI, Anthropic or Google, or an open-weight model such as DeepSeek, GLM, Qwen or Llama), and the code and settings. It is hosted where the client already hosts, or set up for them on AWS, Microsoft Azure, Google Cloud, Vercel or Cloudflare, all providers with SOC 2 reports.

Three controls come with every agent.

So you always know what it can reach, what it can do alone, and what it did.

  • It only sees what it needs.

    It gets its own login, like a new hire. The invoice agent can read accounting, but not payroll.

  • Big calls wait for a person.

    You set the limits. An invoice over $5,000 stops and waits for your approver. The rest goes through.

  • You can see everything it did.

    Every step is logged: what it read, what it changed and who approved it.

Next step

Got a job eating your team’s week?

Tell us what it is and what it costs you each month by hand. On one call we’ll go through the catalog with your business in mind and sketch how the first agent would work for you.

Let’s chat
AI agent use cases

Looking for AI use cases?

Eighteen everyday workflows from six teams, step by step: what automation handles, where an agent fits, and what stays with a person.

See the use cases
How we compare

Plenty of ways to get an agent. One that stays yours.

Agent SaaS is subscription software in a new suit. Building your own takes a team, and consultants leave a plan. We start from an agent that works and make it yours.

Switchboard Agents
Agent SaaS
Build it yourselves
Consulting firms
What you get
An agent running on your systems, set up to your rules.
The same product every customer gets.
A project for people you would have to hire.
A strategy deck.
Before you commit
Your own version, to click through.
A demo of their product.
A spec to sign off.
A proposal.
Fit
Your process, approvals and exceptions, written in.
Their workflow. You change yours to match.
As deep as you have time for.
Their framework, applied to you.
Speed
Running in 3 to 6 weeks.
Fast to buy, slow to fit.
Usually months.
Months of discovery.
Model
Your choice, including your own.
Theirs.
Whatever you can wire up.
A recommendation.
Ownership
Your accounts, your code, your model bill.
Rented, per seat, every year.
Yours, if it ships.
A plan you still have to build.
After launch
Your team runs it, or keeps us on.
A support queue and a roadmap vote.
You are the support team.
The engagement ends.
Questions

What people ask before they start.

What is a Switchboard agent?
Software that does one recurring job from start to finish. It starts on a trigger, reads the systems the job needs, does the steps, and hands anything outside your rules to a named person. Each one starts from a version in our catalog, then gets set up for your business.
Do we have to commit before we see it?
No. Before the build starts, you click through your own version of the agent, with your data, your rules and your approvers in it. You decide whether to go ahead with the build after that.
How is this different from an AI Jumpstart?
A Jumpstart is for when you are still working out where AI belongs, and it ends with a map and a working prototype. Agents are for when you already know the job, and they end with the agent running on your systems. Plenty of clients do a Jumpstart first and pick their agents from what it finds.
Do we need engineers on staff?
No. We do the setup, the connections and the testing. From your team we need the people who do the work today, for a few hours at kickoff and while it runs alongside them, and someone who can grant access to the systems involved.
Which systems can an agent connect to?
Most of what a mid-market company runs: accounting, CRM, HR and payroll, ticketing, shared drives, email and spreadsheets. If a system has no way in for software, we tell you in week one, before anything gets built.
Which AI model does an agent run on?
Whichever suits the job and your rules. That can be a frontier model from OpenAI, Anthropic or Google, an open-weight model such as DeepSeek or GLM running on servers you control, or an account your company already pays for. You can switch later without changing the agent’s steps, rules or approvals.
What happens when the agent gets something wrong?
Anything outside the rules you set goes to a person instead of going through. Before it goes live, the agent runs next to your team and people check every output, so the rules get tuned on your real work before it takes anything on by itself.
Who owns the agent after launch?
You do. It runs on your accounts under the permissions you set, and one person on your team owns it and can pause it. The first 30 days of tuning after go-live are included. After that we can stay on to look after it and add the next one, or hand it over.
What does it cost to run once it is live?
Hosting and the AI it uses, billed to your own accounts by the providers, not through us. We estimate both in week one from your real volumes, so you know the running cost before you go live.
Can you build an agent that is not on the list?
Yes. If the job is close to one in the catalog, we start from that. If it is new, it becomes a custom build through the AI Deployment Studio.