Services · Custom builds

Custom builds: apps, automations and AI agents built around your business

The right build starts with the whole business, not the software. I scope your sales, marketing and operations, flag the areas that could and should be automated with AI, then bring in the right AI or build contact. In some cases, I set it up myself.

At a glance

Promptwise Advisory separates custom builds into two kinds of scoping. Founder Adam Agresta does the business scoping across sales, marketing and operations, flagging if, where, what, how and why AI should help, at A$250 an hour, agreed upfront, and hands over a plain-English brief. The builder then does the technical scoping, quotes the build and delivers it.

Business scoping
Adam Agresta, across sales, marketing and operations
Rate
A$250 an hour, agreed upfront
Output
A plain-English brief of what to build and why
Technical scoping
Done by the builder, who quotes and delivers the build
What gets built
Apps and portals, automations and integrations, AI agents

How a build gets scoped

Two kinds of scoping, two different jobs

I work out whether, where and why AI should help your business. Your builder then works out exactly how to build it and quotes the build.

Step 1 · with Adam

Business scoping

  • I look at your sales, marketing and operations end to end
  • I find where time, leads or revenue leak
  • I flag where AI fits, and where it doesn’t
IfWhereWhatHowWhy

You get: a plain-English brief of what to build and why. A$250/hr, agreed upfront.

Step 2 · with the builder

Build scoping

  • The builder, one of my trusted partners, takes my brief
  • They dig into the technical detail: screens, data, connections, security
  • They work out the effort and timeline

You get: a written scope and quote for the actual build, from the builder.

Step 3 · build

Build and launch

  • The builder builds it and is responsible for delivery
  • I can stay on to keep it tied to your sales and marketing goals

You get: a working tool your team uses, on accounts you own.

General overview

How is a modern build put together?

Think of it like a building. People only see the front door, but a lot sits behind it. Every custom build, from a simple internal tool to a full platform, has these eight parts. AI makes the screens and code faster to build. It doesn’t do the thinking for you.

What people see

The front doorThe screens

What your customers and team actually use: a web app, a phone app, a client portal, a chat window, or a helper inside Slack or Teams.

  • Web and mobile apps (e.g. React)
  • One codebase, different screens per role
  • Customer portals
  • Chat interfaces
  • Works inside Slack, Teams or email

The keysLogins and permissions

Who can get in and what each person can see. Your sales rep sees their deals, your manager sees everyone’s, a client sees only their own.

  • Sign in with Google or Microsoft
  • Roles and permissions
  • Row-level security
  • Two-factor login
  • Audit trail of who did what

What does the work

The brainThe AI layer

The AI that reads, writes, sorts and answers, trained on your own playbooks and data, with a human checking anything that matters.

  • Models like Claude, ChatGPT or Gemini
  • Open-source agents (e.g. OpenClaw) where it fits
  • Your knowledge base fed in (RAG)
  • Guardrails and approved wording
  • Human approval steps
  • Usage and cost caps
  • Testing AI answers before go-live

The engine roomLogic and automations

The rules and steps that run in the background: when a lead comes in, do this; every Monday, send that; if a deal stalls, flag it.

  • Server functions and APIs
  • Scheduled jobs
  • Webhooks (instant triggers)
  • Queues for heavy work
  • No-code automations (n8n, Make, Zapier) for simple flows

The filing cabinetYour data

Where everything is stored safely: records, documents, history. Backed up, and kept in Australia where that matters.

  • Database (e.g. Postgres)
  • File storage
  • Encryption
  • Daily backups
  • Australian data hosting
  • Version history

The pipesConnections to your tools

Plugs into what you already use, so nobody types the same thing twice: your CRM, accounting, calendar, payments, email and SMS.

  • CRM (HubSpot, Attio, Salesforce)
  • Accounting (Xero, MYOB)
  • Payments (Stripe, Airwallex)
  • Calendar, email and SMS
  • MCP connectors so AI can use your tools
  • Each outside tool wrapped in one place, so it’s easy to swap

What keeps it standing

The foundationsHosting, security and reliability

Where it runs, how changes are tested before your team sees them, and how you know straight away if something breaks.

  • Hosting and fast delivery
  • Separate test and live versions
  • Version control (every change saved)
  • One-click publish and roll back
  • Monitoring and error alerts
  • Australian Privacy Act alignment

After the keys are handed overRunning and owning it

You own the code, the data and the accounts. Running costs sit on your own accounts, your team gets trained, and someone looks after updates.

  • You own code, data and logins
  • Hosting, AI usage and licences billed to you directly
  • Support and update plan
  • Team training and adoption
  • Measuring time and revenue gained
Where most DIY AI builds fall over: the last two parts. Getting a demo working is the easy bit. Keeping it secure, reliable and affordable once your team relies on it is where scoping pays for itself.

Go deeper: building custom software in the age of AI →

See what it might look like

What could AI do in your department?

Pick a team to see the kind of AI teammates you could build. These are examples, not a product: every build is scoped to how your business actually works.

Your AI sales department

A team of AI helpers that research, write and keep your pipeline moving, so your reps spend their time on calls, not admin.

Research

Account researcherBuilds a one-page brief on any company and its decision makers
Signal watcherFlags funding, hiring and job changes at your target accounts

Prospect

List builderFinds best-fit companies that match your ideal customer
Lead scorerRanks leads by fit and buying signals

Outreach

Message writerDrafts personalised emails and LinkedIn messagesHuman checks
Follow-up sequencerSends timed follow-ups and stops when someone replies

Meetings

Call prepAgenda, talking points and likely objections before every call
Meeting notesRecords, summarises and updates the CRM after every call

Pipeline

Deal-risk spotterFlags stalled deals and missing next steps
ForecasterWeekly forecast from what’s actually in the CRM

Close

Proposal drafterFirst-draft proposals from your templates and call notesHuman checks
Objection coachSuggests answers from your best past replies

Human checks A person approves before it goes out

Plugs into
  • HubSpot or Attio
  • LinkedIn Sales Navigator
  • Apollo or Clay
  • Gmail or Outlook
  • Fathom
  • Claude
A good first build

Meeting notes into the CRM plus call prep. Quick to build, saves every rep hours a week, and cleans up your pipeline data.

Want to know which of these would actually move the needle in your business? That’s what business scoping is for.

Chat with Adam

Examples only. What’s worth building depends on your business, and the builder quotes after scoping.

What can be built?

Three kinds of build, often combined into one connected system. Each one replaces manual work your team does today.

Apps and portals

Software around your workflow

Web and mobile apps for your clients, members, staff or contractors, designed around how you work.

  • Client and member portals
  • Booking and scheduling tools
  • Purpose-built CRMs and internal tools
  • Staff and contractor apps

Automations and integrations

Connect the tools you already have

Information moves between systems on its own, instead of someone copying it between tabs.

  • CRM syncing with your other systems
  • Automatic follow-ups and reminders
  • Lead routing and alerts
  • Live reporting dashboards

AI agents

Repeatable work, handled

Agents that work across your tools on repeatable jobs, with a person approving anything sensitive.

  • Answering common enquiries
  • Making and changing bookings
  • Processing documents and forms
  • Keeping your CRM up to date

Building in the age of AI

Six steps before anyone builds

AI has made software faster and cheaper to build. It hasn’t made it easier to decide what to build. This is how I approach it.

  1. 01

    Map the whole business first

    The best opportunities usually sit between teams, not inside one tool.

  2. 02

    Find the hours and the leaks

    Put a number on the manual work and the revenue slipping through.

  3. 03

    Buy, connect, then build

    Take the cheapest path that works. Only build what nothing else does.

  4. 04

    Prototype small and prove it

    AI makes a first version quick. A demo is not yet a product.

  5. 05

    Decide how much the AI does alone

    Start with AI drafting and a person checking, then loosen the leash.

  6. 06

    Own it and measure it

    Your code, data and accounts, a maintenance plan and a clear result.

35%of teams have replaced at least one SaaS tool with a custom build
78%expect to build more of their own tools in 2026
49%of teams who shipped a build say it saves 6+ hours a week

Source: Retool, 2026 Build vs. Buy Report, a survey of 817 builders.

Read the guide: building custom software in the age of AI →

Should you buy, connect or build?

Custom isn’t always the answer. I’ll tell you when an existing tool will do the job, and I never recommend a build that a subscription could replace.

Buy off the shelfConnect your toolsCustom build
When it fitsA proven tool already covers most of what you needYour tools work, but someone copies data between them by handYour process is your advantage, or every tool forces workarounds
ExamplesA CRM, a scheduler or an outbound tool from the tool directoryCRM synced with your other systems, automatic follow-ups, live dashboardsClient portals, purpose-built CRMs, AI agents
CostMonthly subscriptionsSet-up work plus your existing subscriptionsQuoted by the builder after their technical scoping

Read the guide: build vs buy →

Free · 3 minutes

Is a custom build worth it for you?

The AI and custom builds check shows where your team is losing hours, and whether automating or building would pay off.

Take the Funnel Health Check

FAQ

Frequently asked questions

What’s the difference between your scoping and the builder’s scoping?

Adam Agresta’s business scoping looks at sales, marketing and operations to work out if, where, what, how and why AI should help, at A$250 an hour, agreed upfront, ending in a plain-English brief. The builder’s scoping is technical: screens, data, connections and security, the effort and timeline, and a written quote for the build.

What can AI do for my department?

It depends on the team. In sales, AI can research accounts, write first drafts and update the CRM after calls. In marketing it can repurpose content and qualify leads. In operations, finance, legal, product, customer success and people it can triage requests, chase invoices, check contracts and answer policy questions, with a person approving anything that matters.

How much does a custom build cost?

Every build is different, so there is no price list. Adam Agresta’s business scoping is billed at A$250 an hour, agreed upfront. The builder then does their own technical scoping from Adam’s brief and quotes the build. Cost depends on user types, integrations, payments, AI features, data migration and compliance needs.

Who builds it?

A trusted AI or build contact that Adam Agresta introduces you to with a written brief. The builder scopes the technical detail, quotes, contracts with you directly and is responsible for delivery, and any referral fee is disclosed in writing first. Adam can stay on to keep the build tied to your sales and marketing goals.

How is a modern custom build put together?

Every custom build has eight parts: the front door (screens), the keys (logins and permissions), the brain (the AI layer), the engine room (logic and automations), the filing cabinet (your data), the pipes (connections to your tools), the foundations (hosting, security and reliability) and running and owning it after launch.

Should I build custom software or buy an off-the-shelf tool?

Buy when a proven tool covers most of what you need, connect your tools when data is copied between them by hand, and build when your process is your advantage or every tool forces workarounds. Adam Agresta’s business scoping tells you which applies before anyone quotes a build.

Tell me what’s slowing you down

Start with a free 30-minute chat. I’ll listen first, then tell you honestly whether to buy, connect or build.

Book a free 30-minute chat