AMAlex Morgan
- Account
- Northlane Fabrication
- Open request
- Pricing review
- Last contact
- 12 Aug, by email
AI
Put AI where the time goes.
We put AI inside the processes you already run, so the routine moves faster and your team stays on the work that needs a person. First, though, we'll check whether AI is the right answer at all.
AI offTry switching AI on
ActivityPattern identified
Activity increased over the last two months.
Meeting brief Ready for review
Prepare meeting brief
- Recent conversations summarised
- Activity change highlighted
- Open request to discuss
Draft follow-up
Next stepsConfirm scope by FridaySend the revised quote
What we build with
Every platform we work with- Salesforce
- Fin
- Anthropic
- Amazon Web Services
- ElevenLabs
Google Cloud
What keeps it useful
Process first. AI next.
The most useful thing we do on an AI project is often crossing things off the list.
Some work is a great fit. Routine, high-volume, repetitive work can be done by an agent. Other work really isn't, like the call that depends on a relationship or the exception nobody's written down. Put AI on those and you don't save time. You just move the human intervention further down the chain.
So before anything gets built, we go through the work with you. For the work that makes the cut, the usual rules apply.
Sort the work
We split it into three piles: what AI can take off your plate, what it can help with while a person decides, and what stays with a person, full stop. Just as important, we rule out the work it shouldn't touch.
Ground it
We ground every answer in trusted business context, not the model's best guess.
Limit it
We lock down exactly what the agent can see, do and say, and we make sure it's enforced.
Keep a person in the loop
We build in handover points, so a person stays involved wherever judgement, risk or empathy matters.
Measure it
We measure the work before the agent starts and again after, and only keep what's actually better.
If you run Salesforce
A coworker in a week. An agent in a fortnight.
Activate
A working capability, not a recommendation to buy one.
In a week, we switch Agentforce on in your org and connect up to five of your existing systems through Data 360, so its answers come from your real business context.
Get startedWhat's included
- Your data checked before anything goes live
- Up to five systems connected through Data 360
- Agentforce enabled for some of your team, or all of it
- A costed path to the work worth automating next
Best when
You want to see Agentforce working on your own data before you commit to anything bigger.
Optimise
Working agents, one cycle at a time.
In cycles of two to three weeks, we tailor pre-built agents to your org and deploy them across more than one part of the business. Each cycle sharpens what's already there and leaves one or two working agents behind.
Book a chatWhat's included
- Pre-built agents tailored to your org
- One or two working agents each cycle
- A usage estimate for every agent
- ABN, GST and local format checks, and Spam Act and Privacy Act consent
Best when
Activate is live, and you know which part of the business to tackle first.
Transform
Longer processes, with people at the judgement calls.
For the longer processes, over four weeks or more, we build orchestrated agents with Agent Script to handle the repeatable path, with checkpoints wherever your team should keep the decision.
Book a chatWhat's included
- Orchestrated agents built with Agent Script
- Human checkpoints at every judgement call
- Ongoing managed Agentforce as the work grows
Best when
The agents are carrying their weight and the bigger processes are next.
Agentforce licensing and usage are billed by Salesforce.
What AI can do
AI works best with well-defined tasks.
Start with routine, high-volume tasks, where it's obvious when a person should take over.
Service
Turn an incoming request into a ready next step.
Sort the request, pull the right context and draft a response for someone to check.
Human owns
Exceptions, tone and the final action.
Sales
Prepare the account before the conversation.
Bring recent activity, open issues and anything worth knowing into one useful brief.
Human owns
The relationship and the recommendation.
Operations
Move information out of documents and into work.
Pull out, check and route the routine stuff, without adding another copy-and-paste step.
Human owns
Sign-off.
A workflow worth testing
The admin shouldn't take longer than the meeting.
The useful AI opportunity isn't "write something clever". It's pulling the context together, drafting the routine notes and flagging the calls that need a person.
Without AI
- Dig out the account and the last few conversations
- Type up the notes from memory
- Write the follow-up and the next steps
- Update the record and chase the right people
With AI in the flow
- Context pulled together
- Notes tidied and next steps drafted
- You review and approve
- Everything lands where it needs to be
What our AI approach covers
It depends what you want it to do.
Find the jobs where AI would remove real work, and rank them by what they're costing you now.
Put AI into the process you already run. It takes the routine steps and a person keeps the judgement calls.
Fin, now part of Salesforce, answers customer questions across chat, email, phone and WhatsApp. We ground it in your Salesforce records and set where it hands over to a person.
AI where your team already works. Slackbot answers questions, summarises threads and pulls in your Salesforce records, and custom agents take on specific jobs, so updates, approvals and handovers happen in the conversation.
Point it at data you trust, so its answers come from your own records.
Pick the right model for the job, then test it in a safe space before it goes anywhere near live work.
Show your people where AI fits, how to work with it and when to step in.
Baseline before, numbers after. Keep what holds up, change what doesn't.
Is this the right conversation?
Worth a conversation if
- You have high-volume routine work and better things for those people to do.
- Your team's already using AI without anyone signing off on it, and you'd like some guardrails.
- You need a business case for AI, not another proof of concept.
Probably not if
- Nobody trusts the data underneath yet.Data and analytics
- You're hoping AI will cover for a process nobody has fixed.Strategy
- You just want someone to tell you which tool to buy.
Not sure? Let's chat anyway.
“You have to be really honest about where your current data position is. We knew ours was challenging.”
James LakeChief Product & Technology Officer, Prime Financial Group
4specialist platforms connected into one client view
Read the story Where AI goes next
AI is only as good as what's under it.
Bring the job nobody wants to do by hand.
We'll work out if AI can help, and what needs to be done to get there.