Enquiry and intake agents
Assistants that answer on chat, web or email, trained on your real history rather than a generic list of frequently asked questions, with defined escalation rules for anything they should not be answering alone.
Artificial Intelligence (AI) automation
Artificial Intelligence (AI) automation
Chatbots, Artificial Intelligence (AI) agents and workflow automation written as custom software and owned by you, not rented from a platform that can change its pricing, its model, or its terms next quarter.
Most Artificial Intelligence (AI) is a subscription with a wrapper
The market is full of chatbots you configure through a dashboard. They demo beautifully and fail in the specific place your business is specific: they cannot see your booking calendar, they invent an answer when the knowledge base is thin, and nothing they do reaches your customer relationship management (CRM) system. When the vendor changes their pricing tier, your customer service changes with it.
The other failure mode is the pilot that never ships. Someone connects a large language model (LLM) to a document folder, the output is impressive in the meeting, and then it meets real enquiries, the ones with typos and three questions at once, and nobody is willing to put it in front of a customer.
What actually works is narrower and more boring: automate one measured, repeated, well understood task; give it real access to the systems that hold the answer; give it a hard handoff to a human when it is not confident; and measure the percentage it resolves unaided. That number is the whole business case, and it is the number we build against.
What's included
Everything in the fixed quote
Workflow automation
The repeated internal sequence, quote, approval, invoice, follow up and reminder, run automatically end to end, with a human checkpoint wherever the cost of being wrong is high.
Document and data processing
Extracting structured data from the documents, invoices, forms and emails your team currently retypes, with validation so bad input is flagged rather than silently written to the database.
Retrieval over your own content
Answers grounded in your documentation, policies and product data with citations back to the source, so staff and customers can check where a reply came from instead of trusting it blindly.
Systems access, not screenshots
Real integrations into customer relationship management (CRM), calendar, inventory and messaging, so the automation can actually complete the task rather than politely describing what someone else should do next.
Measurement built in
Resolution rate, escalation rate, cost per conversation and time saved, reported from day one. If a workflow is not paying for itself we would rather turn it off than keep invoicing for it.
How it's delivered
The same five steps, every time
- 01
Scope call
You describe the problem; we tell you honestly whether we’re the right fit and what it will roughly cost. No pitch deck, no upsell.
30 min - 02
Written scope & quote
Deliverables, timeline, price. One document. You approve it before anything is billed.
2 to 3 days - 03
Build in the open
You see progress in a shared workspace as it happens, not in a monthly PDF. Real working software from the start, reviewed as it goes.
2 to 8 weeks - 04
Handover
Code, credentials, documentation. You own everything. Ongoing support is optional, never forced.
Launch week - 05
Ongoing growth & upkeep
Search Engine Optimization (SEO), content, campaigns and automation upkeep as retained work once the foundation is live. Optional, never assumed.
Monthly
Proof
Aerlou
An always on Artificial Intelligence (AI) front desk that answers calls and texts, books appointments, routes enquiries and drops the full conversation into the client’s customer relationship management (CRM) system. Built as software, not configured in a dashboard.
Every build, in one place
Questions
Answered in advance
Which Artificial Intelligence (AI) models do you use?
Whichever fits the task and the budget, chosen per project rather than by loyalty. Because the integration layer is ours rather than a vendor’s, swapping the underlying large language model (LLM) later is a configuration change, not a rebuild.
What happens to our data?
It stays in infrastructure you own wherever the architecture allows. We scope data residency, retention and what does or does not get sent to a third party model before the build starts, not after.
What if the automation gets something wrong?
Every workflow ships with confidence thresholds and a human handoff. The design assumption is that it will be wrong sometimes; the job is to make sure being wrong is visible and cheap rather than silent and expensive.
How do we know it is working?
You get the resolution rate, escalation rate and volume from launch. Most engagements target a specific percentage handled without a human, agreed before the build so success is not decided in hindsight.
Do we own the automation?
Yes. It is custom software delivered into your repository and your accounts, the same as any other build. There is no per seat licence and no platform that can hold it hostage.
Tell us the problem, not the specification
One conversation, one honest scope, one fixed quote. You brief the founder directly and hear back within a working day.
Start a project