Author: Simon Clarke OCI/DBA Practice Director & IT Director at Namos
AI Only Works When It Solves Something
AI has gone from a futuristic fantasy to an everyday tool, but real business value from it only comes when it’s used to solve actual problems. For Namos, that meant skipping the hype and asking: What are we wasting time on that AI could handle?
To answer that, the team went department by department, collecting real pain points. In the People team, one clear issue emerged: too much time spent answering repeat questions, the kind that are already documented but buried in SharePoint PDFs.
Two days a month were lost answering questions like “Can I expense this?” or “Do I get my birthday off?”
The solution? Build an internal AI assistant that employees can query directly, no extra licensing, no data scientists, no six-figure projects.
Building AI Using What We Already Had
With an existing Oracle Cloud Infrastructure (OCI) tenancy, the team avoided costly new tools. Instead, we assembled a solution from components already available:
- Object Storage for housing policy documents
- Oracle Generative AI agent for answering user queries
- APEX app to deliver a user-friendly interface
This setup was designed to scale internally and later showcase to clients. In real-world use, the AI assistant is now live on the intranet, and the People team got their two days per month back.
What It Looks Like in Action
Employees type natural-language questions into a simple box in the APEX app. The AI replies clearly and includes a citation showing exactly which document the answer came from.
If a question involves inference (e.g., “Can I expense a bag?”), the agent interprets based on context, like recognising that bags count as clothing under the expenses policy, even if not named explicitly.
Fine-Tuning with Background Instructions
A standout feature of OCI’s Generative AI agent is the ability to adjust how it responds by editing the underlying instructions that guide the AI’s behaviour. These background instructions can shape tone, format, verbosity, or even the language of the response.
Need short, punchy answers for a quick-reference tool? Set the instruction to summarise responses in two sentences. Want detailed bullet points for a policy team? Update the instruction accordingly. Supporting a multilingual workforce? Responses can be returned in other languages with a quick adjustment.
This approach means the same data can serve different audiences, just by updating how the AI is instructed to respond.
Comparing Documents — Instantly
The example AI agent doesn’t just answer questions; it can also compare documents. In one internal test, two fictional CVs were uploaded and analysed for differences. This functionality opens up major time savings in legal, HR, or procurement, where policies, contracts or reports need line-by-line comparison.
Integration with SharePoint and Teams
For organisations that store content in SharePoint or rely on Microsoft Teams for internal communication, integration with Oracle’s AI agents is both feasible and practical. Using Oracle Integration Cloud, files and data can be connected directly to the knowledge base, so employees can ask questions and receive answers without duplicating storage or re-uploading documents.
While SharePoint is already supported through proven APIs, similar connectivity with Teams is under active exploration. These integrations mean the AI agent can be embedded seamlessly into existing workflows, wherever your information lives.
The Cost? Less Than a Cup of Tea
Even with 100+ users, the daily running cost for the Namos AI assistant was around £1.40. Compared to £3,000/month for Microsoft Copilot for 100 users, the savings are significant — especially since OCI’s tools don’t charge per user.
By reusing what’s already licensed, and keeping consumption low, the AI agent delivers ROI without needing extra budget approval.
Final Thoughts: How We See AI
At Namos, we don’t view AI as a silver bullet or a shiny trend, we treat it like any other tool. Its value lies in how well it solves problems. That’s why we don’t implement AI for the sake of it; we implement it where it genuinely saves time, reduces friction, or unlocks new capability for our clients and for ourselves.
We believe AI should be accessible, affordable, and aligned with how your business already works, not something reserved for specialist teams or massive budgets. OCI has allowed us to take that belief and put it into action.
Find out more about our AI Services.
If you have any questions please don’t hesitate to reach out to the friendly team at Namos – Contact Us.
