Services
Auranium helps with three things: clarifying the product before investing, building a product people can use, and coaching businesses to deploy AI Agent systems themselves.
Start from your team's current situation. You don't need to pick the right service name first.
Start from where you are
You don't need to pick the right service name first. Choose the situation closest to your team and Auranium suggests where to start.
Product consultingSuggested starting point
Before hiring a build team, clarify the users, evidence and assumptions. The conclusion may be to build, adjust or stop.
Product consultingSuggested starting point
Consulting aligns goals, priority reasons and scope boundaries before every request gets pushed into one version.
Product developmentSuggested starting point
Design, build and test in rounds with a viewable version, so the team can give feedback before scaling.
Product developmentSuggested starting point
Review code, architecture, data, permissions and tests to decide what to keep and what to reinforce before going to production.
Coaching AI AgentSuggested starting point
Four 1:1 coaching sessions: you install it yourself, connect tools, move the system to a server and set up cost control.
Coaching AI AgentSuggested starting point
High-frequency, standardised work is where AI Agents cut cost and time the most. The first three processes are chosen in session 1.
Three service scopes
Clarify the product before hiring a build team.
A fit when
Requirements are still scattered, stakeholders aren't aligned, or decision-makers don't know whether to invest in building.
What Auranium does
Clarify users, evidence, assumptions, priority scope and what still needs validating. The conclusion may be to build, adjust or stop.
Outputs
Design and build a version people can actually use.
A fit when
The team already has goals and scope and needs a website, sales page, digital product or internal tool that really works.
What Auranium does
Design, build and test in rounds with a viewable version so the team can give feedback before scaling.
Outputs
Deploy and run AI Agents in your business yourself.
A fit when
Owners or managers who want AI in real processes but don't yet have the in-house ability to deploy and maintain it.
What Auranium does
1:1 coaching over four sessions. Participants install it themselves, connect tools, move the system to a server and set up cost control.
Outputs
01Product Consultant
For business owners, founders and teams who need to decide what to do first. Auranium works with you to clarify users, the problem to solve, the priority scope and the assumptions to validate.
You receive a set of documents the team can use to align on decisions, whether you build it yourself, with Auranium or with another development partner.
When to use it
This service fits when the investment question is still open, requirements aren't aligned or the scope has no clear boundaries.
There is an initial direction, but it's unclear who will use it and which problem is worth solving first.
Several parties make different proposals, without agreeing on goals or priorities.
The wish list is long, but what to do first and what to postpone isn't decided.
The team needs to separate process problems from tool problems before choosing software or AI.
Core outputs
Each deliverable helps the team see what's known, what's uncertain and the boundaries of the next step.
Record the problem, users, evidence and assumptions to validate.
Describe the main usage flow, what's in and out of scope, and the reason for each limit.
Weigh impact, strength of evidence, risks and dependencies instead of relying on a wish list.
Record the actions, decision conditions, open questions and who needs to confirm.
How we collaborate
You don't need a long specification. The starting point is your current situation and the decision you need to make.
You share your goals, current situation, existing documents and the questions to decide.
Both sides agree the questions, outputs, responsibilities, review method and fee before starting.
Auranium cross-checks data, processes and user feedback within the approved scope.
Both sides review the proposal and record decisions, open items and next steps.
Handover structure template
Example: managing customer support requests — current state, evidence and scope side by side to weigh the next step.
Support requests arrive through many channels; the team needs to know who is handling what.
Cross-check conversation history, how work is assigned and the cases that were missed.
One place to track status could reduce manual handoffs.
Fix how work is assigned, use an existing tool or build a dedicated management flow.
Prioritise intake, owner and status; AI isn't in the flow yet.
Is the bottleneck a missing tool or unclear responsibility?
Additional scope
The items below aren't included in every project by default. Auranium proposes them when needed and agrees them separately before doing them.
An interview or test plan, the questions to answer and the conditions to continue.
Used when a flow or way of working needs to be seen and given feedback before development.
Review integrations, data, dependencies and risks when the decision needs more technical evidence.
Review data, execution rights, exceptions and approval points. The conclusion may be that AI isn't needed yet.
After consulting
Consulting creates no obligation to hire Auranium for development. The documents can be used by your in-house team or another development partner.
Use the documents to align internally, validate further or talk to another development partner.
Once the direction is chosen, Product Development prepares the technical delivery and builds within the right scope.
Product development → 03If the process fits and the owner wants to deploy it themselves, see the AI Agent Coaching programme.
Coaching AI Agent →02Product & AI Development
You can start from a problem, a process, an idea or a prototype. Auranium helps clarify what to build, design the experience, develop the system, apply AI where it fits and take the product to production.
Starting points
Many products start from a manual process, a prototype or changing requirements. The first job is to reduce the uncertainty about what to build.
Data entry, file transfers, checks and approvals make work slow or hard to track.
The feature list is long, but it's unclear which parts are needed to validate the direction.
The use case, data, business rules and human approval points aren't clear yet.
The demo needs architecture, permissions, integrations, testing and operational monitoring.
What can Auranium build?
Pick a product type to see where it's used. Technology changes by project; the starting point is always the users, the process and the smallest useful scope.
Multi-user product
01/10A platform serving many accounts or organisations, with clear sign-up, permissions, operations and administration.
Turn a repeatable process or service into a subscription product.
Self-service experience
02/10A secure space for clients to view data, submit requests, track progress and receive deliverables.
Reduce scattered email back-and-forth and give clients a single point of access.
Internal operations
03/10A system that connects data, tasks, approvals and reports around how the business actually works.
Replace spreadsheets, file transfers and manual handoffs with a trackable workflow.
Multi-sided platform
04/10A product connecting supply and demand, with profiles, search, transactions, controls and administration.
Create a trusted exchange mechanism between multiple user groups.
Bookings and service capacity
05/10Manage availability, service types, bookings, confirmations, rescheduling and back-office operations.
For services with limited resources such as experts, rooms, equipment or time slots.
Browser application
06/10A product with its own workflow, data and permissions, used directly in the browser.
Build business tools, dashboards, self-service platforms or B2B products.
On-device experience
07/10Mobile experiences tied to on-site needs, notifications, camera, location or quick actions.
When the main work happens away from the desk or needs frequent interaction.
AI inside the product
08/10AI is placed in a flow with data, criteria, access control and a way to evaluate output.
Document understanding, search, recommendation, copilot, classification or extraction.
AI inside the workflow
09/10A chain of repetitive steps connected with integrations, business rules, approvals and exception handling.
Routing, drafting, reconciliation, knowledge retrieval, approval or system updates.
Industry-specific software
10/10A product designed around the operations, roles, data and control requirements of a specific industry.
When off-the-shelf tools don't reflect the real workflow and domain constraints.
Starting points
Each starting point addresses a different kind of risk. Pick one to see its scope and outputs.
When the scope is unclear
01 / 06Define the problem, users, process, product scope, AI opportunities, risks and technical direction before committing to Build.
Outputs
When you need to test first
02 / 06Test user flows, experience or technical and AI feasibility before investing in the full system.
Outputs
When the direction is clear
03 / 06UX/UI design, architecture, frontend, backend or mobile, integrations, QA, infrastructure, analytics and production launch.
Outputs
When work is too manual
04 / 06Map the workflow, design AI logic, business rules, approval points, integrations, exception handling and monitoring.
Outputs
When a demo needs real users
05 / 06Review code and architecture, strengthen the data model, authentication, permissions, security, integrations, testing, monitoring and deployment.
Outputs
When the product is live
06 / 06Monitor stability and usage data, fix bugs, improve flows, tune AI and build new integrations.
Outputs
AI, but it must solve a real job
AI sits inside a workflow with data, business rules, access control, exception handling and traceability.
Documents, requests or operational signals are standardised before processing.
Data sources, usage rights and background information are clearly controlled.
Classify, search, extract, draft or recommend based on the goal.
Business conditions decide which work continues, stops or is handed off.
Risky decisions stop at a checkpoint for a human to approve.
Actions are logged, fed back and brought into an audit loop for improvement.
From demo to real product
AI and no-code make demos fast. Once real users arrive, the product needs data, access control, fault tolerance and a clear operating process.
Development process
Every phase ends with a reviewable output. The client takes part in decisions about priorities, data and business rules.
The business problem, users, processes and constraints.
Scope, requirements, success criteria and risks.
User flows, prototypes, UI and validation.
Architecture, development, integration and QA.
Production, analytics, monitoring and handover.
Feedback, metrics, improvements and post-launch support.
Auranium
Analysis, recommendations, Product, Design, Engineering and delivery management.
Client
Domain knowledge, access to users and data, priorities and commercial decisions.
No big build commitment while scope or risk is still vague.
Assumptions, exclusions, dependencies, acceptance criteria and changes are all made explicit.
Code, design and documentation are handed over according to the agreed scope and terms.
Approval, logging and exception steps are designed right into the workflow.
031:1 Coaching Programme
A hands-on 1:1 programme for business owners and managers. After four sessions, you can deploy, configure and run an AI Agent system for your own business. No programming knowledge required.
04
sessions, 90 minutes each
1:1
One-to-one, one person
15
days of support after the programme
Context
Most businesses use AI as Q&A and then copy the result into other tools. AI shortens each step but doesn't change the cost structure of operations.
Using AI in chat sessions
Every session starts from scratch; products, clients and templates must be entered again.
Results stop at text. Sending emails, updating spreadsheets and exporting files are still manual.
A process that runs 20 times a month needs 20 separate sets of instructions.
Running on AI Agents
Profiles, product documents and presentation standards are loaded once and used in every session.
Chat, email, calendar, storage, spreadsheets and publishing channels.
High-frequency, standardised work is handled at significantly lower cost and time.
Analyses requests and makes decisions. Choosing the right model for each type of work is what determines operating cost.
Connections to chat, email, calendar, storage, spreadsheets and publishing channels. They define the scope of work the system can execute.
Company profile, product documents, report templates and presentation standards, loaded once and used from then on.
Training roadmap
Each session completes one deployment phase and ends with a component already running on the company's system.
Session 01 · 90 min
Goal: Understand what AI Agents can and can't do, have a system running on your own device and define where it applies in the business.
Session 02 · 90 min
Goal: Move the control interface from the command line to the company chat and standardise how tasks are assigned so output quality is consistent.
Session 03 · 90 min
Goal: The system runs independently of personal devices, with access control and a recovery plan for incidents.
Session 04 · 90 min
Goal: Make costs predictable and turn stable processes into assets the business owns.
Training principle. From session 2 on, the participant operates the system directly while the trainer guides verbally. The business can maintain the system after the programme, without depending on a vendor.
Outcomes
Install the AI Agent on a personal machine and a private server. Read error logs, restart services and update versions.
Set up departments in the chat channel according to the org chart, assign work and receive results — no new software needed.
Write requests with enough information to get usable results the first time — a core skill independent of any tool.
Understand the cost structure, choose models by type of work, set spending limits and apply optimisations.
Package stable processes into report, email and proposal templates and presentation standards.
Classify work by frequency and standardisation to produce a prioritised list of next processes.
Who should join
How it runs
A 30-minute fit call, a short survey about size, tools and the three most time-consuming kinds of work. Then you receive a preparation checklist.
Four private sessions, one a week. Between sessions there are exercises on real work and a dedicated support channel.
A dedicated technical support channel after the last session, plus a 30-minute review to fix anything still unstable.
Programme fee
Contact
For one participant. Includes the intake assessment, four training sessions, the guide set and 15 days of technical support.
Book a 30-minute callServers and AI models, paid by the business directly based on usage. The estimate is made in session 4.
1:1 training, with content adapted to the business's industry and processes.
Two instalments: when the schedule is confirmed and after session 2.
General theory, multi-company classes, programming knowledge, turnkey deployment.
Tools and programmes
You can assess your AI readiness yourself or join the client referral programme.
Frequently asked questions
On the scope, way of working and responsibility limits of each service.
You can start from your goals, current process and the questions to decide. The scope is proposed after Auranium reviews the inputs.
No. The outputs can be used with your in-house team or another development partner, according to the agreed handover scope.
We can't confirm that before understanding the problem. The approach and scope depend on the questions to answer and the evidence available.
No. Consulting clarifies decisions and risks. Market demand still has to be validated with the right evidence.
Not by default. Prototypes, technical reviews or development are scoped separately if needed.
Auranium reviews the problem, outputs and responsibilities needed, then sends the scope and fee for your approval before starting.
Next step
Clarify your current stage to choose the right kind of support instead of investing in the wrong direction.
Start your journeyWe'll listen, ask the right questions, and tell you honestly if we're not the best choice for you.
Phone number
(+84) 094 4686 093
Zalo
(+84) 094 4686 093
Working hours
Mon-Sat (08:00 - 21:00)
