Discovery
Frame the problem, the value and the risk before anything is built.
- Use case brief with baseline and target
- Risk tier (1 to 4) and autonomy level (A0 to A4)
- Value hypothesis and success measures
Every use case, from a first MVP to agents in production, runs on one life cycle. Nothing moves to the next phase until its gate is signed.
AI projects stall in two places: before anything is built, when nobody agrees on value and risk, and after the demo, when nobody has proved it is safe to depend on. The AIDLC fixes both with a gate at each step.
The same phases, outputs and gates apply whether you buy a fixed-bid MVP, a retained pod or a full capability centre, so work moves between teams without starting over.
Frame the problem, the value and the risk before anything is built.
Make the data the use case needs reachable, clean and permitted.
Choose and tune the model against your own golden set, not a benchmark.
Wire models into workflows, tools and people.
Prove it is safe and accurate before people depend on it.
Run it in production and keep it working.
Set in Discovery and checked at every gate. The higher the tier, the stronger the evidence needed before release; the higher the autonomy, the more checkpoints a person keeps.
Controls map to NIST AI RMF, ISO/IEC 42001 and the EU AI Act. Agents are tested against the OWASP LLM Top 10 before release.
A discovery call, a view of the competitive landscape and a first roadmap, within a day. Free.
Get your AI Blueprint →Way to startWhich use cases pay back first, and what it takes to build them. Ends with a costed roadmap.
Ask about it →Way to startIn three weeks, a register of every AI use, its risk tier and a 90-day plan.
View offering →Way to startA working MVP on your own data, with the numbers to decide whether to scale it.
Ask about it →Way to startTwo days to design an agent: its tools, human checkpoints and autonomy level.
Ask about it →Way to startAn independent check of an AI system that is already live: accuracy, cost, risk and run.
Ask about it →Start small with a fixed-price proof, scale with pods or managed operations, or build a capability centre you own. Rates come from one Stixis rate card, shared with each proposal; nothing on this page is a price.
| Model | How it works |
|---|---|
| Fixed Bid | Defined scope and outcome at a fixed fee; Stixis carries delivery risk. |
| Managed Capacity (T&M) | Named specialists on time and materials, inside your team. |
| Retained AI Pod | A standing cross-functional pod on a monthly run rate, with fixed-bid outcomes. |
| Managed Services | Stixis runs it in production against agreed service levels. |
| India Technology Centre (BOT) | A dedicated centre in Bangalore, built and run by Stixis, then transferred to you. |
| Managed Advisory Retainer | A fixed monthly retainer with a named lead, after a three-month minimum. |
These sit in every proposal and statement of work, so you know where you stand before work starts.
Work product built for you is yours. Pre-existing Stixis tools are licensed to you for the engagement and at handover.
A data processing agreement, data kept where you need it, and no client data used to train shared models.
For generative AI, accuracy is measured against your own golden set, with an agreed target and remedy.
Every use case gets a risk tier and autonomy level, with human checkpoints and an audit trail.
Scope changes are written, costed and approved before work moves.
Leadership and client teams in Dallas–Fort Worth, engineering depth in Bangalore, and delivery in Johannesburg, working to the same life cycle and standards.
Tell us the process you want to change and we'll propose a Discovery plan.