Prompt clinic
Rewrite five weak prompts until each produces usable output first time.
Twelve weeks from AI basics to a working GenAI project you can show your manager.
An XcellAI Academy programme for working professionals and teams: self-paced lessons, live sessions with an instructor, graded labs and a capstone, delivered by XcellHost's AI practice in Mumbai.
A one-off workshop shows what AI can do. Habits change when people practise on their own work every week, get feedback and finish something real. The integrated format is built around that.
A new tool every week, no method
Six prompt patterns that work in ChatGPT, Copilot, Gemini and Claude
Watching demos, never practising
Eight graded labs on everyday work tasks
Customer data pasted into public chatbots
A data traffic-light rule and DPDP basics from week one
Enthusiasm fades after a month
Twelve weeks of live check-ins with a cohort
Nothing to show at the end
A capstone demo and an XcellHost certificate of completion
Not a data-science or model-building course. If your team writes code against AI APIs, ask us about a developer track*.
Each outcome is practised in a lab and checked in the capstone, so it is something you have done, not just heard about.
Tokens, context windows, training cut-offs and why models make things up.
Use six patterns to get consistent, well-formatted output on the first or second try.
Draft, summarise and analyse with Microsoft 365 Copilot or Gemini in Google Workspace.
Use Perplexity and Gemini Notebook (formerly NotebookLM)* and check every claim against a source.
Clean a sheet, chart it and catch the errors AI introduces into formulas and totals.
Set up a custom GPT, Gem, Claude Project or Copilot agent, plus a flow with a human approval step.
Apply a data traffic-light rule, DPDP basics and your company's acceptable-use policy.
Present a capstone with a before-and-after time or quality measure on a real task.
Every module mixes self-paced lessons, a live session and lab time. Open a module to see topics, its lab and how it is assessed. Hours are typical and confirmed in your joining pack*.
Your final mark combines three parts. Pass with 70% overall*, a submitted capstone and attendance at 75% of live sessions* (recordings count when you are excused). One free capstone resubmission*.
Labs use sample or synthetic data supplied by us, never your customers' personal data. Each is graded on a short rubric, and feedback comes back within the week*. Hover or tap a lab to see the rubric.
Rewrite five weak prompts until each produces usable output first time.
Turn a rough brief into a client email, a one-page note and five slides.
From a sample meeting transcript, produce minutes, owners and a follow-up mail.
Summarise a 20-page sample policy with a citation for every key point.
Clean a messy sample sales sheet, chart it and write three findings.
Build an assistant that answers only from three approved sample documents.
Automate: a form arrives, AI drafts a reply, a person approves, then it sends.
Find the hallucinations, the prompt injection and the data leak hidden in a sample workflow.
Module 7 goes deep, but the data traffic light is taught in week one and used in every lab. The course explains the rules; it is not legal advice, and your organisation's own policy always comes first.
The DPDP Rules were notified in November 2025 with phased commencement; most obligations apply from May 2027. Breaches go to the Data Protection Board without delay, with a detailed report within 72 hours.
Taught in M0, M7 · applied in every labOrganisations must report specified cyber incidents to CERT-In within 6 hours. Learners practise spotting and escalating AI-enabled phishing and data leaks fast.
Taught in M7 · Lab 8Principles-based guidance under the IndiaAI Mission that relies on existing laws. The course uses its themes of accountability, transparency and human oversight.
Taught in M1, M7Since 2 February 2025, providers and deployers of AI systems in the EU must take measures for staff AI literacy. Completion records from this course can support that evidence.
Relevant if you serve EU clientsWant an AI policy and governance for the whole organisation? See AI Adoption Framework and ISO 42001 consulting. Regulatory dates checked October 2026.
Self-paced lessons and labs are online in every format. What changes is where the live sessions happen and who is in the room. Dates below are sample intakes, to be confirmed with XcellHost.
Open cohort. Weekly live session, cohort chat and recordings.
Sat 10:00–12:00 IST*Live sessions in a Mumbai classroom; venue confirmed on booking*.
MumbaiOur instructor runs the live sessions at your office anywhere in India.
Travel at actuals outside MumbaiPrivate batch for your team, with labs and capstones on your own use cases.
On quoteStarts Sat 9 Jan 2027* · live Saturdays 10:00–12:00 IST*
Live online · MumbaiStarts Tue 2 Feb 2027* · live Tuesdays 19:30–21:30 IST*
Live onlineStarts Sat 6 Mar 2027* · live Saturdays 10:00–12:00 IST*
Live online · MumbaiStart when your team is ready, minimum batch size agreed on quote*.
A consultant from XcellHost's AI practice who also runs AI rollouts for clients, so examples come from real projects.
Reviews your labs against the rubric, runs office hours and answers questions in the cohort channel.
From XcellHost's security and compliance team; leads Module 7 and checks capstones for data risks.
Helps you pick a task, set the before-and-after measure and prepare for demo day.
Prices are in rupees and exclude 18% GST. They are indicative placeholders — confirm with XcellHost before you enrol. There is no early-bird offer.
Training a whole department? Pair the cohort with the AI Use-case Discovery Workshop (₹49,999*, excl. GST) so capstones target the use cases your leadership has already prioritised.
Build an indicative order for yourself or your team. Add seats to the cart, or send the configuration to us for a quotation — groups of 10 or more usually move to corporate cohort pricing.
Seven questions are scored from 0 to 3; the last one picks a domain course if one fits your role. Your answers stay in your browser.
Answer the questions to see your recommended track.
Static, worked examples — nothing is sent to an AI service. Pick a pattern to compare a weak prompt with a strong one and a sample of the output it tends to produce.
This looks like it contains personal data (an email, phone or ID number). Remove it before pasting into any AI tool.