CEO and CTO that has done this and learned the hard way
What I keep seeing
AI moves fast. Reading is not doing. You need to act.
You bought AI tools. The seats sit empty.
Vendor pitches all sound the same. You need a peer.
Config, security, integration. The hands-on parts, not another deck.
I use Claude. I use ChatGPT. I use whatever wins next. Your work lives in your own fileshares like SharePoint, OneDrive, or Google Drive. You own it. You can switch any time.
Partner
Blue Sheen
2026 - Present
AI advisory firm for mid-size companies. Co-founded with Pravina Pindoria. Senior operators on every engagement. No juniors handed the work. All my consulting work now runs through Blue Sheen.
Co-Founder & CEO
Tallyfy
2014 - Present
Built Tallyfy from scratch - the most easy and cost-effective workflow automation platform available. Invented Workflow Made Easy - a product that anyone can use to document, run and track business processes. Currently running most operations with AI, with deep knowledge of where it works and where it fails.
Co-Founder
SternStella (idea stage)
2025 - Present
Currently at the idea stage.
Partner
BrightKite.ai (idea stage)
2025 - Present
Currently at the idea stage.
Entrepreneur in Residence
Skandalaris Center, Washington University in St. Louis
2025 - Present
Helping deep-tech and all kinds of entrepreneurs with AI, technology, and operations.
Faculty
OneDay (a Woolf college)
2025 - 2026
Taught cohorts of MBA students about lean operations, automation, and effective AI use. Also taught creativity and innovation to early-stage founders.
Process Improvement Consultant
Dachis Group
2007 - 2012
Joined via the Headshift acquisition - Headshift pioneered Social Business consulting and Dachis Group acquired it to scale that work. Clients were City law firms, large corporates, blue chip companies, and charities. Problems mixed technology, culture, and process.
Process Improvement and Systems Integration
Atos Origin
2004 - 2006
BPM, systems integration, and process improvement for UK public sector and enterprise clients including a local authority portal, NHS Choose and Book, and VOSA. Stack: Tibco, BEA WLI/WLP, Java.
University of Bath
Computer Science (BSc)
2004
Most companies are stuck at stage 1 or 2. Here's the full picture.
Shadow AI on personal accounts
Managed AI with proper logins and audit trails
From "I use my own ChatGPT" to everyone on one secure system
Typing prompts at 40 WPM
Voice input at 150+ WPM with structure baked in
From hunt-and-peck typing to just talking
Saving prompts in Word docs
AI asks YOU the right questions
From prompt libraries to AI that knows what to ask
Copy-pasting between AI and your apps
AI acts inside your apps directly
From tab-switching hell to AI that just does it
AI knows zilch about your business
AI loaded with your actual data and rules
From generic slop to answers that know your world
AI only runs when someone remembers to ask
AI fires on its own when things happen
From "oh I forgot to run that" to automatic
Three phases. No fluff. Just the work that gets AI running.
Real testimonials from colleagues and clients on LinkedIn
"As a non-technical founder, I've been lucky enough to be coached by Amit on leveraging AI tools like Claude for my business. How to think, how to plan, how to build, how to execute, etc., etc. If you're lucky enough to get a chance to work with him - just do it! It'll be one of the best investments in yourself and your business. If you really want to learn the true fundamentals in a short period of time, I can't recommend working with him enough. He's also a genuine guy who is easy to work with!
"
Founder at Risecap
"Amit is an exceptional AI expert and outstanding mentor. I worked with him on a language-model product, and his guidance shaped the direction of the project. He combines deep technical expertise with strong product design insight, helping me think more clearly and identify stronger directions. He was patient, communicated complex ideas clearly, and helped me make better decisions with confidence.
"
Research Assistant at WashU
"Amit was my mentor and teacher during my MBA process. I felt like I was decent with AI at the time- but Amit completely transformed how I think about using AI. He explains complex concepts in an easy to understand manner. He is creative in his thinking with practical outcomes. He was also able to help me one on one which was very productive. Highly recommend working with him and just all around great human!
"
Oncology Partner at Guardant Health
"Amit is generous with his time and knowledge to help students across St. Louis to learn and succeed with AI. Very thankful for Amit and the great partnership established with UMSL.
"
Director of IP Management and Commercialization at UMSL
"I had the wonderful opportunity to benefit from Amit's mentorship at the Skandalaris Center for Entrepreneurship at WashU. Speaking with him was mind-blowing and deeply inspiring, and it gave me new insights while opening fresh and original perspectives for my project. Amit shared many great ideas, as well as a full range of practical advice (spanning from customer discovery, problem solving, to AI...) and actionable tips on how to implement them effectively. He was extremely generous with his time, fully committed, and truly dedicated to his mission as both a mentor and a role model.
"
Post-Doc Researcher in Neurosciences & Entrepreneur
"Amit is a whirlwind of ideas. And what is better, he is usually sharing them with others. Every idea, even the ones that will not make it into a project in the end, is bound to contain some inspiring nugget that will eventually get you thinking about scenarios and possibilities for days.
"
Staff Backend Engineer at Deliveroo
"There really is only one Amit. With a unique blend of strong technical capabilities, entrepreneurial spirit and deep understanding of user behaviour, Amit has the enviable ability to sniff out problems and then (seemingly) effortlessly create innovative business solutions with infectious enthusiasm and relentless follow through. Equally as important, Amit is absolutely lovely to work with. He's got an unflappable positive outlook, loads of energy and is not afraid to ask the tough questions.
"
Jodi Schneck
Head Teacher at Jodanno Learning
"Amit is a fantastic, lateral thinker. His entrepreneurial spirit is matched by his kindness, presence and ability to generate great enthusiasm. Amit can teeter between the logical and the dreamy with an envious ease.
"
VP, Design at Aruliden
"Amit and I had a lot of fun working on some really interesting web projects. With lots of hurdles and challenges, I often found Amit to rise to the challenge with creative technical solutions that complimented business requirements. There simply is only one Amit.
"
Jay Khan
Delivery Manager at V&A Museum
"Amit has the right attitude for the new way of creating business solutions so the technology is not in the way of the experience. Amit is a one-off diamond.
"
Andrew Woolfson
Head of Knowledge Management
"Amit is a great guy both professionally and personally. Working with Amit was always an eye-opener because of the incredible knowledge and skill he brings to the table. I have worked with Amit on various occasions where I was lucky enough to witness some of the innovative skill and thinking that comes from combining a thorough sense for business and incredible skill in Programming and Architecting.
"
Director at Microsoft
"Amit is an excellent technical thinker who always manages to keep an eye on the human nature of what is being built; he can be relied on for sharp, relevant insights into both user centred process/design and technical decisions, and it is always a pleasure working with him, as well as spending time together socially away from the office.
"
Digital Lead at The Chancery Lane Project
"Amit is technical capabilities are combined with a strong business insight and understanding of user behaviours which makes him a great project collaborator. His focus is on getting the work done and problems solved. I always enjoyed working with Amit on development projects both for his ideas and his helpful positive attitude.
"
Co-Founder & CEO at Clouz
"Amit has a fabulous attitude and is a very capable technologist. I have always enjoyed working with Amit and watching him solve the world's problems one entrepreneurial idea at a time.
"
Delivery Partner at Hippo
"Amit offers an impressive background of technological and creative knowledge. I was introduced to Amit through a networked link and his recommendation proceeded his entrepreneurial-ship and effective working practice. His ideas are sharp and original, and provide clear insight for what will be the next big thing. I very much look forward to working with Amit and anticipate his success for the future.
"
Professor at Durham University Business School
"Amit has great vision and knowledge of web design and web usability. He has a real intuition and passion for usability in web applications.
"
Founder at Payara
"I recommend Amit highly. He is a very smart and dedicated person who is also a pleasure to work with.
"
Programme Manager at MultiChoice Group
"Amit is a truly unique and colourful person who will bring light into any group of people he encounters. His commitment to anything he turns his mind to amazed me and he is not only willing but loves to put himself in the firing line for any new challenges or experiences he can find.
"
Dave H
Advocate/Advisor at DIGTL Markets
"Amit is an open minded and very creative entrepreneur. He has very good communication skills and has always a pragmatic approach in getting things done.
"
Partner at IBM
"Amit worked with me on a paper researching Multi-Agent Systems. I found him very knowledgeable, reliable and committed. I regularly see Amit contributing to forum discussions and is always willing to help anyone above and beyond the call of duty.
"
Freelance Copywriter at The Word Booth
Claude Enterprise puts 25 settings sections behind one nav, grouped into four blocks that do not match how anybody actually governs AI. Here is the whole tree, what each block controls, and the six settings worth opening on your first day as an Owner rather than the twenty-five you will otherwise scroll past.
Anthropic now lets an Enterprise organisation put its own server in front of every prompt and return allow or deny before the model runs. It is the strongest inline control Claude has shipped. It also never receives raw image bytes, so a screenshot of the document you are trying to stop walks straight through, and the failure mode when your server goes down is a setting somebody has to choose.
The Enterprise console offers one toggle that keeps all model inference in US regions for a 10 per cent surcharge. Underneath it are two independent settings, one of which is a one-way door you set when you create a workspace and can never change. The surcharge is the part everybody reads and the least expensive of the three costs.
Three tools failed within a minute of each other with three unrelated errors, and the folder they blamed had ordinary permissions the whole time. macOS TCC had revoked it. Two tccutil resets fixed that with no restart, because Terminal and Claude Code hold separate TCC identities and each one needs its own grant.
Ask the Claude add-in in Excel to roll a workbook forward into next month and you get a permissions error, which reads like a misconfiguration and is not one. The add-in has no file management at all, and that limit is missing from the unsupported list. Here is what it does do well, why it beat Claude Code on the same spreadsheet, and the settings field that fixes the complaint everybody has.
Anthropic shipped write tools for the Microsoft 365 connector on July 7, 2026, and left them switched off for every organization that had connected before that date. Nothing in the product says so. Here is how to check in ten seconds, the two admin gates in two different consoles, the scope that deserves a second look, and the write path already running in your building that will derail the diagnosis.
Claude Code sessions can now message each other over a Unix socket on your own machine. I had 23 sessions running and only 14 of them could see each other. Here is what the message channel fixes, what it cannot fix, and why the guard I trusted for months was never a guard at all.
Adding an organization-wide instruction block does add tokens to every prompt, and that objection is correct. A controlled 180-call test on Claude Sonnet 5 shows one brevity line cutting output 23.9 percent and returning 14.2x its own cost, while the block carrying it fails to break even uncached.
A shared Claude conversation returns HTTP 200 whether or not you can read it, and the page carries six characters of visible text. Team and Enterprise chats are organization-only by design, attached files never travel with the snapshot, and the one programmatic route is a Compliance API that has no idea what a share link is.
Claude Code loads your CLAUDE.md into every subagent except two. Explore and Plan skip it by design, and no setting changes that. So the agents you fan out most widely are the ones that never see your rules. ETH Zurich measured what the file costs on the occasions it does load.
A git worktree isolates your working directory and your index. It does not isolate the stash stack, which is shared by every checkout in the repository. Here is what I measured on git 2.50.1, why MERGE_AUTOSTASH escapes the problem through an accident of spelling, and how one defensive stash can make a worktree look safe to delete.
Merge queues, merge trains, and speculative merging all rest on one assumption nobody states: that a per-branch green tick means the code works. Our own CI runs no unit tests at all, and master went red twice in two days from pull requests that were each green on their own. Here is what to build before you buy the queue.
Feeding the right background data into prompts so AI isn't guessing.
Making sure AI can grab what it needs, even when data's scattered.
Small model for grunt work, big model for thinking. Match the tool to the job.
Splitting big problems into tiny prompts. One thing at a time.
Building things that actually run in production, not slide decks.
When one prompt won't cut it - chaining dozens together.
Tenant restrictions, device trust, locked-down proxies, MCP allowlisting. The phase-0 work most teams miss.
Baking in checks so AI doesn't go off the rails.
Prompts and instructions that improve themselves automatically.
Most AI fails because the problem was badly defined. I've spent decades on this.
Sometimes a simple script beats a fancy prompt. Knowing when matters.
Running many prompts at once without them stepping on each other.
Tech changes are easy. Getting people on board is the hard part.
When should AI fire? On a schedule? On an event? On a hunch?
I work with everyone from college students to C-suite execs.
