AM Consulting · Selected work
Your teams start working with AI tomorrow, and the system you need goes live in days, not quarters. For most of our clients, the first result was live within the first week.
Companies we have worked with and in
Timeline
Every dot is something we delivered to a client during 2026. Most went live within a week, so you never wait months to find out whether it works. Hover over a dot to see what was delivered.
Confidential work appears as a dot without a link.
Projects
You get a complete system, not just a model: from the data to the decision, including deployment, tests and a proper handover to your team. Here is how that looked for manufacturers, enterprises and startups we worked with.
Danpal · facade panel manufacturer
At Danpal, every 2D production drawing was prepared by hand from a 3D assembly. Now an AI agent does it inside SolidWorks: it produces a drawing, compares it with the reference, fixes it and tries again until it is right. The draftsman gets the drawing in one click from the browser, and is free for work that needs a person.
Plasgad · manufacturing
Plasgad wanted AI visual quality checks. Before they ordered hardware, we ran a 674 call experiment for them, for under a dollar. It showed a 2 megapixel camera is enough: 78 to 80 percent accuracy, and not a single good part rejected by mistake. The 20 megapixel camera on the quote simply was not needed.
Focus · financial planning and wealth management
At Focus, preparing a client proposal was manual work buried in documents. Now they have a guided process from scoping to the finished proposal, so the team spends its time on the client, not on paperwork.
Precise / Lawyal · legal tech
The engine matches every bank transaction to the right invoice and leaves a person only the final approval. A working prototype was ready after 4 working days. We then built it as an engine on AWS inside the Lawyal platform, together with the client's developers, so the knowledge stays with them.
STC · engineering
Instead of reading engineering drawings and copying data out by hand, the system reads them with Claude and returns structured data. All of the infrastructure is defined as code, and the first version was live in less than a week.
Ready Group · software and hardware R&D · advisory
Since 2024 Ready Group has had an outsourced Chief AI Officer from us, who leads their AI team and takes part in the big decisions. Together we planned a three year, $10M AI research lab, from the business model all the way to whether the building's power could carry the servers.
Training
Your teams leave the workshop with an automation that works on your own tools and data, and the file that drives it. Something they use on Monday morning, not another slide deck.
AT&T · 4 programmes
Architects, product managers, project managers and engineers each got a programme built around their daily work. The sites for architects and product managers went live on consecutive days, and the architects' site has a live simulator that shows in advance what routing between models will cost and how fast it will respond.
Salesforce · 3 programmes
The assistants got a four session programme with automations for calendars, travel, summaries and team communication. The developers got five short two hour courses and a workshop on large codebases, with a practice environment that comes up in one command.
AppsFlyer · repeat client
After the first workshop the analysts pull data on their own, and the data team gets far fewer requests. In the second one they build a churn model and a revenue model on a synthetic copy of their own data, and keep about 120 lines of Python they run themselves every month.
SysAid · repeat client
In a single build day, opened by the CMO, everyone on the marketing team built a tool for their own daily work. Then it was the turn of the HR and Finance teams, with the same method.
TraceSpan · hardware engineering
An FPGA team got an eight week plan for bringing in AI, plus a practical proof: a decoder for the IEEE 802.3ah standard, written by AI and checked three independent ways. The plan and the proof were ready within 48 hours.
Our products
When we build for you, you get a team that has already been all the way: from an idea, through real users, to a product running in production.
Dibra · dibra.one
A caption studio where every word can be edited, with Hebrew captions that display correctly from right to left. We tested 12 competitors, and none of them does this. Every model choice was made against transcripts produced by people.
Upshot
Records, transcribes and summarises meetings on your own computer, with no bot joining the call. Built for Hebrew first, available in five languages.
Afterprompt · open source
Scans the history of 22 AI coding tools and finds API keys and passwords left behind. Runs locally on macOS, Linux and Windows.
Deck engine · internal tool
A framework for branded interactive decks, with one command that creates the repository, sets up hosting and goes live. Eight client decks came out of it in the first weeks, and one went live 13 seconds after its first commit.
How it works
An intro call, then we learn your systems, data and constraints. The proposal you get fits you, not a template.
A written plan: what you get, when, and how we will both know it worked.
A brand system, a deck engine, workshop templates and a library of Claude Code skills. Everything repetitive is already done, so your time and budget go into your problem.
Within the first week you see the real thing, not a mockup, and you can steer it as we go until handover.
About
Founder and CEO, AM Consulting
For more than twelve years Avishay has been putting AI into production in places where every mistake costs money. At PayPal he built fraud detection models that protect hundreds of millions of customers, and led the data science team behind a $2B credit portfolio. He then headed data at Lili, and data and AI at the digital bank ONE ZERO. In 2023 he founded AM Consulting, and he brings that experience to every project: systems that have to work correctly, even when they are built fast.
US10866995 US11507631 US12271870US11182795 US12469035US11443224US11249965 US12259860US10762103 US11609929US11422983US10706050US10936677 US11475079US11468129MBA in Econometrics and Quantitative Economics, The Open University of Israel · Graduate studies in Computer Science, The Hebrew University of Jerusalem · Talks: the Israel IT Conference, DataTalks meetup.
What clients say
“Avishay expertly advised us through the entire process of designing a $10M, 3 year plan for establishing an AI research lab. From planning a detailed and viable business model to ensuring the building's electrical infrastructure could support servers loads, Avishay demonstrated exceptional expertise and attention to detail.”
“Leading and developing AI solutions demands a rare combination of strategic vision and professional expertise. Fortunately, with AM Consulting by my side, I'm able to leverage this powerful blend to my advantage.”
Let's talk
You get a written proposal within 48 hours. Training for your teams, or a system we build for you.