Ten client stories from manufacturing, universities, insurance, engineering, research, dental care and a music nonprofit. Each one says what was there before, what we did, and what changed.
people trained hands-on, across companies and universities
25
AI agents and automations running in production
87%
lower hosting costs after one cloud migration
11 apps
brought under one security standard for a college
What we actually do, in plain words.
Most clients come to us for one of these and stay for several. Each one links to a real project where you can see it in action.
Train your people
Hands-on workshops, executive sessions, train-the-trainer programs and courses. Built around your tools and your policies, not generic slides.
See it with , ,
Advise your leaders
Where AI fits, which platform to choose, how to write the policy, and how to govern hundreds of users once the rollout starts.
See it with ,
Build from idea to product
You bring the idea. We write the specs, design the architecture, build it, secure it and hand it over running.
See it with ,
Take over and modernize large sites and apps
Inherited code, many vendors, rising cloud bills. We consolidate it, document it, and add an AI layer on top of what already works.
See it with ,
Secure and audit
Sign-in, credentials, spending limits and access control brought to one standard, and an honest review of how your systems and people use AI, with fixes ranked by risk.
See it with , ,
Automate marketing, outreach and operations
Content pipelines, prospecting campaigns that keep a person in charge of what gets sent, and manual work like onboarding and procurement scanning replaced by tools your team trusts.
See it with , ,
Ten projects, one table.
Read across a row to see how much ground a single client relationship covers. Select any client to open the full story.
Clients and the services Allemnii delivered for each
Client
Train
Advise
Build
Secure and audit
Automate
No
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
No
Yes
Yes
No
Yes
Yes
Yes
Yes
No
No
No
Yes
No
Yes
No
Yes
Yes
Yes
No
No
No
Yes
No
Yes
No
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
No
A filled circle means we delivered that service for the client.
The case studies.
Filter by what we did or by industry. Every result below comes from delivered work, and we only publish numbers we can stand behind.
What we did
Industry
Showing 10 of 10 case studies.
Antique Phonograph SocietyMusic and nonprofit
A sprawling cloud estate rebuilt at an eighth of the cost
A music nonprofit's website, member portal, forum and research library moved from 104 Azure resources to one Google Cloud project the society owns, with a new site and admin portal built on top.
BuildSecure and auditAdvise
Before
$1,300
per month, 104 Azure resources
After
$170
per month, one project they own
Rollins CollegeHigher education
Eleven AI apps, one security standard, one data layer
We took over a portfolio of AI tools built by different hands, brought every one to the same production standard, and kept building new AI apps on top.
BuildSecure and auditAutomateAdvise
Before
11
apps, 11 different rulebooks
After
1
security standard for all
Global insurance companyInsurance
A custom AI coach for Lean Six Sigma, approved by enterprise security
We turned a client's idea into an AI coaching platform that guides operations teams through every phase of a Lean Six Sigma project, and took it through a full enterprise security review.
BuildTrainSecure and audit
Before
Version 1
coaching platform
After
Version 2.0
security reviewed
Construction companyConstruction and engineering
One AI policy, two tiers of training, and governance for a large rollout
A construction company now has a company-wide AI policy, training built from that policy, a system to approve and manage AI skills across the firm, and an automated scanner for public procurement opportunities.
AdviseTrainAutomateSecure and audit
Before
No policy
for AI use
After
1 policy
one approval path for AI tools
Business ConnectManufacturing
A multi-track AI program for a manufacturer
With Davenport University's IPEx, we delivered a customized AI program for a Grand Rapids manufacturer, with separate tracks for the whole team, marketing, supply chain and the executive team.
TrainAdvise
Before
One team
many different AI needs
After
5 tracks
each built for one role
Davenport University, IPExHigher education
Trained the trainers, then automated marketing and outreach
For Davenport University's Institute for Professional Excellence we trained its facilitators to teach AI, automated its marketing channels, built a custom AI agent and set its team up with an outreach platform.
TrainAutomateBuild
Before
Manual
marketing and outreach
After
Automated
channels and campaigns
Rollins College, Crummer Graduate School of BusinessHigher education
A Mini MBA AI program, delivered twice
We teach the AI module of the Mini MBA at the Crummer Graduate School of Business, delivered once as a standard program and once fully customized for a client organization, plus staff workshops and platform advice.
TrainAdvise
Before
Little to none
prior AI experience
After
Hands-on
AI in daily work
University of OklahomaHigher education
Cohort onboarding, from 16 manual steps to one upload
An executive academy got a custom AI readiness assessment, faculty training, and an onboarding tool that turns a participant roster into personalized letters and welcome emails in minutes.
BuildAutomateTrain
Before
16 steps
by hand for every cohort
After
1 upload
letters and emails ready
AxelysResearch
A French-language AI portal analysts keep using
Invention and patent analysts at a Quebec technology transfer organization got French-language AI training and a learning portal built around their daily work.
TrainBuildAdvise
Before
Training
ended with the session
After
Reference portal
still in use
OraSurgDental and healthcare
A secure patient intake app for dental practices
We built a multi-practice intake app that collects clinical intake forms and insurance cards, and audited it end to end before launch because it handles patient health information.
BuildSecure and audit
Before
Paper
forms and copied cards
After
Digital
intake, audited before launch
Antique Phonograph Society, Music and nonprofit
A sprawling cloud estate rebuilt at an eighth of the cost
A music nonprofit's website, member portal, forum and research library moved from 104 Azure resources to one Google Cloud project the society owns, with a new site and admin portal built on top.
Before
104 Azure resources across several vendors, about $1,300 a month, and a developer needed for routine content changes.
After
One Google Cloud project at about $170 a month, a new website, and staff updating content themselves through their own admin portal.
The situation
The society runs a public site, member portal, discussion forum and research library for collectors of early recorded sound. Over the years its hosting had grown to 104 resources across more than one vendor. The board wanted lower running costs, staff-managed content, and every member and payment record moved without loss.
What we did
Migrated the database, about 70 GB of files and the forum server from Azure to Google Cloud, consolidating 104 resources into one project.
Moved roughly 900 member records, several thousand subscriptions and the full order history intact.
Rebuilt the website with memberships, PayPal donations, a rotating homepage showcase and PhonoWizard, a guided feature for the society's members.
Rebuilt sign-in on Google Identity Platform, corrected permissions on about 500 admin-flagged accounts and carried no legacy passwords across.
Kept security in place throughout: web application firewall, HTTPS everywhere, credentials in Secret Manager and automatic certificate renewal.
Built a 26-screen admin portal, wrote the infrastructure runbook and handed over full ownership.
Results
Hosting cost down about 87%, from roughly $1,300 to $170 a month
83 tracked items delivered in about six weeks by a team of four
Member, subscription and payment records migrated intact
Automated build and deploy pipeline in place of manual releases
Inheriting someone else's infrastructure usually means sprawl, unclear ownership and costs nobody has questioned. We consolidated, secured and re-costed a live system without breaking it, and left the documentation behind so the client can run it.
Rollins College, Higher education
Eleven AI apps, one security standard, one data layer
We took over a portfolio of AI tools built by different hands, brought every one to the same production standard, and kept building new AI apps on top.
Before
Provider keys in configuration files, paid endpoints open to anyone with the address, and a code release every time staff access changed.
After
Verified sign-in, usage ceilings on paid routes, managed cloud credentials instead of stored keys, and access lists staff edit themselves.
The situation
The institute had built AI tools for research, classroom simulations, proposal drafting and lead capture. Most were prototypes or inherited code, each on its own framework and hosting. There was no shared standard for sign-in, keys or spending limits, and CRM data sat outside any place an AI agent could safely use it.
What we did
Reviewed every system's dependencies, endpoints, sign-in paths and database access, and fixed findings at the root rather than patching around them.
Standardized passwordless sign-in verified on the server, and replaced stored AI keys with managed cloud identity on least-privilege accounts.
Migrated and hardened the public website on Firebase App Hosting with rate limiting, encryption with client-managed keys, and separate website and data projects.
Built a governed BigQuery data layer fed by scheduled HubSpot snapshots, ready for an analytics agent to query trusted data.
Built new AI apps, including a daily briefing that scans 45+ news sources and scores each story against the institute's priorities, a voice-to-CRM tool where a person approves every record, a simulation builder that turns a plain-language scenario into a playable exercise, an RFP builder that flags missing dates and signatures, and a risk reviewer for vendor contracts that cites evidence for every finding.
Results
11 applications, a public website and a data warehouse on one production standard
22 classes of security finding closed across the systems reviewed in depth
Critical dependency advisories cleared, with one app going from 18 advisories to 5
8 applications consolidated onto a single AI provider
Staff now manage access themselves, without a developer
When apps are built separately, each drifts into its own way of handling access, keys and cost. We put one standard across all of them without forcing a rebuild, and control now sits with the client instead of whoever wrote each app.
Global insurance company, Insurance
A custom AI coach for Lean Six Sigma, approved by enterprise security
We turned a client's idea into an AI coaching platform that guides operations teams through every phase of a Lean Six Sigma project, and took it through a full enterprise security review.
Before
An early coaching platform, and process improvement coaching that depended on the time of a few experienced coaches.
After
An AI coach on Google Vertex AI and Gemini that walks each person through Define, Measure, Analyze, Improve and Control, cleared by the client's security team.
The situation
A global insurance company brought us in on two fronts: AI training content for its annual company retreat, and a coaching platform for its operations team. The platform had to meet enterprise security, privacy and procurement standards before anyone could use it.
What we did
Delivered AI training content at the company's annual retreat.
Built version 2.0 of the coaching platform: the coach guides a learner through each DMAIC phase, runs the statistical tools and explains the results in plain language.
Built it on Google Vertex AI and Gemini, hosted on Google Cloud with encrypted user interactions.
Moved the platform from stored API keys to managed Google credentials.
Completed the vendor security questionnaire and a conversation isolation review with the client's security team, so no user can see another's coaching history.
Results
Custom coaching platform built for the operations team
Full security and compliance review completed with the client's internal team
Training content delivered at the annual retreat
Statistical toolset now being expanded for the next release
Stack
Google CloudVertex AIGeminiEncrypted storageManaged credentials
Why it matters
A tool for an operations team has to clear the company's security review before anyone uses it. This one was designed for that review from day one, which is why it went from idea to approved platform.
Construction company, Construction and engineering
One AI policy, two tiers of training, and governance for a large rollout
A construction company now has a company-wide AI policy, training built from that policy, a system to approve and manage AI skills across the firm, and an automated scanner for public procurement opportunities.
Before
No written AI policy, no shared training, and no single way to approve or retire AI tools as the firm adopted them.
After
A written AI usage policy, literacy and practitioner training based on it, and one governance process for every AI skill in use.
The situation
The firm needed a company-wide AI usage policy its people could follow, training that taught those real rules, and a way to keep control as a large AI rollout grew. The work started in 2026 and has become an ongoing retainer.
What we did
Wrote the AI usage policy with the leadership team, reviewing and refining it together.
Designed and delivered two tiers of training: AI literacy for everyone and deeper practitioner training, both built from the firm's own policy.
Built the skill governance system for the firm's AI rollout: owners, moderators, an inventory of every skill, checklists and handoff documents.
Built RFP Radar, an automated workflow that scans procurement sites, removes duplicates, scores each opportunity and sends a curated report.
Ran a live AI discovery workshop and plant tour for one of the firm's own clients.
Results
Company-wide AI usage policy in place
Two-tier training built from that policy
One system to catalog, approve and retire AI skills firm-wide
Procurement opportunities scanned and scored automatically
Grown into an ongoing retainer
Stack
Governance playbooksWorkflow automation
Why it matters
The training used the firm's own policy as source material, so people learned their real rules, not generic content. Governance came before scale, not after something went wrong.
Business Connect, Manufacturing
A multi-track AI program for a manufacturer
With Davenport University's IPEx, we delivered a customized AI program for a Grand Rapids manufacturer, with separate tracks for the whole team, marketing, supply chain and the executive team.
Before
Different teams needed very different things from AI, and one generic session could not serve the shop floor, marketing and leadership at once.
After
Each group trained on the AI uses that matter in its own job, with a shared foundation in responsible use.
“The trainer really heard and listened to the people in the room. It was people first the entire time.”
Jereme LambertCEO, Business Connect
The situation
Business Connect, a manufacturer in Grand Rapids, Michigan, wanted its people to use AI in their actual jobs. We designed the program with Davenport University's Institute for Professional Excellence so each group got training it could apply the next morning.
What we did
Delivered foundational AI literacy for the whole team.
Ran a marketing training track on content, campaigns and research.
Ran a supply chain optimization track built around the company's own processes.
Led an executive strategy session on where AI should and should not go.
Closed with responsible AI and ethics, so every track shared the same guardrails.
Results
Five training tracks delivered, each customized to a role
Executive team aligned on AI priorities
Delivered in partnership with Davenport University's IPEx
Stack
Microsoft CopilotChatGPTClaudeRole-based exercises
Why it matters
Manufacturers do not need AI theory. They need people in marketing, supply chain and leadership who each know what to do with it on Monday.
Davenport University, IPEx, Higher education
Trained the trainers, then automated marketing and outreach
For Davenport University's Institute for Professional Excellence we trained its facilitators to teach AI, automated its marketing channels, built a custom AI agent and set its team up with an outreach platform.
Before
Facilitators new to teaching AI, marketing channels run by hand, and prospecting done one email at a time.
After
Facilitators teaching AI to IPEx clients, automated marketing channels, a custom agent, and outreach campaigns with clear reporting.
The situation
IPEx delivers professional education to companies across Michigan. It wanted its own people to teach AI, use it daily, and grow its client base without adding staff.
What we did
Delivered a train-the-trainer program so IPEx facilitators can teach AI to their own clients.
Trained the internal team to apply AI in daily operations.
Automated IPEx's marketing channels.
Built a custom AI agent through a three-phase process: requirements discovery, data source review and an implementation demo.
Set up an outreach platform with ideal customer profiles, campaigns and branded lead generation reports.
Wrote the content for an ethical use of AI e-learning module for educators, and co-facilitated an AI readiness workshop at Tech Week Grand Rapids.
Results
IPEx facilitators now deliver AI training to their own clients
Marketing channels automated
Custom AI agent delivered
Outreach campaigns running with branded reporting
Joint programs delivered to IPEx clients, including Business Connect
Stack
Custom AI agentOutreach platformMarketing automationArticulate Rise (with design partner)
Why it matters
The best training partner leaves you able to teach it yourself. This relationship started with training and now covers marketing, outreach and joint delivery to companies across Michigan.
Rollins College, Crummer Graduate School of Business, Higher education
A Mini MBA AI program, delivered twice
We teach the AI module of the Mini MBA at the Crummer Graduate School of Business, delivered once as a standard program and once fully customized for a client organization, plus staff workshops and platform advice.
Before
Participants with little or no prior AI experience, and staff unsure which AI platform the school should adopt.
After
Participants practicing prompting, research, agents and data analysis on real tasks, and a platform choice backed by clear evaluation questions.
The situation
The business school wanted a practical AI program for executives starting from zero. It ran the program as a standard offering, then again customized for one of its client organizations. In parallel, the school's staff wanted hands-on training and help choosing an AI platform.
What we did
Taught Strategic AI: Scaling Operations with Ethical Guardrails and Leading with AI in the Mini MBA.
Built six live demos: thinking mode, canvas, deep research, agents, data analysis, and image and video.
Delivered a second run fully customized for a client organization.
Ran an AI staff workshop and a lunch and learn on AI agents.
Led AI platform selection with structured discovery questions.
Results
Program delivered twice, once fully customized
Prompt library and practice decks participants keep after the session
Staff trained on AI agents and daily use
Relationship continued as an ongoing AI advisory retainer
Stack
Microsoft CopilotChatGPTGeminiClaude
Why it matters
Most participants started with little or no AI experience. The program was paced for them and built on doing, not watching.
University of Oklahoma, Higher education
Cohort onboarding, from 16 manual steps to one upload
An executive academy got a custom AI readiness assessment, faculty training, and an onboarding tool that turns a participant roster into personalized letters and welcome emails in minutes.
Before
A 16-step manual process for every new cohort: assessment letters, single-use passwords and welcome emails assembled one participant at a time.
After
An admin uploads the roster, picks the cohort, and gets each participant's personalized PDF letter and a ready-to-send email draft, with a full audit log.
The situation
The academy runs leadership cohorts where every participant receives a personalized assessment letter, a single-use password and a welcome email. Doing this by hand took hours per cohort and invited mistakes.
What we did
Built a custom AI Readiness Assessment for the academy.
Trained faculty on AI adoption and applications.
Built the onboarding tool: sign in, upload the master workbook, choose leadership level and cohort.
The tool reads every participant, generates a personalized PDF letter for each, and prepares one welcome email per person as a draft with attachments in place.
Returning participants keep their existing password, so none are wasted, and every action is logged.
Results
16-step manual process replaced by one upload
Onboarding time cut from hours to minutes per cohort
Every letter and email traceable through an audit log
Partnership extended to the academy's own executive clients
Stack
Google CloudGoogle Drive APIGmail APIPDF generationAudit logging
Why it matters
A team should spend cohort week with participants, not with spreadsheets. The tool removed the busywork and kept a person approving every email before it goes out.
Axelys, Research
A French-language AI portal analysts keep using
Invention and patent analysts at a Quebec technology transfer organization got French-language AI training and a learning portal built around their daily work.
Before
Analysts were interested in AI but had no shared guide on which tools to use, or on which data was safe to put into them.
After
A French portal with a platform guide, a data-security router tied to the AI usage policy, step-by-step recipes and a 23-prompt library.
The situation
Axelys wanted its invention and patent analysts to use AI in real work. The training had to be in French, follow the organization's own AI usage policy, and stay useful long after the session ended.
What we did
Delivered a generative AI training in French, with participant and trainer guides and live demonstrations.
Built a French-language learning portal around the daily workflow of an invention and patent analyst.
Added a platform-selection guide for choosing the right AI tool.
Built a data-security router aligned to the AI usage policy, so analysts know what data can go where.
Wrote step-by-step recipes and a 23-prompt library covering researcher meetings, patent search and final reports.
Results
A standing reference tool the team keeps using after training
Training, guides and tools delivered fully in French
23 prompts covering the analyst workflow end to end
Stack
Web learning portalPolicy-based data routerPrompt library
Why it matters
Most AI training ends when the session does. This portal stays open on analysts' screens because it was built around their own work and rules.
OraSurg, Dental and healthcare
A secure patient intake app for dental practices
We built a multi-practice intake app that collects clinical intake forms and insurance cards, and audited it end to end before launch because it handles patient health information.
Before
Patient intake relied on forms and manual copying of insurance cards at the front desk.
After
Patients complete intake on their own device, documents and insurance cards are captured once, and each practice's data stays separate.
The situation
A dental surgery group wanted to modernize patient intake across practices. Because the app handles patient health information, security had to be part of the build from the first line of code.
What we did
Designed and built a multi-tenant intake app, with each practice's data isolated from the others.
Added capture for clinical intake documents and insurance cards.
Ran a full security audit of the application before launch and closed the findings.
Set a rule that all demos and testing use synthetic data, never real patient records.
Results
Digital intake in place of paper forms
Practice data isolated by design
Full security audit completed before launch
Stack
Multi-tenant web appDocument captureEncrypted storage
Why it matters
In healthcare, a fast app that leaks data is a failure. We treated the security audit as part of delivery, not a box to tick afterwards.
real results
"The trainer really heard and listened to the people in the room. It was people first the entire time."
Jereme LambertCEO, Business Connect
We also teach where we train.
Our team teaches inside university programs and speaks at corporate and university events. The classroom keeps our training honest: if a method does not hold up in front of students and executives, it does not reach your team.
Rollins College
Mini MBA AI program at the Crummer Graduate School of Business, staff workshops, and AI platform advice for the college.
Davenport University
Train-the-trainer program for the Institute for Professional Excellence (IPEx), plus programs delivered to its corporate clients.
University of Oklahoma
AI readiness, faculty training and automation for an executive academy.
McGill University
Teaching and talks on applied AI for students and professionals.
We speak at corporate retreats, executive briefings and university events, including Tech Week Grand Rapids, where we co-facilitated an AI readiness workshop with Davenport University's IPEx.
Every person on our team is an AI engineer.
From founders to engineers
Our whole team holds Masters and PhD degrees. The person who scopes your project is an engineer, not a salesperson.
Founding team from Google and IBM
Before Allemnii, our founders implemented AI and cloud systems for customers across North America.
Built on Google Cloud
Firebase, Cloud Run, Firestore and Vertex AI are our home stack, and we work across Microsoft Copilot, Claude and other platforms our clients already use.
Security first
Every build goes through a security review before launch, and we audit the AI practices of systems built by others.
Bring us the project you have been putting off.
A large website that needs an AI layer, a team that needs training, a policy nobody has written yet, or an idea that needs to become a working product. A 30 minute scoping call shows where to start.
