Who they are
Juice Plus+ is one of direct selling's enduring names: founded in 1970 with a mission to inspire healthy living around the world, and best known for its whole-food-based nutrition — the Fruit, Vegetable and Berry capsules that bring the nutrition of thirty plants to a daily routine. Its products are recommended and sold by a global field of independent Partners.
The business runs in 24 countries — nine storefronts, many languages and payment cultures, and one global compensation plan across all of it.
The challenge
Juice Plus+ ran for more than twenty years on a platform built in-house — and years of under-investment had caught up with it. Technical debt had accumulated to the point where the platform no longer met compliance standards, and the technology leadership of the day concluded it had to be replaced outright, on a hard deadline: a new platform, live by September 2024, built in nine months. A leading global consulting firm and an offshore development partner were engaged to deliver it, on an off-the-shelf core — Shopify storefronts, with a packaged commission engine behind them. The complexity of the Juice Plus+ model had to be absorbed somewhere: nine production storefronts serving 24 countries, against a single global compensation plan.
The launch went badly wrong — worst of all for the field. Partners lost functionality they had relied on for years. They lost visibility: the reports they ran their businesses on no longer existed. They lost control: customers could no longer be managed by their Partner, and were left to self-serve in a storefront. And they lost attribution — the link between customer and Partner that the entire earnings model depends on needed significant work.
The subscription model broke hardest. Juice Plus+ ships product in four-month supplies and charges customers monthly — and the new platform mis-handled nearly every part of that: duplicate shipments, installment charges that were never collected, customers overcharged, subscriptions auto-cancelled without notifying the customer or their Partner, and subscriptions stuck in states from which they would never process again. Customers were lost, orders were lost, and Partners could not see any of it happening. The storefronts themselves confused customers — especially in Europe, where direct debit remains a primary way to pay.
The commercial impact on the company and its field was substantial. But the deeper damage was trust: a significant portion of the field lost faith in the platform and left the business for easier opportunities. By the February 2025 Board meeting, the Board was not satisfied the situation was under control — and reached out to AHM Technology for an independent evaluation.
What we did
AHM Technology engaged in March 2025 for a two-month assessment. Within the first month the picture was clear: the situation was dire, and two failures mattered above all — attribution, and subscription handling. The findings were shared with the technology leadership, CEO and Board.
The engagement escalated from there. First, a request to work alongside the existing technology team on the most critical fixes, which were in place by early summer. Then, when the technology leadership of the day chose to leave the business, a request to take ownership of technology entirely — the organization, the platform, and the plan to fix it. From August 2025, the rebuild was run with AI at its core: Anthropic's Claude in the hands of a much smaller, much more senior engineering team.
One thing this engagement is not: a platform license. The solution is Juice Plus+'s own — a dedicated platform, built as professional services, owned and operated by the company, on its own codebase. It is informed by everything AHM Technology has learned building direct sales technology — but the platform is Juice Plus+’s: a dedicated platform, built for them.
Independent Assessment
Engaged by the Board in March 2025 to establish what had actually happened, what was recoverable, and what needed replacing. The assessment was scoped at two months; the severity was evident within one. Findings were shared with the technology leadership, CEO and Board, with two failures prioritized above everything else: attribution, and subscription handling.
The true state established in weeks — and attribution and subscriptions named first
Fix the Worst, Fast
From April, AHM Technology worked alongside the incumbent technology team, adding a small senior unit — a product manager, an architect and an engineer. The focus was disciplined: understand the business requirements the field actually had, restore trustworthy data in the business intelligence system while internal reporting was broken, fix attribution with a specialist agency, and fix the subscription logic with internal resources.
The subscription rescue shows how the work was run. The failures were invisible by design — no logging, no alerts, no monitoring: a subscription would fail silently and be forgotten. So the first thing built was visibility. A healthy subscription was defined precisely, and every subscription that failed to meet that definition was found.
Each failure was then investigated and classified into buckets of issue types, traced to root cause with the engineers, and fixed — prioritized by customer harm and frequency: anything that charged a customer incorrectly first, product shipped too early second, and so on down the list. Many issue types took multiple rounds of fixes; the implementation underneath was that poor. Two permanent changes went in alongside: detailed logging, so every failure could explain itself against the healthy-subscription definition — and automatic reprocessing, so a failed subscription would retry instead of dying silently.
The critical failures were resolved by early summer 2025.
Every failing subscription found, classified, root-caused and fixed — worst harm first
Take Ownership of Technology
With the critical fires out, the existing technology leadership chose to leave the business — and Juice Plus+ asked AHM Technology to step in and run technology. From August 2025 the approach changed: trusted development agencies were brought in, engineers who had genuinely adopted AI — Claude in their daily practice — joined the team, and the plan was rebuilt around the critical items holding back the field and impacting customers.
A delivery organization rebuilt around AI
Rebuild the Back Office First
The Partner back office was not doing what the field needed, so the rebuild started there: the reports Partners run their businesses on, customer management returned to the Partner, and visibility into everything the platform had been hiding from them.
The field's tools rebuilt around what Partners actually do
Fix the Commerce Experience
In parallel, the customer experience was reworked: simpler, more intuitive, and content-rich — product detail, reviews and the conversion fundamentals the storefronts launched without. The early work was about simplifying and standardizing the experience — and some of the problems were e-commerce fundamentals. Customers were being asked to choose their payment method on the product page, before they had even reached the checkout, where every online shopper expects that choice to live. Errors at that level, left behind by the original implementation, had to be corrected before anything more ambitious.
Commerce fundamentals corrected — and a storefront customers understand
Transform the Technology Organization
The rebuild needed a different organization, not just a different roster. Structures were changed and roles redefined; where individuals were not right for what came next they were replaced, and strong new talent was brought in. Just as much effort went into the people who stayed — coached to take control and drive the change rather than watch it, empowered to be accountable, and expected to own the success of the transition: fixing the platform for Partners and customers was treated as a serious obligation, not a project.
The contractor bench was the largest structural change. The inherited model — a large bench of contractors spread across the platform — was rotated out deliberately from September 2025 to March 2026, with contractors holding critical knowledge retained until their replacements were built, and replaced by the smaller, senior, AI-equipped team.
The result was a 70% reduction in contractors, with output more than doubled and no delivery dates missed.
A smaller, stronger, accountable organization — 70% fewer contractors, double the output
Bring the Cost Base Under Control
The platform's costs were out of proportion to its architecture: a substantial annual cloud spend on top of two SaaS platforms, plus multi-year commitments to software the platform could not use.
Costs were re-architected alongside the platform, cutting the technology operating budget by 30% — multi-million cost savings identified across cloud computing, software licenses, contractors and other areas.
Technology operating budget down 30%
The Day-One Features, Finally
With the core stable by February 2026, the work moved to what should have existed at launch: the back-office reports, single sign-on, smooth Partner onboarding, guest checkout, share-a-cart and more — delivered through June 2026. More recently, Google Pay, Apple Pay and PayPal have been added to the storefronts.
The platform the field should have been given at launch
AI Transformation — Reimagining the Partner Experience
With the platform stable and the day-one features delivered, the work changed character: from fixing what was broken to reimagining what running a Juice Plus+ business should feel like when AI is native to the platform. The answer being delivered is simple to state and radical in practice: a Partner should never have to mine a report again.
At the center is an AI assistant built into the Partner back office and exposed in the mobile app — where Partners actually are. Every morning it delivers a daily digest: what happened in their business overnight, what is going well, and what needs their attention today. Through the day it surfaces real-time updates the moment something needs action — a subscription about to fail, a customer whose order pattern has broken, a team member going quiet. The Partner does not go looking for any of this. It comes to them, on their phone, in plain language.
Beyond the Partner's own business, AI now works the front of the funnel too. An AI prospecting assistant engages prospective customers the moment they show interest — any hour, any time zone — answering their product questions naturally and in depth. When the conversation ends, the Partner receives a summary of who the prospect is, what they asked and what they care about, so the human follow-up starts warm instead of cold. The assistant never replaces the Partner; it makes sure the customer learns about the products, recommends what is right for them, and even brings them to the product pages so they can add to cart and purchase — while keeping the Partner informed of the engagement throughout.
Customers get their own AI on the storefronts: a commerce chatbot that helps them understand the products, ask questions in their own words, interrogate the clinical research behind the range, find the right information videos — and, when they want a human, connect them with a Partner. The things a great Partner does in person, available to every visitor, at any hour.
And through all of it runs the simplest capability of all: a Partner can just ask. The reports still exist — but increasingly they are the archive, not the interface. In plain language, on a phone, a Partner asks things like:
- “Who should I follow up with today, and why?”
- “Which of my customers’ subscriptions need attention this week?”
- “Who usually reorders around now but hasn’t?”
- “How is my team tracking this month, and who is close to their next milestone?”
- “Which customers have I not spoken to in sixty days?”
- “What changed in my business this week that I should know about?”
A Partner’s business, run from a phone — with AI doing the mining
What began as a two-month assessment became a request to fix things — and has ultimately become the reimagining of how a Partner runs their business, more productively.
AI-powered innovation
AI was not a feature added to this program. It is how the program was delivered — and it is becoming how the business thinks.
Claude Code Across All Development
The entire replatforming has been built with AI. Anthropic’s Claude Code is adopted across all development — every engineer, every day, on every part of the platform — with senior architects checkpointing what it produces and holding it to the design.
This is not a productivity tool at the edge of the process; it is the delivery model. It is what allowed a team 70% smaller to more than double the output of the model it replaced, deliver the majority of a full platform rebuild in nine months, and do it without a significant incident or a missed date.
AI Over Business Intelligence — Answers for Leadership
The business runs on Snowflake and Power BI, and that investment stays. What changes is who can use it: a natural-language layer sits on top, so leadership can ask questions of the company’s own data directly and get back not just numbers but detailed analysis: what happened, why, and what it suggests doing.
A leader asks how subscription retention moved since the fixes, which markets are ahead of plan and what is driving the gap, or where Partner activity is trending — and refines from the answer, instead of commissioning another dashboard and waiting. The reporting queue becomes a conversation.
Results & impact
A failed transformation reversed, a platform rebuilt, an organization restructured, and the field's confidence won back — with the majority delivered in nine months.