Most AI stories are demos. This one is production. Foundry Works runs agent teams in live operation for two named enterprise manufacturers, Lenovo and Brother, handling real buyer conversations and real channel support today. This case study goes inside both deployments: the challenge each client faced, the agent architecture built to solve it, and the outcomes it produces. Every claim here traces to what is published on the Foundry Works site. Where a specific number would strengthen the page but is not on record, it is left as qualitative proof rather than invented.

What Foundry Works built for Lenovo and Brother

A live agent deployment means an agent team that is in production and doing defined work, not a pilot or a demo. For Lenovo, Foundry Works built a three-agent architecture around the buyer journey: it greets buyers with a question rather than a form, delivers tailored product intelligence in real time, and hands sales warm, pre-qualified opportunities with full conversation context. For Brother, Foundry Works deployed an agent team directly into the reseller channel, giving more than 5,000 partners and their customers expert-level product guidance on demand, at the point of need, without adding headcount. Both are described on the Foundry Works site as live and operating now. Lenovo is a global technology manufacturer running the system across direct and partner channels; Brother International runs a custom deployment delivering measurable efficiency gains. These are production systems, not experiments.

Key takeaways

  • Foundry Works runs agent teams in live production for two named enterprise manufacturers, Lenovo and Brother, described on the site as live and operating now rather than as pilots.
  • The Lenovo deployment uses a three-agent architecture (Guide, Expert, Closer) that removes friction at first contact, delivers product intelligence in real time, and hands sales warm, pre-qualified leads with full context.
  • The Lenovo system delivers dramatically lower cost per qualified lead versus gated content, with always-on coverage across time zones and markets, per the published results.
  • The Brother deployment gives a channel of more than 5,000 resellers consistent, expert-level product support at the point of sale, and scales through volume spikes with no additional cost or resource.
  • Both deployments prove the same point: agent architecture that ships beats a demo, because the value is not conversation, it is work done in production.

Lenovo: three agents around the buyer journey

Lenovo is a global technology manufacturer, and the brief was B2B lead generation at scale. The deployment is described on the Foundry Works site as live now.

The challenge

Lenovo needed to scale product education and lead capture across direct and partner channels without scaling headcount. Traditional gated assets and contact forms were producing high bounce rates and low-quality leads that required extensive sales development time to qualify and progress. In other words, the old funnel spent human hours turning cold form-fills into anything worth a conversation.

The approach

Foundry Works built a three-agent architecture designed around the buyer journey. It removes friction at first contact, delivers personalised product intelligence in real time, and captures a lead only after genuine value has been exchanged. The result is warm, pre-qualified opportunities handed to sales with full context.

The buyer journey

  1. The buyer lands on a product page and is greeted with a question, not a form.
  2. They share their challenge in their own words.
  3. They receive tailored product recommendations instantly.
  4. They explore specs, configurations, and case studies on demand.
  5. They are offered a natural next step, and share their details to book it.
  6. Sales receives a warm, pre-qualified lead with full conversation context.

The results

Per the published results, the Lenovo deployment delivers:

The Foundry Works enterprise page adds that the Lenovo ISG deployment (data storage division) is a custom AI system with branded document generation, multi-channel deployment, and comprehensive analytics, which handles product queries, generates specification sheets on demand, and routes complex requests to human representatives. It is described as live and operational.

Brother: expert product support across a 5,000-partner channel

Brother is a technology manufacturer, and the brief was channel partner enablement. The deployment is described on the Foundry Works site as live now.

The challenge

Brother UK operates through a network of thousands of reseller and channel partners, each with varying product knowledge, sales capability, and bandwidth to engage end customers effectively. Delivering consistent, expert-level product education at that scale through traditional training and support is not sustainable.

The approach

Foundry Works deployed an AI agent team directly into the partner channel, available to resellers and their customers at the point of need. Product expertise on demand, at scale, without adding headcount. Partners get a sales tool that knows the full catalogue. End customers get expert guidance without waiting for a callback.

How it works in practice

  1. A reseller or end customer starts a conversation at the point of need.
  2. The agent identifies whether they are a partner, prospect, or existing customer.
  3. It delivers the right information: product, pricing tier, or support.
  4. It handles follow-on questions without any human escalation needed.
  5. It surfaces a qualified opportunity and routes it to the right sales contact.

The results

Per the published results, the Brother deployment delivers:

The Foundry Works enterprise page also lists Brother International as a custom AI deployment described as an active engagement delivering measurable efficiency gains.

Before and after: what changed in production

Built only from the outcomes on record, the pattern across both deployments is the same. Specialist agents each own one stage, and work that used to depend on human hours or human availability now runs continuously.

DimensionBeforeAfter (in production)
Lenovo first contactGated assets and forms, high bounce, low-quality leadsA question, not a form; friction removed at first contact
Lenovo lead qualityCold form-fills needing heavy sales development time to qualifyWarm, pre-qualified opportunities with full conversation context
Lenovo cost and coverageCost tied to gated content and human qualificationDramatically lower cost per qualified lead; always-on across time zones
Brother channel knowledgeThousands of partners with uneven product knowledgeConsistent, expert-level support across 5,000+ resellers
Brother scaleTraditional training and support that does not scaleInstant expertise at point of sale; handles spikes at zero added cost
Team modelHuman headcount added to grow capacityThree-agent team per deployment; capacity grows without headcount

Every cell above is drawn from the published Lenovo and Brother descriptions. No figure has been added that is not on record.

Why this counts as proof, not theatre

The reason these deployments matter is that deployment is the hard part. AI theatre is a chatbot with no workflow, a deck with no deployment, or a pilot that never reaches operations. Lenovo and Brother are the opposite: named enterprise manufacturers with agent teams live in production, each agent expert at one stage, each system connected to real buyer and partner journeys. Foundry Works is not pre-revenue; these are described as production systems generating recurring monthly value.

This is also why the same Guide, Expert, Closer thinking runs through the wider offer. To see the operating model behind it, read agent teams running in production, and for the category definition and how these deployments fit the bigger picture, start at the AI agent development company hub. A dedicated commercial explainer on what Foundry Works builds and owns, and a comparison of custom agents versus off-the-shelf tools, are planned and will link into this case as they ship.

Frequently asked questions

Are the Lenovo and Brother AI deployments live or pilots?

Both are described on the Foundry Works site as live and operating now, not as pilots or demos. Lenovo runs a production three-agent system for lead generation across direct and partner channels. Brother runs an agent team live across its UK reseller network. Foundry Works states it is not pre-revenue and that these are production systems.

What did Foundry Works build for Lenovo?

Foundry Works built a three-agent architecture around the buyer journey: a Guide that qualifies through conversation, an Expert that delivers tailored product intelligence in real time, and a Closer that captures the lead in the context of a next step. The enterprise page adds branded document generation, on-demand specification sheets, multi-channel deployment, and routing of complex requests to human representatives.

What results did the Lenovo agent deployment produce?

Per the published results, the Lenovo system delivers a dramatically lower cost per qualified lead versus gated content, always-on coverage across time zones and markets, a three-agent team where each agent is expert at one stage, and live operation handling real buyer conversations. Specific percentage figures are not published, so they are described qualitatively here.

How does the Brother deployment support 5,000 resellers?

Foundry Works deployed an agent team into the Brother partner channel: a Partner Concierge for enablement, a Product Expert for technical guidance, and a Sales Enabler for opportunity capture. It gives more than 5,000 resellers and their customers instant, consistent product expertise at the point of sale, and scales through volume spikes with no additional cost or resource.

Does the client own the AI system Foundry Works builds?

Foundry Works operates on a scope-and-build model rather than per-seat licensing, so clients own the agents once deployed. There is no vendor lock-in and no recurring licence fee for capability already paid for. A client can keep Foundry Works embedded as an operations partner or take full ownership internally.

What makes these deployments different from AI theatre?

AI theatre is a chatbot with no workflow, a deck with no deployment, or a pilot that never reaches operations. The Lenovo and Brother systems are the opposite: named enterprise clients with agent teams live in production, connected to real buyer and partner journeys, with each agent expert at one stage of the workflow. The proof is that they are running, not demonstrated.

About the author

Jason Sibley is the founder of Foundry, the company behind Hello Foundry and Foundry Works. He leads strategy across both, setting direction and keeping the work tied to real client outcomes rather than activity. His background spans sports marketing, technology and Web3, building engagement and growth systems for clubs, brands and platforms. Alongside Foundry he runs Cleo Group and Zenko Protocol, and he writes much of the company's thinking on AI agents, marketing and the economics of AI. Foundry runs on the same connected, agent-driven model it builds for the local businesses it works with.