Mindset AI

We start with your KPIs

Every engagement begins by defining the results that matter to you — then we help you hit them. Real metrics, real timelines, real accountability.

Learning Curve Group: From an AI tutor proof of concept to one platform for learner, customer and staff support

Industry: Education & Training

Challenge: Learners increasingly expect timely and accessible support, including outside standard working hours. At the same time, LCG wanted to improve the consistency of guidance and reduce pressure on teams without introducing a series of disconnected AI products, each requiring its own content, controls and user experience. The requirement was therefore broader than deploying a single chatbot. LCG needed an approach that could support an immediate learner use case while establishing a scalable model for future agents across learner, customer and staff services.

Solution: Learning Curve Group integrated Mindset AI into its LMS, eAssessorPro, to create EVA, an AI Support Tutor grounded in approved course content and learner guidance. By placing the AI Support Tutor within a system already used by learners, support is available in the context of their learning rather than through a separate destination. EVA draws on course content and learner guidance to provide relevant, contextual assistance. This implementation also established a "one platform, many agents" model. Instead of building separate solutions for every requirement, LCG can reuse the same underlying platform and approved knowledge approach, while configuring each agent for a clear audience and purpose.

  • 19% difference in completion rates — For learners supported by EVA, the AI Support Tutor inside the LMS
  • 20% difference in active conversion — In the learner group using EVA
  • 12% fewer non-starters — Reduction in learners who never start their programme
  • Live chat replaced — A Mindset AI website agent now handles the previous live chat service

Hult Ashridge: Ava live across 33 client programs, built with less than 15 hours of engineering time

Industry: Corporate Executive Education

Challenge: Executive learners on Hult Ashridge programs receive coaching sessions, but the gaps between sessions are where development stalls. By mid-2025 learners were filling that gap with ChatGPT: generic answers with no knowledge of Hult Ashridge's proprietary research or coaching frameworks. Hult Ashridge Executive Education, part of EF Education First, partners with global organizations to deliver leadership and organizational development solutions that create lasting change. Backed by triple-accredited Hult International Business School, it has been delivering transformational leadership and organizational development since 1959, and runs 2,000+ programs a year across all sectors. Emma Thompson, Head of Technology, leads an engineering team that maintains two platforms: Xceed (program delivery and cohort management) and Leadership Live (a content library with research, video, audio and interactive exercises). A learner finishes a coaching call, wants to revisit a concept or apply a framework to their own situation, and has nowhere to go except a static content library or a generic chatbot that knows nothing about their program. Leadership Live had the content. Xceed had the program structure. Neither had a conversational layer that could connect a learner to the right material in context. The content existed; the access path did not.

Solution: Built Ava, a virtual learning assistant, into both Xceed and Leadership Live on the Mindset stack. Each program gets its own agent scoped to its own content, with no data shared between clients, and nothing in the existing MongoDB, Vimeo or xAPI stack had to change. The core ask was quick intervention. As Emma framed it: a participant has a coaching session in two weeks, has a challenge right now, and needs an answer in the moment. Ava fills that gap between scheduled touchpoints. Phil's team connected MongoDB to the knowledge bank ingestion API, creating program-specific knowledge banks from the existing content library (HTML, MP3, MP4, SCORM and interactive exercises). Ingestion runs through the API, so knowledge banks update automatically as the editorial team publishes each month. Multi-tenant access is configured through the Sessions API. Each program gets its own session configuration, with an external tenant ID mapped to Hult Ashridge's internal client identifiers, so a learner in one program never sees content scoped to another. Agent behaviour is controlled in the Agent Management Studio, and Leadership Live runs a single "plug and play" agent across the full content library for all clients. Phil built Ava into both platforms using the Mindset Agent Chat SDK: guide-on-the-side in Xceed, full-screen or sidebar in Leadership Live. The agent runtime sits on top of the existing architecture without replacing any component, and analytics feed back into the existing xAPI and Luzmo pipeline for client-facing metrics.

  • 33 programs live — Ava in 33 client program experiences within weeks of launch
  • <15 hours of engineering — Total engineering time to deliver the first agent
  • NPS 76% → 80% — Moving upward on pilot programs, with strong gold-tier client feedback
  • Zero stack changes — Existing MongoDB, Vimeo and xAPI stack left untouched
"Our clients were already using ChatGPT between coaching sessions. We had RFPs for AI coaching and nothing to answer them with. We needed something that could give each program its own agent with access to the right content, across 2,000 programs, without any data leaking between clients, and that could sit inside both our platforms without us having to rebuild them. I also wasn't willing to add five vendors to get there. I wanted one partner who covered the lot."
— Emma Thompson, Head of Technology, Hult Ashridge

Mindtools Kineo: AI coaching for nearly 20,000 learners across 200+ organisations, built by a three-person team

Industry: Learning & Development

Challenge: Mindtools Kineo's 3,000-item content library is a core part of its value, but when a manager needs coaching between meetings, that same library becomes a blocker: too much content, too hard to find, at exactly the wrong moment. Learners weren't reaching the right content at the point of need. Mindtools Kineo has produced management and leadership content since 1996. The library runs to over 3,000 articles, videos, toolkits and structured short courses, serving 200+ client organisations (including the NHS, British Council, Volvo, Whitbread and Center Parcs) across 160 countries. The L&D market is moving fast. Content expectations have shifted, and Mindtools Kineo was also integrating recently acquired products. The first two use cases were clear: find a resource or topic quickly, and get a specific question answered at the point of need. A manager handling a difficult conversation needs an answer in the moment, not a browsing session.

Solution: Built Ask M: Coach, an AI coaching agent grounded in the full content library, on the Mindset stack. One agent configuration serves the web Content Hub, a standalone mobile app and the Skills Accelerator, with each client organisation's content entitlements enforced per session. A line manager types a question ("how do I handle an underperforming team member?") and the agent responds with coaching guidance grounded in the library, citing and linking to specific articles, videos and toolkits. It launched in March 2025. The Skills Accelerator, rolled out to professional and enterprise tier clients in January 2026, is a structured three-step learning journey for managers. Managers complete a 40–45 minute assessment across 12 core management skills, are directed to a targeted e-learning suite based on their gaps, and then land on an apply stage where Ask M: Coach is the primary feature. Pre-configured prompts around common challenges such as delegation and giving feedback connect directly into the agent. Both products run on a single agent, with system prompts set up in the Agent Management Studio. Different client organisations see different content based on their entitlements, configured per session via the Sessions API.

  • Nearly 20,000 learners — Have used Ask M: Coach since launch
  • 73,000+ coaching messages — Sent to the agent across web and mobile
  • Two products, one agent — Three surfaces from one configuration, built by a three-person technical team
  • Clean IT security reviews — GDPR compliance and prompt-injection protection across 200+ client organisations
"We needed a toolkit that gave us the flexibility to deploy and integrate in different ways across web, mobile, and new products. It had to be something our product and content teams could run directly after implementation, not something that created a permanent engineering dependency. We built the agent experiences we needed, our non-technical teams run the day-to-day, and engineering supported with implementation so the core platform roadmap was not sacrificed."
— Michaela Heigl, Executive Director Performance Enablement Solutions, Mindtools Kineo

Future Talent Learning: 95.2% learner retention and 6 hours saved per coach each month

Industry: Learning & Development / Apprenticeships

Challenge: FTL needed to scale coaching support across 1,300+ apprenticeship learners without growing their coaching team proportionally. Coaches were spending significant time on routine learner questions, and FTL needed to prove an AI-assisted model quickly — within a three-month window.

Solution: Built a per-course AI tutor on the Mindset stack that handles learner questions around the clock, surfacing only the conversations that genuinely need human coaching. Deployed across all active courses in under three months, integrating directly with existing content without a custom infrastructure build.

  • 95.2% learner retention — Across 1,300+ apprenticeship learners using the AI tutor
  • 6 hours saved — Per coach each month, redirected to high-value conversations
  • 116k+ interactions — Learner conversations handled by the AI tutor
  • <3 months to deploy — From kickoff to live across all active courses
"The Mindset platform let us scale coaching support without scaling our team. Our coaches are spending their time on high-value conversations, and our learners are staying engaged."
— Rubecca Ali, Head of Learning Technologies, Future Talent Learning

: Shifting a 2,600-course catalogue to product-led growth

Industry: Maritime & Energy Safety Training

Challenge: A maritime safety training marketplace had 2,600 courses structured for traditional browse-and-search, making it difficult for professionals to find the right certification pathway under time pressure. Discovery was passive and the provider wanted to shift to a conversation-driven, product-led growth model.

Solution: Replaced catalogue browsing with an AI-powered discovery conversation on the Mindset stack, mapping learner needs to the right courses through 14 defined lifecycle actions. The new experience debuted at five customer conference sessions to accelerate commercial adoption.

  • 2,600 courses — Entire catalogue surfaced through conversational discovery
  • 14 lifecycle actions — Defined to guide learners from first interaction to certification
  • 5 conference sessions — Customer-facing events anchored around the new AI experience
"The shift from browse to conversation fundamentally changed how our customers discover training. It's the kind of experience we'd always wanted to build but never had the platform to do it quickly."
— , CTO, Maritime Training Marketplace

: A 20-product agent architecture proven in 90 days

Industry: Healthcare Workforce & Clinical Education

Challenge: A healthcare workforce and clinical education platform needed to validate a complex multi-product AI architecture — 20 separate agents, each scoped to a different product domain — within 90 days. The architecture required sub-three-second retrieval across all agents without compromising domain isolation.

Solution: Built 20 product-specific agents on the Mindset stack, each backed by a dedicated MCP server connected to the relevant API. Achieved 2–3 second retrieval across all agents and proved the full architecture in production within the 90-day window.

  • 20-product architecture — Proven in production, one dedicated agent per product domain
  • One MCP server per API — Clean isolation between every product domain's data and logic
  • 2–3 second retrieval — Across all 20 agents in the live environment

: Course completion up 3.7 points in 90 days with a per-course AI tutor

Industry: Apprenticeships & Vocational Training

Challenge: An apprenticeship training provider needed to improve completion rates across vocational programmes, where learner disengagement and knowledge gaps were going undetected until assessment. They needed visibility into where learners were struggling — at scale — without increasing tutor workload, and with zero tolerance for safeguarding risk.

Solution: Deployed a per-course AI tutor on the Mindset stack that surfaces knowledge gaps through conversation, with safeguarding guardrails built in from day one. Within 90 days completion rates had improved measurably and the system had flagged genuine curriculum gaps at scale.

  • +3.7 points — Course completion improvement in 90 days
  • 17.4% of questions — Surfaced genuine knowledge gaps for curriculum review
  • Zero safeguarding incidents — Full safeguarding compliance maintained throughout
"We needed something that could scale across all our courses without adding tutor load — and that took safeguarding seriously from the start. This delivered both."
— , CTO, Apprenticeship Training Provider

Major Learning Platform: 900% increase in course enrollments

Industry: Online Education

Challenge: Learners were overwhelmed by 5,000+ courses without clear starting points. Completion rates were low and certificate sales underperformed due to limited perceived value.

Solution: Built a recommendation agent on the Mindset stack connected to the full course catalog. The agent asks clarifying questions, surfaces personalised recommendations in seconds, and was live without a custom build.

  • 900% increase — In course enrollments among users who interacted with the AI agent
  • 3.4x higher revenue — Generated by AI agent users versus the overall test group

BreakPoint: 30,000+ employees deployed across enterprise clients

Industry: Leadership Development & Wellness

Challenge: Traditional senior management training diluted as it cascaded through teams. Employees struggled with content accessibility and engagement. Break Point needed to scale teaching without relying on hierarchical information distribution.

Solution: Built white-labelled web and mobile agents using the Mindset stack — unlimited agents, guided conversational search, integrated knowledge banks, and analytics. Deployed fast without building the agent infrastructure from scratch.

  • 30,000+ employees — Deployed nationally across enterprise clients including Shell, Manchester University, and Bentleys
  • Strong NPS — Among senior leadership teams
  • Daily engagement — Users actively asking questions every day
"The AI ability is dynamic, its individual. I can put a question in there to stimulate an answer thats really tailored to me, and thats what I find so powerful."
— Ollie Ollerton, Founder, Break Point Academy

Training Industry: 1,400 signups in 3 months with an 11% conversion rate

Industry: Learning & Development

Challenge: Training Industry had 200,000 monthly visitors browsing free content then leaving without converting. Thousands of articles, podcasts, and videos made discovery overwhelming, and legacy content was buried.

Solution: Built Tia (Training Industry Assistant) on the Mindset stack — a white-labelled agent embedded in their membership portal in weeks, with real-time content syncing from WordPress, Wistia, and Podbean.

  • 1,400 signups — Free trial signups in the first three months
  • 11% conversion — From free trial to paid membership at $30/month
  • ~150 new subscribers — Generating tens of thousands in new recurring revenue

Persimmon Life: 40% of growth managers' time saved in month one

Industry: At-Home Health & Beauty

Challenge: Regional growth managers spent 60–80% of their time answering repetitive FAQs from 500 trained nurses. Resources were scattered across Google Drives and learning platforms, and healthcare compliance required accuracy guarantees.

Solution: Built compliance-ready agents on the Mindset stack with accuracy controls and healthcare guardrails. Connected directly to existing Google Drive repositories with automatic content sync — no custom infrastructure required.

  • 40% time saved — For regional growth managers within the first month
  • 50%+ of target — Achieved within the first month of launch
  • Adoption doubling monthly — Nurse usage growing every month
"The feeling from our providers is just, thanks. They're thankful they have a space for answers, where going for a marketing answer they go here, for a business answer you go there and inventory answer you go somewhere else. So the overwhelming word is thanks."
— Rachel, Regional Growth Manager, Persimmon Life

5app: Went from a 9-month timeline to weeks with the Mindset stack

Industry: E-learning SaaS

Challenge: 5app needed to deliver meaningful AI innovation quickly while managing competing priorities and a complex multi-tenant SCORM environment. They lacked dedicated engineering resources for a custom AI build.

Solution: Mindset AI provided Agent Chat SDK, multi-tenant APIs, SCORM ingestion pipeline, agentic workflows, and MCP server integrations. 5app built VeeCoach, an AI coach integrated into their product.

  • 9-month timeline to weeks — Built coaching, compliance, and course discovery without starting over each time
  • <2 hrs/month — Ongoing maintenance managed entirely by the product team
  • Increased ACV — New pricing model increased Average Contract Value
"We went from a 9-month timeline to weeks — the Mindset stack let us build multiple use cases — coaching, compliance, and course discovery — without starting over each time."
— James Cranwell, Head of Product, 5app

HACT: £800k+ recurring revenue and 80 customers upsold

Industry: Social Impact & Housing

Challenge: HACT faced pandemic-driven behavior shifts requiring digital transformation. Manual processes prevented scalability, and the organization had limited digitization experience while needing cost-effective solutions.

Solution: Built the Social Value Insight tool on the Mindset stack in 6 weeks — automated data collection, simplified customer journeys, and a repeatable framework they could reuse for future tools without starting over.

  • £800k+ recurring revenue — Moved to a scalable recurring revenue model
  • 80 customers upsold — Converted into managed accounts within two months
  • 6-week setup — Solution fully configured and launched
  • 5 days/month saved — After two months of implementation
"We built the Social Value Insight tool in 6 weeks. The feedback from our customers has been outstanding."
— Jacqui Bateson, Managing Director, HACT

AchieveUnite: $1M in enterprise opportunities with 20% partner retention boost

Industry: Partner Training & Enablement

Challenge: AchieveUnite needed products supporting a recurring revenue model. Their vast collection of books, assessments, and learning materials required on-demand delivery at the point of need for enterprise clients like HP, Lenovo, and Atlassian.

Solution: Built multiple specialist agents on the Mindset stack — each acting as a topic expert across their IP, research, and content. Embedded into partner systems as a workplace tool without building agent infrastructure from scratch.

  • $1M in opportunities — Enterprise pipeline generated
  • 40% faster onboarding — Accelerated partner enablement speed
  • 20% retention boost — Increase in partner retention rates
"We built multiple specialist agents — each acting as a topic expert — and surfaced them into partner systems. We now have a $1M enterprise pipeline and 20% higher partner retention."
— Jessica Baker, Co-founder & Chief Program Officer, AchieveUnite