From Boutique Expertise to Enterprise Scale: How D4S Partnership Built SOLAS with CoachBot

The Partner

D4S Partnership (D4SP) is a UK based, independent consultancy specializing in leadership development, executive coaching, and behavioural and organizational change. Their team brings together organisational and behavioural psychologists, ICF-accredited PCC and MCC executive coaches, business mentors, HR professionals and facilitators, working across financial services, luxury goods, education, government bodies, charities, and other sectors.

D4SP describes their own difference as a refusal to play it safe: authentic partnering over off-the-shelf delivery, a willingness to constructively challenge clients in service of better outcomes, and a focus on the whole picture behind an organization's people challenges rather than a narrow brief. That combination - an evidence-based rigour paired with a genuine appetite to question convention is also what shaped how they approached AI.

When D4SP began exploring AI in coaching, the question was not whether AI belonged in their offering. It was how to bring it in without compromising the standards that built their reputation. Managing Partner Steve Hendry has written publicly on this question, arguing that not every AI that can hold a conversation can hold a coaching one, and pushing on what AI coaching would actually have to do to deserve the name. That same scrutiny shaped D4SP's choice of technology partner.

The Challenge

D4SP saw a gap that most of the coaching industry recognizes, but few have solved. There is a strong business case for coaching at that level, but D4SP realized the return on investment becomes far more compelling early on - supporting those at more junior levels of management, where the benefits have decades rather than years to compound. Coaching remained concentrated at the top often because it has been too costly and time-intensive to extend to the levels below. AI created a genuine opportunity to change that equation, but only if it was done in a way that avoided compromise on three important principles D4SP believe in:

• Their intellectual property. Frameworks and methodology built up over nearly two decades of practice needed to become a scalable product, not a training set someone else could copy.

• Client trust. Any AI introduced into their client relationships had to earn that trust, not exploit it.

• Professional standards. Coaching delivered by or alongside AI needs to be held to the same standards as coaching delivered by a person, including adherence to ICF principles.

This is what led D4SP to CoachBot as their AI intelligence partner, and specifically to CoachBot's done-for-you service rather than building agents independently. That choice allowed D4SP to stay focused on what they do best (delivering human-led leadership and talent development), while CoachBot's AI coach engineering team led them through the build.

Building SOLAS AI: A Methodology First Approach

Working with CoachBot, D4SP built their foundational AI coaching companion, which they named SOLAS (the Irish and Scottish Gaelic word for ‘light’ or ‘guidance’). The name reflects how D4SP wanted the tool to feel: not a chatbot focused on giving answers, but a coaching companion that helps someone find their own way forward.

One deliberate choice sets SOLAS AI apart from many AI coaching tools on the market: it has no human or animal persona. Where some providers give their bots a name and personality designed to feel like a person, D4SP made the opposite call. SOLAS is presented plainly as what it is: a structured tool, not a simulated colleague. For a firm whose entire proposition rests on keeping humans at the centre of development, that distinction matters more than it might first appear.

D4SP describes SOLAS's development as a four-stage journey.

• Concept. D4SP identified the coaching access gap first-hand: the majority of employees never receive any coaching at all, often because, historically, it has been too costly and time consuming to extend beyond senior leadership. D4SP framed SOLAS from the outset as a coaching companion grounded in coaching best practice, not a generic chatbot with a coaching skin.

• Partner. D4SP selected CoachBot as their technology partner and began integration design, testing conversational flow and refining the user experience together.

• Develop. D4SP built out the coaching conversation architecture, which included persona, tone and ethical safeguards, all aligned to D4SP's own values, principles and people development philosophy.

• Launch. D4SP prepared SOLAS for rollout with pilot organizations, building the commercial model and the narrative to bring it to HR audiences.

The build itself followed a structured process that was completed in approximately four weeks, with dedicated focus from the partner side, resulting in a custom, proprietary conversational AI coaching agent based on D4SP's own methodology and style.

D4SP integrated their own coaching and training frameworks, along with hard guardrails they were not willing to negotiate (e.g. no therapy, no health advice). CoachBot's team, having built hundreds of coaching agents, brought judgment on which methodologies and standards actually hold up when translated into an LLM, based on experience and current research.

What distinguishes this process is where it starts; not with the technology, but with a simple question posed jointly with the partner: what does ‘good’ mean, in the specific context of AI coaching? That answer becomes a partner-specific evaluation framework. CoachBot's ‘LLM-as-judge’ evaluation tooling then tests each agent against that framework at scale, using synthetic conversation data generated for the intended audience. Every evaluation run produces a detailed report, scores, and insight into what is working and what needs refinement. CoachBot's AI coach engineers use those results to refine the agent before it ever reaches a human user in beta.

The outcome for D4SP was two-fold: a ready-to-deploy AI tool for their clients, and objective, measurable evidence of that tool's value they could present to client organizations during the vendor selection and buying process.

From Foundation to Fit for Purpose

A foundational agent is a starting point, not an endpoint. D4SP initially piloted SOLAS AI as a general-purpose coaching companion but quickly found that specialization is what actually landed with clients. The model that emerged kept a common structure at the start and end of every session, with the in-session conversation built around a specific coaching methodology or use case. They found this was much more effective than promoting a singular, multipurpose bot that attempted to be all things to all people. D4SP has since adapted SOLAS AI across a range of these more focused applications:

AI-led client follow-ups between coaching sessions, extending the coaching alliance naturally so clients are never left without support - or worse, left with a generic AI substitute when their coach is not available.

A standalone tool supporting clients working through imposter syndrome. This use case plays to a genuine strength of AI: clients often find it easier to speak openly to an AI agent about something they would hesitate to raise with a human coach. CoachBot supports this by keeping conversations private by default, with clients choosing what, if anything, to share with their coach.

Client-specific versions, customized with pre-approved organizational context in each client's own knowledge base within CoachBot Studio. On top of the default two-layer memory system that lets each agent remember what an individual client has shared, D4SP can add organizational values and terminology, client-specific transformation goals, industry challenges, onboarding material, and company-specific guardrails. That level of customization turns a good tool into one an organization trusts enough to adopt at scale.

Compliance and Ethics: A Deciding Factor, Not a Checkbox

D4SP has been clear that the partnership with CoachBot rests on shared values as much as shared technology. Four things stood out to D4SP when choosing a partner:

• Coaching fidelity. CoachBot's conversational AI is purpose-built for coaching, not a generic chat model repurposed for the job. That alignment with coaching methodology and ethics, including ICF standards, was non-negotiable for D4SP.

• Values alignment. A shared commitment to ethical AI use and genuine human development, not technology adoption for its own sake.

• Innovation readiness. A platform built to evolve, giving D4SP the flexibility and pace to keep developing SOLAS as their needs change.

• Scalability with consistency. The ability to reach thousands of professionals at once without a drop in quality or governance.

Underpinning all four is CoachBot's infrastructure: built enterprise-ready from the start, with full GDPR compliance and data protection standards benchmarked against the highest global bar, including the EU, German & UK data protection frameworks that many multinational and public sector clients require as a baseline. For D4SP, that meant they could tell clients that their data, and their clients' data, was handled to a standard built for organizations with the strictest compliance requirements, not retrofitted to meet them after the fact.

For coaching providers and HR leaders evaluating AI partners, this is often the quiet deciding factor. It rarely makes the headline slide, but it is usually the first thing a legal or procurement team checks, and the first thing that can end a deal if it is missing.

The Result: A Landmark Win

This combination (proprietary methodology, rigorous evaluation, and genuine customization) has translated directly into commercial outcomes for D4SP. Alongside upskilling their own team on AI literacy and maturing their AI strategy from concept to practice, D4SP secured what is, to date, one of the most strategic new client wins in the firm's recent history: a comprehensive two-year leadership development program, comprising more than 200 leaders, with a large, well-known UK organization.

The win came after a long, competitive vendor selection process. By the client's own account, the deciding factor was not price or a features list. It was D4SP's demonstrated leadership in AI-augmented development. When D4SP brought their own client-specific CoachBot to an onsite presentation and demonstrated it live, the client saw, concretely, what "hybrid coaching done well" actually looks like. That live demonstration is what closed the deal.

"With CoachBot, we can deliver truly hybrid coaching (human where it matters, AI where it scales) and do so responsibly. It gives us a practical path to L&D for the AI era, where organizations develop both human talent and agent enabled workforces." Steve Hendry, Managing Partner, D4S Partnership

Why This Matters for Coaching Providers and HR Leaders

D4SP's journey is a useful reference point for two audiences.

For coaching providers, it shows a path to scale without choosing between proprietary methodology and AI adoption. A done-for-you build, grounded in a partner-specific definition of quality and in-depth testing before rollout, means the resulting product still sounds and behaves like your practice, not a generic chatbot wearing your logo.

For HR leaders, it is a concrete example of what "responsible AI in coaching" looks like in practice: private-by-default client conversations, adherence to ICF-aligned standards, org-level customization and guardrails, and a measurable evaluation process behind the tool before it ever reaches an employee. That is the difference between AI as a novelty and AI as infrastructure you can actually govern.

Lewin Keller

Founder & CEO

Ex. Google, DoiT, Accenture · ACTP Coach · Investor & Advisor