The proof, in the FAQ

Reading the case studies, answered in full.

How to read the write-ups as one body of work rather than one at a time: what the attribution note covers, and what the set is and is not evidence of.
questions in this group, each answered in full
18
pages the answers are written on, every one linked
1
questions across the whole FAQ
1293

18 questions on reading the case studies, answered by Tenhaw, a UK AI consultancy and AI delivery partner based in London. Nothing here is a summary: each answer is the exact text from the page that owns it, and every group links back to that page for the context around it.

Elsewhere in the FAQ
18 questions

All 12 engagements

Answered on All 12 engagements, and rendered here in the same words.

Read the page these answers live on →

Why does an AI consultancy lead with delivery case studies?

Because agentic transformation runs on delivery discipline, and the page is honest about which is which. Two of the twelve engagements are ones where AI was the deliverable: the live specialty insurance build and the HSBC Voice Insights proof of concept. The other ten are product delivery, PMO and operating-model work inside organisations like HSBC, Anglo American and YOOX NET-A-PORTER, and every one of those write-ups names what transfers to agentic work and what does not. Agents amplify whatever operating culture they land in, so the delivery track record is not padding: we think it is why the agentic work landed.

Which of the twelve engagements were delivered under the Tenhaw banner?

Eight of them. Anglo American, Yondr, Greggs, Colart, Tecknuovo, Globelynx and YOOX NET-A-PORTER were delivered under the Tenhaw banner, as is the live specialty insurance engagement, and several of those predate incorporation and were delivered by James Rooney personally on contract. The other four, the three HSBC write-ups and Discovery, were James's own delivery and transformation roles inside those organisations. The page keeps the two framings apart rather than blending them into one undifferentiated track record, and we will walk you through which is which on a call.

Are the results in these case studies from production systems or pilots?

Each write-up labels its own status in the same breath as the result. The HSBC Voice Insights study says it reached proof of concept, and that 1.5 million hours a year is a projection rather than a saving anyone has banked. The live insurance engagement's two-week build is described as a working proof of concept, with month three productionising it against the client's security standards. The HSBC Global Payment Solutions operating model was designed, piloted and validated against real feedback, with global rollout due in 2026 and not yet run. When production lands, it is published here, dated.

Why is one of Tenhaw's case studies anonymous?

The current agentic engagement is with a London specialty insurance business that has not consented to being named, so it is written up unnamed. The detailed architecture, the model and platform choices, the prompt and review approach and the working history of the build sit under client confidentiality alongside the name. What is published is what can be: the shape of the engagement, the pipeline stage by stage, and the results. You should expect the same treatment of your name and your estate if you engage us.

Is Tenhaw running an agentic engagement right now?

Yes. The live engagement is an Embedded Agentic Lead role in the London specialty insurance market, three months in and ongoing. Month one was an audit that took us across the whole programme, month two built a working proof of concept in two weeks, PDFs in and business intelligence out on Azure, covering ground that had previously taken the business roughly twelve months, and month three is standing up an adjacent agent-led engineering team to productionise it on a four-to-six week target. The client is confidential, so the write-up is unnamed, but the work is current and the fortnight was pair-programmed with the client's own engineer.

What measurable results are in Tenhaw's case studies?

Several, each stated in the write-up it comes from. Globelynx cut delivery lead times by 60% within six months. Discovery+ hit a CEO-set launch date across six development teams. Greggs got data proving that tackling tech debt accelerated delivery. HSBC's executive team completed annual planning ahead of schedule for the first time in years, with blockers cleared in days rather than weeks. The live insurance engagement rebuilt roughly twelve months of stalled work as a working proof of concept in two weeks. Where a number is a projection rather than a banked saving, the write-up says so.

What sectors do Tenhaw's case studies cover?

Four of the twelve sit in financial services: three HSBC engagements and the live specialty insurance one. Five are retail, media and consumer: Discovery, Greggs, Colart, YOOX NET-A-PORTER and Globelynx. Two are industrial and energy: Anglo American's hydrogen mining venture and Yondr's data centre business. The Tecknuovo engagement governed a 19-project portfolio that included HMRC, MOD and Thames Water work, which is where the public-sector exposure sits. If yours is not on that list, the delivery and operating-model methods have crossed sectors repeatedly; ask on the call which engagement yours most resembles.

Has Tenhaw run delivery across multiple countries and time zones?

Yes, and two of the case studies were structured around exactly that. The Anglo American engagement ran digital teams across the UK, Australia and the USA, with delivery run asynchronously by necessity and throughput data feeding Monte Carlo forecasts, so alignment did not depend on shared office hours. The Yondr engagement ran remotely across the UK, USA and Singapore, reinforced with on-site workshops, and had development teams producing consistent sprint output within three months. Distributed delivery is a demonstrated part of this track record rather than an aspiration.

What is the largest scale in Tenhaw's track record?

HSBC Global Payment Solutions is the largest operating footprint: James co-led the design and piloting of a target operating model for 500 teams and a $450 million budget, due for global rollout in 2026. Alongside it sit executive ways of working across 150+ teams and a $102 million budget at HSBC, coordination of five agile teams on YOOX NET-A-PORTER's £1 billion re-platforming, and the delivery engine behind Anglo American's £40 billion hydrogen business case. The write-ups are precise about what those numbers mean: scope rather than authority, and piloted rather than rolled out where that is the truth.

What would an engagement like the insurance case study cost?

Its write-up states the published equivalent directly: £70k–£85k a month for an Agentic Build Team of three people under partner oversight, or £18k–£35k a month for programme oversight only. Every engagement price derives from the published rate card, partner at £1,560 a day, senior practitioner at £1,250 and associate at £950, excluding VAT, at twenty billable days a month. It is the one engagement on the page carrying a published equivalent, and the same rate card is what a comparable shape would be priced from today. The rates are on the pricing page, so there is no bespoke quote to wait for.

How long did the engagements in these case studies run?

From three months to eighteen. Three months is the shortest, covering HSBC's executive ways of working and the HSBC Voice Insights proof of concept, and the live specialty insurance engagement is three months in and still running. The longest is Anglo American's hydrogen venture at eighteen. Six months is the most common shape on the page: Yondr, Greggs, Colart and the HSBC Global Payment Solutions operating model, with Tecknuovo and Globelynx at nine, Discovery+ at ten and YOOX NET-A-PORTER at twelve. Length follows the outcome rather than a contract term: engagements are retainer-shaped, with the exit date agreed at kickoff and 30 days' notice either side.

Which case studies cover operating model design rather than delivery?

Four of the twelve. The hub sorts the work into three disciplines, and the target operating model group holds HSBC Global Payment Solutions, HSBC executive ways of working, Tecknuovo and Yondr: structure, roles, decision rights and governance, designed and piloted at scale rather than drawn on a slide. Two engagements sit under AI and agentic, where the AI itself was the deliverable, and six under product and delivery. Each engagement appears under one discipline only, so the three groups sum to the twelve. If operating model design is what you are shopping for, start with the Global Payment Solutions write-up: 500 teams, a $450 million budget, piloted and validated before any rollout.

What did clients keep once these engagements ended?

The capability, and several write-ups say what that looked like. By the end of the ten months on Discovery+, both the Discovery+ and Eurosport teams could run the forecasting and tracking practices autonomously. Tecknuovo kept a fully functional portfolio office and a junior delivery team coached on the live portfolio rather than in a classroom. Yondr could plan holistically, with predictable output across development, support and security. On the live insurance engagement the client's own engineer, who pair-programmed the entire two-week build, put their confidence at 70% that they could run the process unaided. Exit dates are agreed at kickoff, so leaving well is the design rather than the ending.

What kind of problem does Tenhaw usually get called into?

Delivery that has stopped being predictable, and rarely with the technology at fault. Read the twelve openings together and the pattern is consistent: a specialty insurance programme stalling in the gaps between product, engineering and platform; a payments division of 500 teams where every region ran its own processes; a CIO's executive team with no shared way to track strategic initiatives; global teams at Yondr delivering so erratically the business could not plan past a quarter; three merged teams at Colart that stopped shipping; an app team at Greggs stood up in a rush and no longer trusted by the rest of the business. Nobody in those situations was underperforming individually. The connective tissue was missing.

Do you publish client names in case studies?

Only where the client has agreed to it. Nine of the ten organisations behind these twelve engagements are named, because each was content to be. The tenth is written up as a business in the London specialty insurance market and nothing further: the architecture in detail, the model and platform choices, the prompt and review approach and the working history of the build sit under confidentiality alongside the name, and that is treated as binding rather than negotiable. Nothing about your name or your estate goes on this site unless you have agreed to it, and you should expect the same treatment of both if you engage us.

Has Tenhaw built AI inside a client's own infrastructure?

Yes. The live specialty insurance build ran from a blank repository on Azure inside the client's own regulated estate, with a security review roughly every fifth prompt, and the output lands in an internal dashboard the business already uses rather than a tool that only exists while we are there. The business logic layer stayed the client's own rules, the client owns all work product, and the fortnight was pair-programmed with one of their engineers. It is a working proof of concept, and month three productionises it against their security standards. The HSBC Voice Insights proof of concept ran inside the bank too, one of James Rooney's delivery and transformation roles rather than Tenhaw-banner work.

What can we check about Tenhaw before we book a call?

Four things, and none of them needs us. Nine of the ten organisations behind these twelve engagements are named, so anyone in your network who was in the room can be asked. The method is published in full and free to adopt without hiring anyone, and the engineering standard our engineers build to is an open-source handbook of 72 rules your reviewers can hold the work against. The entity is on the public register: Tenhaw LTD, registered in England and Wales, company 12735685, incorporated 10 July 2020, linked from the contact page. And the insurance write-up carries its published equivalent, £70k–£85k a month for an Agentic Build Team, so you can cost that shape yourself.

Has Tenhaw ever told a client not to build something?

Yes, and the clearest case is on this page. Anglo American's hydrogen venture existed to find out whether mining could run on hydrogen-powered trucks, and the decisive finding the digital teams produced was that those trucks were not yet cost-competitive, so leadership invested with eyes open rather than late. The same runs through the smaller calls: the Tecknuovo write-up says a handful of projects rarely justifies a portfolio office, and the Yondr one says a single team that needs its ceremonies run properly should hire a coach instead of us. On the £44,000 four-week audit, a recommendation to stop is a valid outcome, and the audit is standalone with no obligation to continue.

If the sources do not answer it, a call will.

Book thirty minutes
The rest of the FAQ

1293 questions, grouped by subject

Every question answered anywhere on tenhaw.com sits in one of 51 groups. This is one of them.

All 1293questions, and every group →

Or ask the question directly and skip the categories.

Book thirty minutes
book a call

Still have a question?

A 30-minute discovery call with James Rooney. Bring the question this page did not answer. You'll leave with a rough scope whether you engage us or not.

most start with a fixed-price AI Readiness Audit · £44,000 · 4 weeks · working prototypes

// pick a slot · cal.com/tenhaw/professional-servicesLIVE CALENDAR

Calendar not loading? Open it on cal.com or email hello@tenhaw.com.