Direction and diagnosis

When this is the right starting point

Few organisations arrive at a technology decision cleanly. You arrive under pressure: a board asking what the AI plan is, a competitor announcing something clever, a budget cycle demanding a number before anyone can say what it buys. Meanwhile the operational problems that would actually justify the investment sit in plain sight, unexamined, because they are familiar.

This is the moment direction and diagnosis was built for, and it is a better starting position than it feels. The pressure means your organisation is ready to move, and the problems worth solving are already there waiting. What you need next is solid ground. Start here if any of these sound like your situation:

• You are being pushed to adopt AI, but no use case stands out as clearly worthwhile, and you suspect some of the candidates are theatre.
• Operational problems are spread across teams and systems, and you cannot yet tell which one to tackle first, or what fixing it would be worth.
• You do not know whether your data, systems and governance could support the change you have in mind, and nobody internally can give you a disinterested answer.
• A supplier, product or venture is making a technical claim that will cost real money to believe, and you want it independently assessed before you commit.
• There are more plausible projects than there is budget, and you want the ranking decided by evidence rather than by advocacy.

What these situations share is not a technology gap. It is an evidence gap: a decision that matters, and not enough verified ground under it. Closing that gap is the whole job of this work.

And the ground is worth claiming, because the prize for deciding well keeps growing. What AI can do is rising faster than most organisations' ability to take value from it, so the distance between a well-chosen move and a fashionable one widens every quarter. An organisation that can tell the difference, with evidence, moves faster than one that cannot, and does it without gambling.

The forms your answer can take

The shape of the work follows the decision in front of you, not a fixed programme. Most engagements draw on one or more of these forms, combined and weighted to fit.

Focused discovery

Structured investigation of a specific problem or opportunity. We spend time with the people who live with the problem, map the workflow and systems around it, quantify what it costs, and establish what would genuinely change if it were solved. Discovery ends with a recommendation you can act on the next morning.

Readiness review

A clear-eyed examination of what your intended change would rely on: the state and reachability of your data, the systems it must connect to, the skills in the team, and the governance that would keep it safe. The output distinguishes what is ready, what needs work, and what the gaps would cost to close.

Technology, product and data strategy

A practical direction that serves the operation and its commercial goals, expressed as decisions and sequences rather than diagrams and slogans. It covers what to build, buy, integrate and retire, and in what order, with the reasoning attached so it survives leadership scrutiny.

Governance and prioritisation

A durable way to compare competing initiatives: how proposals are evidenced, how value and risk are weighed, who decides, and how the queue is re-ranked as facts change. This is how the loudest-voice problem gets replaced with something defensible.

Technical and commercial due diligence

Independent assessment of a technical claim, product or venture: what is actually built, how well, what it depends on, and whether it can carry the plan attached to it. Findings are reported as evidence, assumptions, risks and unknowns, each priced in effort rather than adjectives.

The first worthwhile use case

Identification of the first project that is worth doing and safe to learn on: real value, legible success measures, reachable data and a contained blast radius. First projects calibrate an organisation's trust in technology, so this choice is made deliberately.

How we get you to a decision

Whatever its form, the work runs on one discipline: the criteria for judging the answer are fixed before the answer exists. First we frame the decision: what is actually being decided, by whom, by when, and what evidence would settle it. Where the decision will be judged against numbers, the success thresholds are agreed with you in advance, in writing, and the eventual recommendation answers to them exactly as written.

Then we build the evidence where the problem lives: time with the people who do the work, examination of the systems and data involved, analysis of the commercial numbers, and direct technical inspection where a claim is being tested. Every finding carries a grade for how we came to know it, from measured at first hand down to estimated with stated assumptions. Conclusions rest only on the strongest grades, and assertions that matter get a verification action, not silent acceptance.

Finally we put the decision back in your hands, in a form your board or investment committee can interrogate: the options, the evidence behind each with its grade, what we could not verify and what closing each gap would take, and our recommendation with its reasoning exposed. Then we work the recommendation through with your leadership in person, because the goal is a decision that moves.

Scope and price are agreed before we start, sized to the weight of the decision. And when the decision needs measurement rather than judgement, the productised version of this work is the Evidence Sprint: ten days measuring the workflow that matters, ending in a go, change or stop answer.

What you take away

Every engagement ends with material you keep and can act on without us:

• A decision report with a short memo: the problem as evidenced, the options considered, the recommendation and the first step, written to be challenged, with a two-page summary your board can absorb in minutes.
• The evidence base: what we examined, what we found and how each finding is known, what we could not verify, and what each assumption would cost if wrong.
• A prioritised queue where the work compared multiple candidates, with the ranking criteria made explicit so it can be re-run as facts change.
• A route into delivery if the decision is to build: what Define would need to settle, sized realistically, whether or not we are the ones who build it.

And behind all of it, people you can name: this work is led personally by Ballista's founders, whose names go on the recommendation. You can see who they are before you write to us.

The four tests your decision passes

Before any recommendation hardens, it has to pass four tests. Is the commercial value real: would solving this problem be worth enough, to people who can be named, for the investment to make sense? Is it technically feasible: can the required capability be built or integrated dependably within your constraints, or does it rely on optimism? Does the operational data exist: can it be reached and trusted, and is it richer in the slide deck than in the database? And is the adoption risk priced: are the people who must use the change positioned to understand, shape and accept it?

A recommendation that fails one of these tests is not softened. It is changed, or it is stopped.

When the recommendation is to pilot something, two further things are fixed before any pilot starts. Success criteria and stop conditions are agreed in advance: what result would justify continuing, what result would end the experiment, and who decides. A pilot without predefined criteria cannot succeed, because nothing counts as success; it can only fade. And the road to production is priced into the decision: most of the work between a promising pilot and a dependable capability is data engineering, integration, governance and measurement, and a recommendation that prices only the pilot is a down payment disguised as a plan.

The point of all this rigour is what it buys you: when a recommendation passes, you can commit to it wholeheartedly, take it to a board without flinching, and build on it knowing the ground holds.

Where it leads

Direction and diagnosis is the Advise and Define end of the Ballista journey. When the decision is to build, the same team that formed the evidence carries it into technology and product delivery, so nothing is lost in translation between the people who recommended and the people who deliver. When the gap is leadership rather than a project, fractional leadership continues the same accountability in a different form. When the decision is to stop, the engagement ends there, with a budget saved for a better opportunity. And when the decision is to go, you go with the evidence behind you and the team that built it already at your side.