How Marketing Agencies Should Prepare for London Tech Week's Infrastructure Signal in UK

Direct answer: London Tech Week's 2026 programming signaled that the market is moving past chasing individual AI product trends and toward investing in durable technology infrastructure. For UK marketing agencies, that means the next competitive edge won't come from adopting one more AI tool for client campaigns — it will come from the operational systems, data pipelines, and automation running underneath your own agency, and the credibility you can demonstrate when clients ask how you actually work.

The specific signal comes from Republic Europe's coverage of London Tech Week insights, published in August 2026, which described a broader shift in the conversation away from short-term product trends and toward infrastructure-level technology investment. This is a meaningful pivot in tone. For the past two years, most public discussion around AI in the UK tech scene has centered on which tools to try, which chatbot to plug into a workflow, and which model to swap in next. Republic Europe's framing suggests the more serious players at London Tech Week 2026 were talking about something less flashy but more consequential: the plumbing that makes any of those tools reliable, secure, and repeatable at scale. That distinction matters enormously for a UK marketing agency, because agencies live and die by their ability to deliver consistent, high-quality output across many clients simultaneously — and consistency is an infrastructure problem, not a tool-selection problem. This post lays out what that signal actually means, why it should change how agency leaders in the UK are thinking about their own stack, and what a sensible first move looks like.

What Does "Infrastructure Over Trends" Actually Mean?

It helps to be precise about what's being contrasted here, because "infrastructure" can sound abstract next to something concrete like "AI tools."

A trend-driven approach to technology looks like this: a new AI writing assistant launches, your team tries it for a week, some people like it, it gets bolted onto whatever process already existed, and six months later half the team has quietly gone back to old habits because nothing about how the work actually flows has changed. The tool arrived, but the surrounding system — how a request enters the agency, how it gets routed, what data it touches, how the output gets checked and delivered — stayed exactly the same. You've adopted a feature, not a capability.

An infrastructure-driven approach looks different. It starts with the actual pipeline of work: how a client brief becomes a deliverable, where the bottlenecks and manual handoffs sit, and what data needs to move between systems for that pipeline to run without a person babysitting every step. Only after that pipeline is mapped does the question of "which AI tool" even come up — and often the answer is a combination of a few well-integrated agents and automations rather than a single flashy product. This is precisely the distinction Republic Europe's coverage of London Tech Week 2026 was pointing at: the market conversation has matured past "what's the newest AI thing" toward "what does the underlying system need to look like for AI to actually deliver reliable value."

This is a real, structural shift, not marketing language. It tracks with what's been visible in enterprise software generally over the past year: fewer standalone AI point-solutions being pitched as the answer, more emphasis on orchestration, data readiness, and the connective tissue between systems.

It's worth being honest about what we don't know here. A precise figure for how much UK investment is shifting from product-trend spending to infrastructure spending isn't publicly available for this specific angle — Republic Europe's coverage described the directional shift in the conversation at London Tech Week 2026, not a quantified market breakdown. What can be reasoned honestly from the general pattern is that when a market's most visible platform for showcasing new technology starts foregrounding infrastructure over point products, it usually reflects buyers and investors having already gone through a cycle of disappointment with shallow tool adoption. That disappointment is familiar to anyone who has watched a team enthusiastically adopt an AI writing tool, only to see usage taper off within a quarter because nothing about the surrounding workflow actually changed. The infrastructure framing is, in effect, the market naming that failure pattern and pointing toward the fix.

Why This Signal Matters Specifically to Marketing Agencies in the UK

It would be easy for a marketing agency to read "infrastructure investment" and assume it's a message meant for enterprise software vendors or fintech companies, not for a creative and media shop. That would be a mistake, for a few concrete reasons tied to how UK agencies actually operate.

Agencies Run on Repeatable Processes, Not One-Off Projects

Unlike a product company shipping a single application, a marketing agency's core asset is its ability to repeat a quality process across dozens of clients, campaigns, and reporting cycles every month. Client onboarding, campaign briefs, content approvals, reporting dashboards, media buying reconciliation — these are all recurring workflows with a predictable shape. That predictability is exactly what makes them suitable for infrastructure-level automation rather than one-off tool experiments. If your agency's internal systems are still a patchwork of spreadsheets, disconnected SaaS tools, and manual copy-paste between platforms, you are structurally behind the curve the London Tech Week conversation is describing — regardless of how many individual AI writing or image tools your team has bookmarked.

UK Clients Are Getting More Technically Literate

Marketing buyers in the UK — particularly at mid-market and enterprise companies — have spent the last two years hearing about AI constantly, in the press and from their own boards. Many are now asking sharper questions than "do you use AI," and more like "how is your reporting automated," "how do you handle our data across your systems," and "what happens if a key person on your team is out — does the process still run." An agency that can answer those questions with a real description of its own infrastructure has a credibility advantage in new-business conversations that a competitor relying on ad hoc tool usage simply cannot match.

The Competitive Set Is Widening

Independent UK agencies are no longer just competing against each other. They're competing against larger agencies with dedicated ops teams, against in-house marketing functions at client companies that have quietly built their own AI-assisted workflows, and against a growing pool of smaller, highly automated shops that can deliver agency-quality output with a fraction of the headcount because their internal infrastructure does more of the repetitive work. Infrastructure isn't a nice-to-have differentiator in that landscape — it's close to table stakes for holding margin.

The Cost of Standing Still Is Not Neutral

There's a temptation to treat this as an optional upgrade an agency can get to eventually, the way a team might eventually redesign its logo or refresh its brand deck. That framing understates the risk. Every quarter an agency runs its reporting, approvals, and client communication manually is a quarter where a competitor running the same functions on automated infrastructure is quietly pulling ahead on margin, turnaround speed, and the ability to take on more clients without proportionally more headcount. Because these gaps compound — a faster reporting cycle this quarter becomes a faster new-client onboarding process next quarter, which becomes capacity for another account next year — the agencies that treat this signal as urgent now will find the gap much harder to close by the time it becomes visible in lost pitches.

What Changes in Practice for Your Agency's Website, Stack, and Delivery

Translating "invest in infrastructure, not trends" into something an agency owner can actually act on means looking at three layers: internal operations, the client-facing website and product, and the vendor relationships that support both.

From Campaign Tools to Owned Infrastructure

Most agencies already run several AI-adjacent tools — a copywriting assistant, an image generator, maybe a reporting dashboard with some automated summarization. The infrastructure shift means moving from "a collection of tools people use inconsistently" to "a defined internal system that routes work automatically and only asks a human to step in at the right checkpoints." Concretely, this often starts with the unglamorous internal processes: how status reports get compiled, how approvals move between account managers and clients, how alerts fire when a campaign metric moves outside an expected range. We've written previously about how this plays out in practice in AI in Internal Tools: Automating Reports, Approvals, and Alerts, which walks through exactly the kind of internal automation that turns scattered tool usage into something closer to real infrastructure.

The Client Conversation Shifts Too

Once your internal systems are more automated and reliable, the story you can tell prospective clients changes. Instead of "we use several AI tools to speed up our work," you can describe an actual operating model: how a brief flows in, what gets automated, what a human reviews, and how reporting gets delivered without a 48-hour lag while someone manually pulls numbers together. That's a materially more convincing pitch in a market where, per the Republic Europe coverage of London Tech Week 2026, buyers and investors are increasingly skeptical of surface-level AI adoption and looking for evidence of real operational depth.

Your Website and Client Portal Need to Reflect the Same Standard

If your agency is telling clients a story about infrastructure and reliability, your own public-facing site and any client portal need to hold up to that same scrutiny — fast, well-structured, and functional rather than a template that hasn't been touched since 2022. This matters more than agencies sometimes assume: prospective clients form an impression of your technical competence from your own site before they ever hear your pitch. It's a similar dynamic to what specialist studios face — see how it plays out for a very different vertical in Website Development for Architecture and Interior Design Studios — where the site itself has to demonstrate the same standard of craft the business claims to deliver for clients. For a marketing agency, that standard now increasingly includes visible technical competence, not just design polish.

What UK Marketing Agencies Should Do About It Now

The temptation after reading a signal like this is to do something dramatic — a full stack overhaul, a company-wide AI initiative announcement, a slide deck for the next client pitch. Resist that. The agencies that will actually benefit from this shift are the ones that treat it as a sequence of small, deliberate infrastructure investments rather than one large disruptive project.

Start With One High-Friction Internal Process

Pick the single internal workflow that causes the most friction, delay, or error right now — client reporting is a common candidate, as is campaign approval routing or media spend reconciliation. Map it end to end: where does data enter, what systems does it touch, where do humans currently intervene manually, and where does it break down. Automating that one process, properly, with the right checkpoints for human review, tells you more about what infrastructure investment actually requires than any amount of reading about the trend.

Treat AI Agents as Infrastructure Components, Not Point Tools

The distinction London Tech Week's coverage is drawing matters here directly. An AI agent that sits inside a defined workflow — pulling campaign data, flagging anomalies, drafting a first-pass report, and routing it to the right person for review — is infrastructure. The same underlying model used as a standalone chatbot that a few team members remember to open occasionally is not. This is the core of what AI Agents & Automation is built to address: designing agents that are wired into your actual operational pipeline rather than bolted on as an extra tab in someone's browser.

Plan for the Stack You'll Need in 12–18 Months, Not Just This Quarter

Because infrastructure decisions are harder to unwind than tool subscriptions, it's worth thinking slightly further ahead when you make them. If your agency is likely to need a client-facing dashboard, a mobile-accessible reporting view, and an internal desktop tool for account managers, building each of those separately and disconnected is exactly the kind of fragmented approach the infrastructure shift is moving away from. There's a more durable pattern here, covered in Multi-Platform Software Strategy: Web, Mobile, and Desktop From One Codebase, where a single well-architected codebase serves multiple surfaces instead of each platform becoming its own maintenance burden.

Build the Review Checkpoints Before You Build the Automation

It's tempting to focus entirely on what gets automated and treat human review as an afterthought bolted on at the end. The more durable approach reverses that order: decide first where a person absolutely must check output before it reaches a client — a media spend reconciliation figure, a piece of client-facing copy, a strategic recommendation — and design the automation around those checkpoints from the start. Agencies that skip this step tend to either over-automate and ship an embarrassing error to a client, or lose confidence in the system after one mistake and quietly abandon it. Neither outcome is necessary if the checkpoints are designed deliberately rather than added reactively.

Don't Underestimate the Data Cleanup Step

Almost every agency that starts mapping a workflow for automation discovers the same thing: the data feeding that workflow is messier than anyone assumed. Client naming conventions differ across platforms, campaign IDs don't match between the ad platform and the internal tracker, and historical records have gaps from tools that were swapped out two years ago. None of this is a reason to delay — it's simply part of the work, and it's far better to surface it early, during the scoping of a single workflow, than to discover it halfway through building a larger system. Agencies that budget realistic time for this step tend to have a much smoother rollout than those that assume the automation layer can simply be dropped on top of existing data as-is.

Expect the Second Workflow to Move Faster Than the First

Agencies that go through this process once tend to find the second and third automated workflows considerably faster to build than the first. Much of the initial effort goes into things that get reused — understanding how client data is structured, setting conventions for where human review sits, establishing how the agency wants alerts and notifications to behave. Once that groundwork exists, extending automation to a second workflow, like campaign approval routing after reporting is already automated, is usually a smaller and more predictable project. This is a useful thing to communicate internally at the outset, since it's easy for a team to judge the whole effort by the time and friction of the first rollout alone.

How This Kind of Work Typically Gets Scoped

Agencies asking about this kind of infrastructure work often want a sense of scale before committing to a full engagement. The table below reflects how this typically maps to service tiers — treat it as a starting reference point, not a fixed quote, since actual scope depends on your existing systems and how many workflows you're automating.

Tier	Typical scope for a marketing agency	Investment
Essential	Automating one internal workflow (e.g., reporting or approvals) with a single AI agent	$1,000
Growth	Multiple connected agents across reporting, alerts, and client communication, plus a refreshed client-facing portal	$2,000
Enterprise	Full internal automation suite, multi-platform client dashboards, and ongoing infrastructure support	$4,000+

Key Takeaways

• London Tech Week 2026 coverage from Republic Europe points to a market-wide shift from chasing individual AI trends toward investing in durable infrastructure — a message that applies directly to how agencies run themselves, not just to enterprise software vendors.
• UK marketing agencies compete on process consistency, which makes internal infrastructure a direct lever on margin and quality, not a side project.
• Clients are asking sharper, more technical questions about how agencies actually operate — agencies with real automated systems can answer those questions credibly; agencies with a patchwork of tools cannot.
• The most effective first step is automating one high-friction internal workflow properly, rather than announcing a broad AI initiative with no operational backbone.
• AI agents deliver the most value when they're wired into a defined workflow with clear human checkpoints, not deployed as standalone chat tools nobody consistently uses.
• Client-facing infrastructure — your website, portal, and reporting surfaces — should reflect the same technical standard your pitch claims to deliver.

Getting infrastructure right is less about any single tool and more about how your agency's systems fit together end to end. If you want help mapping where automation would actually reduce friction in your operation, book a meeting with our team.

Frequently Asked Questions

What did London Tech Week 2026 actually say about infrastructure versus trends?

Coverage from Republic Europe on London Tech Week 2026 described a broader shift in the conversation away from short-term product trends and toward infrastructure-level technology investment. The emphasis was on durable systems and operational depth rather than which individual AI tool is newest.

Is this signal relevant to small UK marketing agencies, or only large ones?

It's arguably more relevant to smaller agencies, since they have less slack to absorb inefficient internal processes and stand to gain proportionally more from automating recurring workflows like reporting and approvals. Infrastructure investment scales down well because it targets specific friction points rather than requiring enterprise-level budgets.

What counts as "infrastructure" for a marketing agency specifically?

For an agency, infrastructure typically means the systems connecting client briefs, campaign data, reporting, and approvals — including any AI agents automating parts of that flow, the data pipelines feeding them, and the client-facing portals or dashboards built on top. It's the operational backbone, not any single tool.

How is this different from just using more AI tools?

Using more AI tools without connecting them to a defined workflow tends to produce inconsistent adoption — some team members use a tool, others don't, and nothing structural changes. Infrastructure investment means designing the actual process so that automation happens by default, with clear points where a human reviews or approves output.

Why would a client care whether my agency has good internal infrastructure?

Clients increasingly ask how deliverables get produced, not just what the deliverables look like, because reliability and turnaround time depend directly on your internal systems. An agency that can describe a real operating model gives a client more confidence than one that describes a list of tools it occasionally uses.

What's the first workflow most UK agencies should automate?

Client reporting is the most common starting point because it's recurring, time-consuming, and has a predictable data source, making it well suited to automation with clear checkpoints for human review before anything goes to a client.

Does this mean agencies should stop trying new AI tools altogether?

No — it means new tools should be evaluated by whether they can be wired into an existing workflow with real business impact, rather than adopted because they're new. Trend-chasing and infrastructure investment aren't mutually exclusive, but infrastructure should come first.

How long does it typically take to automate one internal workflow?

It depends on how well-defined the existing process is and how many systems it touches, but a single focused workflow — such as automated reporting with alert thresholds — is typically a matter of weeks once requirements and data sources are mapped, not months.

What's the difference between an AI agent and a chatbot in this context?

A chatbot is a standalone interface a person has to remember to open and use. An AI agent, as used in agency infrastructure, is embedded in a defined workflow — pulling data, taking action, and flagging exceptions automatically — without requiring someone to initiate each interaction.

Should agencies build this infrastructure in-house or bring in outside help?

It depends on whether the agency has existing engineering capacity. Many agencies find it faster and more reliable to bring in a partner experienced in AI Agents & Automation for the initial build, then maintain and extend it internally once the core system is running.

What data privacy considerations apply when automating client reporting?

Any automation touching client campaign data should respect the same data-handling agreements already in place with clients, including access controls and retention limits. This is worth clarifying explicitly with any automation partner before connecting live client data sources.

How does this affect how agencies should structure new-business pitches?

Pitches can now credibly include a description of the agency's actual operating model — how briefs move through the system, what's automated, and where humans check quality — rather than a general claim of "we use AI." That specificity tends to land better with technically literate buyers.

Is a client portal necessary, or can reporting stay in email and spreadsheets?

Email and spreadsheets can work at small scale, but they don't hold up as an agency grows its client roster, and they don't communicate the same operational credibility a real-time portal does. A portal also removes a recurring manual step from the reporting workflow itself.

What's a realistic budget range for a first automation project?

For a single internal workflow like automated reporting or approvals, this typically falls under an Essential-tier engagement around $1,000. Broader efforts spanning multiple workflows and a client-facing portal move into Growth-tier territory around $2,000.

How does multi-platform strategy relate to this infrastructure shift?

If an agency needs its automation and reporting accessible on web, mobile, and potentially desktop, building each separately multiplies maintenance cost. A single well-architected codebase serving all three surfaces is more consistent with the infrastructure-first approach the market is moving toward.

What happens if an agency ignores this shift entirely?

Nothing happens immediately, but the agency risks falling behind competitors who can deliver faster turnaround, more consistent quality, and a more credible technical story to prospective clients — a gap that tends to compound over a year or two rather than appear suddenly.

Does this apply equally to agencies focused on paid media, content, and creative work?

Yes, though the specific workflows differ. Paid media agencies benefit most from automated spend reconciliation and anomaly alerts; content and creative agencies benefit most from automated approval routing and asset delivery tracking. The underlying infrastructure principle is the same across all of them.

How do I know if my agency's current stack is "infrastructure" or just "tools"?

A useful test: if a key team member left tomorrow, would the workflows they touched keep running with minimal disruption? If the answer is no because too much depends on manual habits and tribal knowledge, the stack is tools, not infrastructure.

What's the risk of moving too fast on infrastructure changes?

The main risk is disrupting live client workflows by changing systems without adequate testing or a rollback plan. This is why starting with one contained workflow, rather than a full stack overhaul, is the safer sequence.

Can smaller agencies compete with larger ones purely through better infrastructure?

To a meaningful degree, yes. Automation reduces the per-client overhead that would otherwise require additional headcount, letting smaller agencies deliver comparable consistency and turnaround without matching a larger competitor's staff size.

How does this connect to AI guardrails and compliance concerns in the UK?

As agencies automate more of their client-facing output, having clear review checkpoints becomes more important, not less, particularly for regulated client sectors. Infrastructure investment should include defining where human sign-off is mandatory before anything reaches a client.

What role does reporting automation play in client retention?

Faster, more consistent, error-free reporting is one of the more visible ways clients experience an agency's operational quality month to month. Automating it reduces the chance of the kind of reporting delays or mistakes that quietly erode client confidence over time.

Is this trend specific to London, or does it apply UK-wide?

The signal came out of London Tech Week specifically, but the underlying shift toward infrastructure investment reflects a broader pattern across the UK tech and services market, not something confined to London-based agencies.

What's a reasonable way to measure ROI on an infrastructure investment?

Track the time previously spent on the manual version of the automated task, the error or rework rate before and after, and any change in client-reported turnaround time. These are more meaningful measures than trying to attribute new business directly to the infrastructure change alone.

Do AI agents replace account managers or reduce headcount need?

The realistic outcome is that agents absorb repetitive, low-judgment tasks — data pulls, first-draft reports, status checks — freeing account managers to spend more time on client strategy and relationship work rather than administrative tasks.

What should an agency ask a vendor before starting an automation project?

Ask how the vendor handles review checkpoints for anything client-facing, how they scope a workflow before building it, what happens if the process needs to change later, and whether the resulting system will be something your team can maintain independently.

How does this affect agencies serving international clients from the UK?

International clients often have even higher expectations around reporting transparency and turnaround consistency, since time zone gaps reduce opportunities for quick manual clarification. Reliable automated infrastructure matters more, not less, in that context.

What's the difference between Essential, Growth, and Enterprise tiers for this kind of work?

Essential typically covers a single automated workflow, Growth covers multiple connected agents plus a refreshed client-facing portal, and Enterprise covers a full automation suite with multi-platform dashboards and ongoing support. The right tier depends on how many workflows need attention and how far client-facing systems need to go.

Should agencies automate client communication, or just internal processes?

Internal processes are the safer and more common starting point because errors are caught before reaching a client. Client-facing communication automation is worth pursuing once internal workflows are stable and the review checkpoints for external content are well defined.

How often should an agency revisit its infrastructure once it's built?

Treat it as a living system rather than a one-time project — revisiting it roughly every two to three quarters to check whether new client requirements, team changes, or tool updates have introduced new friction points worth automating.

What's the biggest mistake agencies make when trying to modernize their stack?

The most common mistake is buying multiple disconnected AI tools in response to hype rather than mapping the actual workflow first and automating it end to end. This produces the exact "trend-chasing" pattern the London Tech Week coverage described as outdated.

Can this infrastructure work be done gradually alongside normal client delivery?

Yes, and it generally should be. Automating one workflow at a time while client delivery continues as normal is both lower risk and easier to validate than pausing operations for a larger overhaul.

What technical skills does an agency team need to maintain this kind of system?

Day-to-day maintenance of a well-built automation system generally requires less specialized skill than the initial build, since the system is designed with clear monitoring and simple intervention points. Most agencies can maintain it with existing operations staff after a proper handover.

How does this relate to AI visibility and how agencies show up in AI-generated search results?

Operational credibility and a well-structured, fast website both support how confidently an agency's expertise gets represented when AI tools summarize or cite it, since thin or generic sites give search and AI systems less substantive material to draw from.

What's a sign that a marketing agency has genuinely adopted infrastructure thinking?

A clear sign is being able to describe, in specific terms, how a client brief moves from intake to delivery, including exactly which steps are automated and where a human checks the output — rather than a vague reference to "using AI."

Is it worth building infrastructure before winning more clients, or after?

Building it while client volume is still moderate is generally easier, since there's more room to test and adjust without disrupting a large active client base. Waiting until growth forces the issue often means implementing under more pressure.

How does automated alerting fit into agency reporting workflows?

Automated alerts flag when a campaign metric moves outside an expected range, allowing an account manager to respond proactively rather than waiting for a scheduled report to surface an issue days later. This is one of the more immediately valuable automations for client retention.

What happens to agency pricing models as infrastructure improves?

Some agencies use efficiency gains to improve margin on existing pricing, while others use the improved turnaround and consistency to justify premium pricing for retained services. Which approach fits depends on the competitive position of the individual agency.

Does this shift affect freelancers and very small agency teams?

It affects them differently — a solo operator or very small team benefits most from automating the handful of repetitive tasks that otherwise consume disproportionate time, since there's no larger team to absorb the manual workload.

What's the relationship between this infrastructure trend and AI agent frameworks in general?

The broader move toward orchestration and connected systems across the software industry mirrors what London Tech Week's coverage described for the market generally — less focus on any single agent framework, more focus on how agents fit into a working operational pipeline.

How should an agency handle a client who's skeptical of AI-driven processes?

Being specific and transparent about where automation is used, what a human reviews before anything ships, and how errors get caught tends to address skepticism better than avoiding the topic. Most client concerns are really about quality control, not the technology itself.

What's the realistic timeline from starting this work to seeing measurable results?

For a single automated workflow, most agencies see measurable time savings within four to eight weeks of launch, since the baseline manual process is usually well understood and easy to compare against.

Can this kind of infrastructure work be undone if it doesn't work out?

A well-scoped, modular automation — built around one workflow with clear boundaries — can be adjusted or rolled back without disrupting the rest of the agency's operations, which is another reason to avoid large all-at-once overhauls.

How does website performance factor into the infrastructure conversation?

A slow, outdated agency website undercuts the same credibility that automated internal systems are meant to build, since prospective clients form an impression of technical competence from the site itself before any pitch happens.

What's the difference between automation and full AI decision-making in this context?

Most sensible agency automation keeps a human in the loop for judgment calls — creative approval, strategic recommendations, client-facing tone — while automating the mechanical steps around data collection, formatting, and routing. Full autonomous decision-making is rarely the right target for client-facing work.

Should agencies mention this infrastructure work publicly, like in a blog post or case study?

Once a workflow is automated and delivering measurable results, describing it specifically — what changed, what improved, without overstating it — tends to be more persuasive to prospective clients than general claims about AI adoption.

How does this affect agencies that already outsource development work?

Existing outsourcing relationships are a reasonable starting point for this work, provided the vendor understands agency-specific workflows like campaign reporting and client approval cycles rather than only generic software development.

What's the connection between infrastructure investment and agency scalability?

Automated, well-connected internal systems mean an agency can take on additional client volume without a proportional increase in operational headcount, which is the core mechanism by which infrastructure investment supports scaling.

Is there a risk of over-automating and losing the personal touch clients value?

Yes, if automation replaces client-facing judgment and communication rather than the mechanical steps behind it. The safer approach automates data handling and reporting mechanics while keeping strategic and relationship-facing work in human hands.

What should an agency do in the next 30 days in response to this signal?

Identify the single internal workflow causing the most friction, document its current manual steps, and get a scoped estimate for automating it — a concrete, bounded first step rather than a broad strategic pivot.