At the height of Britain’s AI startup frenzy, Contextly looked like it was everywhere.

The London-based company appeared constantly in founder circles throughout 2024 — on LinkedIn, at Shoreditch demo nights, inside “top AI startups to watch” lists, and across the increasingly crowded ecosystem of podcasts, launch threads, and AI newsletters that emerged after the generative AI boom reshaped venture markets.

From the outside, it resembled a familiar post OpenAI success story:

  • fast growth,
  • visible founder branding,
  • aggressive hiring,
  • AI-native positioning,
  • investor interest.

Internally, according to multiple people familiar with the business during that period, the reality was significantly less stable.

By early 2025, the company had entered what one former employee described as “full survival mode”.

The rise of London’s AI gold rush

Contextly launched in late 2023, during one of the most overheated periods in recent UK startup history.

Investor appetite for AI businesses had surged dramatically following the mainstream adoption of large language models. According to Dealroom (2026), UK AI startups raised more than £4.2 billion in venture capital during 2025, cementing Britain as Europe’s largest AI funding market outside the United States.

That funding boom reshaped East London almost overnight.

Suddenly every coworking space around Old Street seemed to contain founders building:

  • AI assistants,
  • AI workflow tools,
  • AI search layers,
  • AI productivity platforms,
  • AI copilots.

Contextly positioned itself directly inside that momentum.

Archived product material and investor-facing documents reviewed by EP+ describe the startup as an “AI-native operating layer for knowledge work”, combining summarisation, meeting intelligence, workflow automation, and collaborative research tooling into a single platform aimed primarily at consumers and small teams.

The market responded quickly.

Public launch campaigns generated substantial engagement online, while archived hiring listings suggest the company expanded from fewer than five employees to nearly 20 within its first year.

At the time, rapid expansion itself was often interpreted as validation.

That assumption would later become problematic.

Growth without retention

One of the least discussed realities of the AI startup cycle is how difficult it has been to convert consumer interest into durable behaviour.

Usage spikes have frequently been mistaken for sustainable businesses.

Internal metrics referenced in investor updates reviewed by EP+ suggest Contextly experienced rapid early adoption, reportedly surpassing 100,000 users within months of launch. Conversion rates, however, remained weak relative to infrastructure costs.

Several people familiar with the company described a widening gap between public visibility and underlying commercial performance.

“The attention was real,” one former contractor said. “The retention wasn’t.”

This pattern has become increasingly common across Europe’s consumer AI market.

According to PitchBook (2026), enterprise-focused AI startups have materially outperformed consumer AI businesses on revenue stability, retention, and investor confidence over the past 18 months.

The underlying economics are partly structural.

Unlike traditional SaaS companies, many generative AI startups face unusually high operating costs linked to inference, API usage, cloud compute, and model-serving infrastructure. In practice, growth can increase losses if monetisation does not scale proportionally.

By late 2024, according to two individuals familiar with the company’s operations, infrastructure expenses had become a growing internal concern at Contextly.

The hidden pressure behind AI infrastructure costs

Much of the public conversation around AI startups still focuses on product capability.

Far less attention is given to operational economics.

Several former employees described periods of intense internal focus on cloud-cost optimisation, including prompt restructuring, usage restrictions, and efforts to reduce inference expenses tied to customer activity.

One individual familiar with the company’s infrastructure strategy described engineering priorities during that period as “increasingly financial rather than product-led”.

Industry-wide data suggests Contextly was unlikely to be alone.

Analysis published by Andreessen Horowitz in 2025 noted that many early-stage AI startups were operating with gross margins significantly below traditional SaaS benchmarks due to dependency on third-party model infrastructure.

At the same time, investor expectations were changing rapidly.

The market that rewarded narrative momentum in early 2024 became considerably more disciplined by the beginning of 2025.

The funding environment turns colder

During the early generative AI boom, startups often raised capital primarily on future potential.

By 2025, investors increasingly demanded evidence of:

  • monetisation,
  • retention,
  • defensibility,
  • operational efficiency.

Particularly in London’s AI ecosystem, investors had begun seeing near-identical pitches repeatedly.

“Every deck started blending together,” one London venture investor told EP+ separately while discussing the broader market. “AI workflow layer, productivity automation, enterprise copilot — eventually investors stopped rewarding positioning alone.”

Beauhurst (2026) data indicates bridge rounds and internal financings among UK AI startups increased significantly throughout 2025 as fundraising conditions tightened.

People familiar with Contextly’s operations during that period described increasing concern internally about runway and future financing prospects.

According to financial planning material reviewed by EP+, the company’s projected cash position had deteriorated substantially by early 2025.

One former employee described morale inside the business as “visibly anxious”.

Another characterised the atmosphere more bluntly:

“Everyone externally thought the company was scaling. Internally, people were trying to work out whether layoffs were coming.”

The enterprise pivot

The company’s eventual shift toward enterprise software appears to have emerged less from long-term strategic planning than commercial necessity.

According to two individuals familiar with the process, early interest from a professional services firm using Contextly informally inside its operations triggered internal discussions around enterprise deployment opportunities.

The requirements differed substantially from the startup’s original consumer-oriented product strategy.

Instead of lightweight productivity tooling, enterprise customers demanded:

  • permissions systems,
  • deployment controls,
  • compliance functionality,
  • audit trails,
  • integration capabilities.

That transition required significant internal restructuring.

Archived product roadmaps and hiring patterns suggest consumer-facing development slowed considerably during the first half of 2025 while enterprise functionality accelerated.

Several staff departures followed.

People familiar with the company described the period as operationally chaotic, with shifting priorities and uncertainty around long-term direction.

“There’s a tendency in startup culture to romanticise pivots,” one former employee said. “This didn’t feel romantic. It felt reactive.”

The contract that changed the company’s trajectory

By mid-2025, the strategy began showing signs of commercial viability.

Documents reviewed by EP+ suggest Contextly secured a sizeable enterprise agreement with a UK-based professional services organisation during Q2 2025, materially improving the company’s revenue outlook and runway position.

The economics of the business reportedly changed quickly afterwards.

Enterprise clients offered:

  • larger contract values,
  • lower churn,
  • more predictable usage,
  • improved revenue visibility.

More importantly, the company finally appeared to possess a clearer monetisation pathway.

The shift mirrors a broader trend across Britain’s AI startup market.

According to Dealroom (2026), enterprise AI startups in the UK increasingly account for a disproportionate share of later-stage funding activity as investors prioritise sustainable revenue models over purely consumer adoption metrics.

In practical terms, the market has become less interested in viral growth — and more interested in durability.

A more difficult phase for UK AI startups

Contextly’s trajectory reflects a wider recalibration taking place across London’s startup ecosystem.

The first wave of AI enthusiasm rewarded visibility, speed, and narrative positioning. The next phase appears considerably less forgiving.

Investors are scrutinising:

  • margins,
  • infrastructure exposure,
  • retention,
  • pricing power,
  • enterprise defensibility.

Several venture investors speaking broadly about the sector described increasing fatigue around undifferentiated consumer AI products, particularly those dependent on external foundation models without meaningful proprietary advantages.

That does not necessarily mean the market is collapsing.

It does suggest expectations are changing.

The companies most likely to survive Britain’s AI reset may not be the loudest or most visible. Increasingly, they appear to be the ones capable of converting technical excitement into financially sustainable operations.

From the outside, Contextly now appears considerably quieter than during its peak visibility phase in 2024.

According to current company materials and hiring activity reviewed by EP+, the business remains active, with a smaller operational footprint and a clearer enterprise focus.

In many ways, that may represent a healthier version of growth than the one that initially attracted attention.