aerostack

Search Aerostack

Norwegian

Rydoo Accounts Payable AI invoice automation at Norwegian

MidAmerica St. Louis

GoodMaps indoor navigation at MidAmerica St. Louis Airport

Lyon Saint-Exupéry Airport

Stan robotic valet parking at Lyon Saint-Exupéry

London Heathrow

Facial-recognition eGates at Heathrow

Almaty International Airport

Q-Gate biometric passport-control terminals at Almaty

Pittsburgh International Airport

Arrow Analytics EZ-Board computer-vision carry-on analytics test at PIT Gate A7

Federal Aviation Administration

FAA SMART airspace and trajectory management platform

Dubai International

Dubai Police AI systems for passenger screening, unattended baggage and camera analysis at Dubai airports

Airport intelligence / London Heathrow

Heathrow: where AI enters airport operations

From aircraft turns to passenger connections and runway visibility, Heathrow illustrates a shift from collecting operational data to using it to support decisions. But public evidence of deployment is not the same as proof of better performance.

London Heathrow Airport Terminal 5 in October 2025
London Heathrow Airport Terminal 5 in October 2025. Photograph: CedarGlen86 / CC0 1.0 (cropped to fit). Shows the airport, not an AI implementation.
The picture at Heathrow

Three different problems. One common ambition.

Airports rarely suffer from a shortage of information. The harder problem is recognising what matters early enough to act. Heathrow's reported AI projects reveal three approaches to that challenge: watching ground activity as it unfolds, forecasting passenger demand, and interpreting runway images when direct visibility is poor.

They should not be treated as equally mature. Computer vision is reported in use at Terminal 5 stands; passenger forecasting has a documented implementation history but an uncertain present-day configuration; runway-clearance monitoring remains a trial in the evidence reviewed. None of those descriptions, alone, proves that the airport is operating more efficiently.

The distinction matters: Aerostack reports what the sources establish, separates proposed benefits from demonstrated results, and keeps the original material open to inspection.

Selected implementations

Where the operational interest lies

01 / Deployed

When the turnaround becomes visible

The operational problem

A late turnaround can cascade into the next departure. The difficulty is knowing which task is holding up the aircraft while there is still time to intervene.

What the sources report

Assaia says its ApronAI computer-vision system has been installed with British Airways at 17 Terminal 5 stands, using 52 cameras to identify activity including baggage handling, refuelling and boarding. A larger rollout was announced, but the number of additional gates actually operating has not been disclosed.

Why it matters

The case for an operations manager is earlier visibility across competing ground-handling tasks. Visibility alone, however, is not evidence that delay minutes or stand occupancy have improved.

What remains unproven

No Heathrow-specific turnaround performance improvement attributable to ApronAI is recorded in the sources reviewed.

Sources and verification (3)

mediacentre.heathrow.com · Evidence grade A

Assaia newsroom · Evidence grade B

Runway Girl Network · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Full implementation record →

02 / Deployed

Forecasting the passenger before the queue forms

The operational problem

Transfer passengers and security demand do not arrive evenly. Staffing and connection interventions become harder when forecasts rely only on historic averages.

What the sources report

A UCL impact case study says forecasting models developed with Heathrow were implemented in 2017, estimating connection times and passenger arrivals so operational teams could adjust resourcing and warn of connection risks.

Why it matters

This is a practical example of AI supporting an operational decision, rather than replacing it: forecasts inform staff allocation and intervention.

What remains unproven

UCL reported back-tested cost reductions of 12–54% against a legacy approach, not independently measured live savings. Continued use in the original form has not been established.

Sources and verification (1)

UCL REF 2021 impact case study · Evidence grade A

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Full implementation record →

03 / Pilot

An extra pair of eyes when the tower is in cloud

The operational problem

Low cloud can reduce controllers' ability to see whether an arriving aircraft has cleared the runway, constraining capacity.

What the sources report

NATS and Searidge tested the AIMEE camera-based AI system in Heathrow's Digital Tower Laboratory. It aims to indicate runway clearance, with the controller retaining the operational decision. EUROCONTROL described non-operational trials involving more than 50,000 arrivals.

Why it matters

For airport leaders, the prospect is capacity resilience in poor visibility. EUROCONTROL discussed recovering up to 20% of capacity lost in low cloud as a potential, not an achieved result.

What remains unproven

The available records do not establish an operational go-live, regulatory approval or measured capacity recovery.

Sources and verification (4)

EUROCONTROL · Evidence grade A

NATS / Searidge press release · Evidence grade A

Euronews · Evidence grade B

Searidge (vendor) · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

The wider register

Other activity at Heathrow

These records provide useful context but have not been selected as the three principal operational stories. Some represent agreements, infrastructure programmes or technologies whose AI-specific role needs clearer evidence. Listing them is not an endorsement of their claimed outcomes.

Facial-recognition eGates at Heathrow · Deployed · limited scope

Heathrow is one of the UK airports with facial-recognition eGates, according to GOV.UK and CN Traveler, which reported it on 20 May 2026.

From 6 October 2026 the Home Office began switching eGates to contactless mode, starting at East Midlands Airport with British citizens. Every eGate was due to be upgraded by early 2027, and no date had been given for Heathrow, the Home Office said on 6 October 2026.

Sources and verification (2)

GOV.UK / Home Office and Border Force · Evidence grade A

Condé Nast Traveler · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Open the full record →

Heathrow selects AIRHART to replace its airport operational database · Contracted

Heathrow has selected AIRHART, the platform of Smarter Airports, as the new base for its core operational systems. Netcompany announced the selection on 24 February 2026 and will support implementation. Heathrow Chief Operating Officer Javier Echave is quoted backing the programme, which Netcompany describes as a multi-year agreement.

Netcompany lists three capabilities in scope. A next-generation airport operational database, or AODB, the central record of flight and resource data that other airport systems read from, will replace Heathrow's existing systems. Alongside it come Airport Collaborative Decision Making, the A-CDM process through which the airport, airlines, ground handlers and air traffic control share turnaround milestones, and an Airport Operations Plan that Netcompany calls predictive and continuously optimised.

On AI, the published material stays general. Netcompany refers to "AI-driven decision support" and Future Travel Experience, in March 2026, to "predictive AI" for orchestrating operations, but neither names the models, the data they would be trained on or where their output would reach Heathrow staff. FTE reports that Heathrow joins Copenhagen Airport and Munich Airport in adopting this kind of orchestration platform.

Migration from the current systems is the next stage. Heathrow and its suppliers have not given a go-live date or said how the migration will be sequenced, so the record stays at contracted.

Sources and verification (3)

netcompany.com · Evidence grade A

Future Travel Experience · Evidence grade B

Future Travel Experience · Evidence grade C

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Open the full record →

Smiths Detection HI-SCAN 6040 CTiX scanners across all Heathrow terminals · Deployed

Passenger security lanes in all four Heathrow terminals have been replaced with CT scanners, Heathrow announced on 23 January 2026 at the close of a £1bn upgrade. Smiths Detection confirmed on 30 January 2026 that the scanner is its HI-SCAN 6040 CTiX. Heathrow calls itself the largest airport in the world to fully roll out next-generation CT security scanners, and Smiths makes the same claim.

Computed tomography builds a three-dimensional image of each bag instead of the flat views produced by conventional X-ray, and Smiths pairs that volumetric image with automatic threat recognition software intended to reduce manual bag checks. Smiths says the scanners meet ECAC's C3 standard, which it describes as the highest performance standard for explosives detection systems for cabin baggage. For passengers, Heathrow says liquids and electronics can stay in hand luggage, and Smiths specifies liquids of up to 2 litres.

Heathrow chose the equipment in March 2021. Smiths' announcement of that selection also covered iLane automatic tray-return systems, its Checkpoint.Evo plus lane-management software and a ten-year equipment service agreement.

With every terminal converted, the open questions concern performance. Lane throughput and alarm rates for the new lanes have not been published by either company.

Sources and verification (4)

mediacentre.heathrow.com · Evidence grade A

Smiths Group news · Evidence grade B

Smiths Detection press release · Evidence grade B

smithsdetection.com · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Open the full record →

Oracle Fusion Data Intelligence analytics on Heathrow's finance and HR systems · Deployed

Heathrow's finance, HR, supply chain and asset management functions run on Oracle Fusion Cloud applications, and the airport has added Oracle Fusion Data Intelligence on top of the ERP and HCM data they hold. Fusion Cloud went live in January 2022, Capgemini announced on 24 January 2022, in a programme Oracle's blog calls Project Magenta. Oracle's press release of May 2026 says Heathrow uses the analytics layer for governed access to sensitive data and embedded AI and machine-learning features.

Procurement gives the most concrete example. According to an Oracle Analytics blog of September 2025, Heathrow used OCI Document Understanding, Oracle's document-extraction service, to code invoices automatically, so that dashboards could flag errors and discrepancies that staff had previously chased by hand, such as invoices that did not match ERP transactions or carried wrong VAT numbers. The same blog describes HR predictive analytics run over a decade of workforce records, with salary history, leave of absence and department transfers emerging as indicators of staff likely to leave.

Oracle sets the work against Heathrow's size: almost 85 million passengers last year, over 90,000 people employed on site and more than £200 billion a year in trade, by its May 2026 figures. Its September 2025 blog gives more than 83 million passengers and £200 billion of trade in 2024. These are Oracle's numbers, not measures of what the software has achieved.

Alan Petrie, who now leads corporate data and analytics at Heathrow, told Oracle in 2025 that early pilots of Oracle's AI assistant for FDI pointed towards staff querying data in plain language. In a second Oracle Analytics blog interview, published on 13 October 2026, he says AI agents are "being embedded in many business processes" in Fusion Applications, answering staff questions such as how many days' leave they have left or what Heathrow's maternity policy is. Most of the rest of that interview sets out plans, not systems already in use.

The same post says Petrie's team recently won the EMEA Oracle Excellence Award for Data Analytics and AI; both the award and the announcement are Oracle's. No independent assessment of finance or HR outcomes has been published. Because FDI inherits the Fusion security model, it also sets the access rules for finance and HR data that other AI work at the airport would draw on. This is Aerostack's interpretation; Heathrow has not said so.

Sources and verification (4)

blogs.oracle.com · Evidence grade B

Oracle press release · Evidence grade B

Oracle Analytics blog · Evidence grade B

capgemini.com · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

British Airways Mission Control, an Emu Analytics digital twin for Heathrow ground operations · Deployed

Inside British Airways' Integrated Operations Centre at Heathrow, controllers work from Emu Analytics' Flo.w software, known internally as Mission Control. Emu says it merges real-time aircraft movement data with BA's AODB, ADS-B feeds, ground-vehicle telematics and crew systems into one view of the airside operation. Richard Anderson, BA's director of global operations, told Future Travel Experience in May 2025 that it tracks every aircraft movement by the second, which BA uses to find delay hotspots and to plan tows between its engineering base at the east end of the airport and its customer operation at the west.

The only figure from outside the vendor is OAG's. In June 2025 OAG reported that BA's share of Heathrow flights delayed by at least 60 minutes had fallen from 12% earlier that year to 8%, and linked the improvement to the digital twin and towing work. OAG's post does not say how much of the change it credits to the tool.

Emu's case study, accessed in October 2026, says Mission Control has contributed more than $20 million in annual savings and over 200,000 minutes of delay avoided, across BA's hub and international operation rather than Heathrow alone. It costs four pathways that together come to about $3.94 million a year; the largest, $1.42 million for stand and turnaround work, multiplies roughly 2,800 interventions by four minutes saved, a EUROCONTROL benchmark of $127 per minute and a 50% share credited to Flo.w. Emu also reports a reduction of about 25–30% in tug idle time.

All of the savings figures are Emu's. BA has not published its own, and no next phase has been announced.

Sources and verification (4)

oag.com · Evidence grade B

Future Travel Experience · Evidence grade B

Emu Analytics · Evidence grade B

emu-analytics.com · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

BEUMER baggage system and Flow Twin planned for Heathrow Terminal 2 · Reported · unverified

BEUMER said on 11 January 2024 that it had the contract for Heathrow's new Terminal 2 baggage system, designed for 31,000 bags a day.

BEUMER said the Terminal 2 upgrade would soon use its Flow Twin digital twin, Airport World reported on 22 April 2025. Heathrow said on 19 December 2025 that construction would start in 2026.

Sources and verification (3)

mediacentre.heathrow.com · Evidence grade A

Airport World (ACI) · Evidence grade B

beumergroup.com · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Hallie, Heathrow's Salesforce Agentforce passenger agent on WhatsApp · Deployed

Passengers who message Heathrow on WhatsApp reach Hallie, an AI agent built on Salesforce Agentforce. Peter Burns, Heathrow's director of digital, e-commerce and marketing, told Business Insider in May 2026 that it launched in March 2025. It answers questions on flights, gate directions, amenities and security queue lengths from Heathrow's Service Cloud knowledge base and live data, and Burns said it does not search the open web.

Salesforce's June 2025 release reports a 90% chat resolution rate without transfer to a human agent and more than 50% fewer chat exchanges to reach an answer. It also reports a 40% improvement in digital contact efficiency, while saying in the same release that Heathrow "expects to improve" that efficiency by up to 40%, so the figure is Salesforce-reported and partly forward-looking. Burns told Business Insider that phone calls fell from 70% of customer enquiries before Hallie to 10% by March 2026.

The two sources disagree on the size of the knowledge base behind the agent. Salesforce cites 800 knowledge articles; Business Insider reports a database of about 500. Aerostack publishes both rather than choosing one.

Burns said Hallie would reach Heathrow's website and app later in 2026, and that kiosks in the terminals are being considered, with no plan announced. No independent measure of answer accuracy or passenger satisfaction has been published.

Sources and verification (2)

Business Insider · Evidence grade B

Salesforce press release · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

BigBear.ai master service agreement with Heathrow for potential AI projects · Contracted

BigBear.ai announced on 2 August 2024 that it had signed a master service agreement with Heathrow "to develop and deliver advanced technologies" at the airport. The company had referred to the agreement a day earlier in exhibits to an 8-K filing with the US Securities and Exchange Commission dated 1 August 2024.

A master service agreement sets terms for future pieces of work; it does not by itself commit either side to a project. BigBear.ai's release says this one "allows for a wide range of potential projects" aimed at security, operational effectiveness and the traveller experience, and points to its Vision AI portfolio, including image-based anomaly and threat detection, alongside advanced analytics and digital twin capabilities.

No project, site or go-live under the agreement had been announced by BigBear.ai or Heathrow as of 15 October 2026, and neither has published a timetable for one. All descriptions of scope come from BigBear.ai.

Sources and verification (3)

BigBear.ai press release · Evidence grade B

BigBear.ai press release (Business Wire) · Evidence grade B

SEC (EDGAR) · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Here to Help staff assistant built in Microsoft Copilot Studio · Deployed

Under Heathrow's Here to Help programme, staff from non-operational roles spend time in the terminals answering passenger questions. In a blog post of 31 May 2024, Microsoft described as live a custom assistant built for them in Copilot Studio, its low-code tool for building chat assistants on sources an organisation chooses. It runs inside Microsoft Teams and draws its answers from Heathrow's public website, internal FAQ documents and bus timetables.

The users are occasional. Microsoft says Here to Help colleagues "may only work in the terminals 4 or 5 times a year", and presents the assistant as a replacement for manual searches and calls to other colleagues.

Two and a half years on, that blog remains the only public source. Nothing published since confirms the assistant is still in use. Heathrow and Microsoft have released no figures on answer accuracy or staff uptake.

Sources and verification (2)

Microsoft Copilot Studio blog · Evidence grade B

Microsoft Copilot blog · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Heathrow passenger-flow forecasting on the Databricks Data + AI Platform · Deployed

Heathrow forecasts passenger flow through the airport on the Databricks Data + AI Platform, running on Microsoft Azure. Databricks announced the work in September 2024 and said the airport relies on over 26TB of data covering passenger numbers, flights, airlines, seasonal performance, traffic and weather.

Within a year of the move, Databricks said, the time to produce forecast insight fell from two weeks to four hours and the margin of error from 30% to 10%. Its customer page adds that the job once needed two people and now needs one. The same page says Heathrow treats a security queue longer than five minutes as a breach of customer experience, and uses the forecasts to schedule predictive maintenance, cleaning and service interruptions into quieter periods.

Heathrow records its trained models in MLflow, the open-source tool for tracking models and promoting them from development to production, and in 2024 was looking at Delta Sharing to pass data to airlines, ground handlers and other Heathrow companies. Databricks has not described the forecasting method itself.

Every figure here is Databricks' own, published in September 2024 with named Heathrow staff, and none has been measured independently. There has been no public update on the forecasting work since then.

Sources and verification (2)

Databricks press release · Evidence grade B

databricks.com · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

Heathrow security-lane rostering on Azure Machine Learning forecasts · Reported · unverified

Heathrow's security planning team had moved from spreadsheets to Azure Machine Learning to forecast how many passengers will reach security at each terminal in 15-minute intervals, Business Insider reported in May 2019. Those forecasts feed the rosters that set how many security lanes are open.

Stuart Birrell, Heathrow's chief information officer at the time, told Business Insider that the predictions go to 20,000 staff across 400 companies at the airport, and that they help UK Border Force staff immigration. The stack described in that interview combines SQL Azure storage, Azure Stream Analytics for real-time event processing, Azure Data Lake Analytics and Power BI dashboards.

Microsoft's UK Stories site carries an earlier account from Birrell, first published in October 2018 and last updated in April 2024. In it he says Heathrow focuses its data on Azure and Power BI for planning, scheduling and communication with Border Force, and puts the airport's direct workforce at around 7,000 and the number of people working across the airport each day at 76,000.

The newest first-hand source for this record is Business Insider's interview of May 2019. Neither source publishes forecast accuracy or queue outcomes, and no Heathrow statement from 2025 or 2026 confirms the system is still in use, so the record is marked unverified.

Sources and verification (2)

Business Insider · Evidence grade A

Microsoft UK Stories · Evidence grade B

A source supports a reported claim; it does not automatically demonstrate measurable operating benefit.

What another airport can take from this

Heathrow's examples point to distinct operational questions: can camera observations prompt faster turnaround interventions; can forecasts improve staffing and missed-connection decisions; and can visual AI support runway awareness without eroding controller authority?

The useful next question for any airport considering a similar system is not simply who supplies it. It is which operational decision changes, what baseline is used, and which measured result could establish improvement.

Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.

Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.