At London Tech Week on June 9, 2026 the tech ecosystem pivoted visibly toward cross company AI agents as startups and incumbents showcased systems designed to act across enterprise boundaries. One of the most watched debuts was ArchAstro, a team built by veterans from Microsoft Meta and Stripe, which released a public beta of an autonomous data system that promises to orchestrate workflows across multiple corporate data stacks. The announcement marked a broader shift from single vendor assistants to agent architectures that coordinate services, pipelines and compliance across organisational boundaries.
What cross company agents aim to solve
Enterprises have long struggled with siloed data, fractured workflows and costly integration projects. Cross company agents propose to reduce friction by serving as autonomous coordinators that can access authorised APIs, reconcile differing data schemas and execute multi party processes such as supplier onboarding invoicing and joint analytics. For business leaders the appeal is practical: fewer custom integrations lower run rates and faster collaboration yields quicker time to value for joint initiatives with partners and customers.
That promise translates into concrete gains for everyday workers when agents remove repetitive tasks. Procurement specialists may see purchase orders auto matched and approved within defined risk thresholds. Customer success teams can pull verified partner data into a single view without manual exports. For people who spend much of their day stitching systems together, an effective agent becomes a force multiplier rather than a toy for technologists.
ArchAstro and the new wave of autonomous data systems
ArchAstro presented a public beta that links to multiple cloud data warehouses, identity providers and document stores while enforcing policy driven access and lineage tracking. The demo showed an agent that ingests a cross company dataset requests explicit consent where required and runs a validated analytics pipeline to produce a joint report, all while logging provenance for auditors. The team emphasised governance primitives as central to the design, aiming to balance autonomy with traceability.
Other startups showcased complementary approaches: service meshes for agent orchestration, standardised message brokers to translate domain models and secure vaults for multi party secrets management. Collectively these projects point to a new tooling layer that sits between business logic and raw infrastructure, offering higher level primitives for safe cross company automation.
Technical and regulatory challenges ahead
Despite promising demos, deploying agents across corporate boundaries raises thorny technical and legal questions. Data privacy and consent are immediate obstacles when agents move or combine records belonging to different controllers. Intercompany SLAs and liability models must be rethought for autonomous actions that can trigger financial or reputational harm. Security architects at the conference stressed the need for cryptographic attestations, fine grained authorisation and auditable decision logs to ensure accountability.
Regulators also face new questions. If an agent mediates a cross border data transfer or makes a credit decision, which jurisdiction applies and who is responsible for redress? These are not theoretical queries; companies piloting agents will likely encounter regulators and legal teams early, creating a feedback loop that shapes product design and deployment strategies.
Business model implications and market reactions
Vendors pursuing cross company agents are experimenting with various commercial models. Some plan usage based pricing tied to API calls and compute. Others propose value sharing where vendors capture a fraction of efficiency gains achieved through automation. Platform companies hope to earn recurring revenue through governance subscriptions and certified connectors. Investors at the event showed strong interest, viewing agents as a potential new layer of enterprise software that could command high margins if they solve critical coordination problems.
Meanwhile potential customers weigh trade offs. Large enterprises with complex ecosystems may benefit most but also face higher governance burdens. Smaller companies may gain access to partners they could not integrate with previously, but they may worry about vendor lock in or pricing that scales with success.
Practical use cases highlighted
-
– Supply chain reconciliation where agents match shipments invoices and payments between manufacturers and logistics providers, reducing manual exceptions.
– Collaborative analytics for cross border clinical trials where patient level data is harmonised and aggregated under strict consent and provenance controls.
– Joint marketing campaigns that require synchronous asset distribution, performance reporting and revenue sharing across agencies and publishers.
Human dimension and workforce impact
Speakers emphasised that agents do not simply replace people; they reassign human effort to higher value work. A product manager we spoke with described a pilot where an agent handled routine reconciliation tasks, allowing finance staff to focus on exception strategy and supplier relationship management. Yet the human transition requires investment in training and governance roles that monitor agent behavior and resolve complex edge cases.
There are also equity considerations. Firms must design reskilling pathways to avoid widening gaps between technical and non technical employees and to ensure that human oversight roles are accessible and well compensated. Organisations that pair technical pilots with clear workforce strategies will likely see smoother adoption and better outcomes.
Standards, interoperability and the role of open protocols
Interoperability is foundational for cross company agents. Several panels called for open protocols for consent exchange data schemas and action authorisations so agents can operate predictably across vendors. Industry bodies and standards organisations are starting to convene working groups to define shared ontologies and exchange formats that prevent a patchwork of incompatible agent ecosystems.
Open standards also matter for competition. If a few dominant platforms control the plumbing, smaller firms may struggle to offer trusted connectors. Conversely, community led standards could democratise access and encourage innovation in specialised agents for sectors such as healthcare energy and logistics.
Security and auditability as design principles
Security architects argued that auditability must be a first class citizen. Systems that produce immutable evidence of decisions, cryptographic proofs of data origin and tamper resistant logs will be easier to certify for regulated industries. Several vendors unveiled features for explainable decision trails and human in the loop checkpoints designed to ensure that autonomous actions are reversible and accountable.
The emphasis on auditability also reflects a pragmatic need to gain buyer trust. Enterprises will accept autonomy only when they can interrogate an agent s reasoning and trace outcomes to authorised inputs and policies.
What to watch next
Adoption will hinge on three things: robust governance tooling that eases legal and compliance friction, open standards that enable cross vendor cooperation and concrete pilots that demonstrate measurable ROI while maintaining audit trails. Early adopters will likely be consortia of companies with aligned incentives, such as trade groups and regulated industries with clear collaboration needs.
For observers interested in technical developments the IEEE and other standard bodies are likely to publish guidance in the coming months, while industry consortia will produce reference implementations to accelerate safe experimentation.
Where to follow developments
For ongoing coverage of technical standards and enterprise pilots consult major industry publications and standards organisations that track interoperability efforts. Technical readers may find detailed working papers and protocol drafts on sites maintained by standards bodies and academic labs focused on multi party computation and secure data exchange. For regulatory signals the European Data Protection Board and national regulators will publish guidance as real world pilots surface compliance questions.
Readers can explore standards discussions and policy briefings at organisations such as IEEE and follow regulatory updates through national data protection authorities.
A reasoned outlook
The rise of cross company AI agents showcased at London Tech Week points to a practical ambition in enterprise software: reduce the cost of cooperation while maintaining accountability. The technology is promising and the demos are compelling, yet the path to safe, large scale adoption will run through governance frameworks, legal clarity and standards that ensure agents act transparently. If those pieces fall into place, cross company agents could free people from repetitive work and enable new forms of collaboration that were previously too costly to manage. For now the industry is at an important inflection point where engineering, law and human centred design must converge to move pilots into production responsibly.
