Public Sector × Project-Based Work
Best Public Sector AI Agencies for Project-Based Work (September 2026)
Public Sector AI agencies with verified project-based client evidence — ranked by depth of documented work, then editorial quality.
Each listed agency documents project-based engagements on its own site or in a published case study; the model is never inferred from a service list.
Ranked agencies for project-based clients
Rankings updated · Newest review
Top 30 of 33 agencies with documented project-based work.
Dublin, Ireland
Client evidence: 20 documented clients overall
Best for: Irish organizations already running Dynamics 365 or Power Platform that want Copilot and Azure AI advice from a Microsoft partner.
Berlin, Germany
Client evidence: 19 documented clients overall
Best for: German enterprises and federal or city administrations taking a machine-learning or LLM use case from feasibility to an operated system.
ontolux
#3Berlin, Germany
Client evidence: 13 documented clients overall
Best for: German public bodies, broadcasters and publishers that need tagging and search over large German-language text collections.
Madrid, Spain
Client evidence: 10 documented clients overall
Best for: Spanish utilities, public hospitals and large employers bringing a forecasting, clinical-data or Spanish-language modeling problem.
Babel
#5Madrid, Spain
Client evidence: 9 documented clients overall
Best for: Spanish and Portuguese banks, insurers and public administrations wanting AI delivered by a large integrator with the data work around it.
London, United Kingdom
Client evidence: 8 documented clients overall
Best for: Pharma, research and public-sector buyers with an NLP problem on unstructured documents that has to survive review.
Berlin, Germany
Client evidence: 8 documented clients overall
Best for: Enterprises putting voice or chat in front of real customers, where a wrong answer reaches the public.
Bismart
#8Barcelona, Spain
Client evidence: 6 documented clients overall
Best for: Organizations whose AI plans depend on fixing the data layer first, especially in healthcare, public services and tourism.
MS iHub
#9Ljubljana, Slovenia
Client evidence: 6 documented clients overall
Best for: Manufacturers and online retailers commissioning a first scoped AI feature or MVP from a small Ljubljana team.
Mercury Labs
#10London, United Kingdom
Client evidence: 6 documented clients overall
Best for: UK public bodies and publicly funded research programs building AI tools that keep expert reviewers in charge.
Helsinki, Finland
Client evidence: 5 documented clients overall
Best for: Regulated Finnish and Nordic enterprises that need an AI system governed, certified and operated, not just delivered.
Ghent, Belgium
Client evidence: 5 documented clients overall
Best for: Enterprises adding AI to a product or support workflow they already run, with the delivery team embedded alongside their own.
Adnovum
#13Zurich, Switzerland
Client evidence: 5 documented clients overall
Best for: Swiss institutions wanting a conversational AI system built into their own cloud and then operated for them.
App4You
#14Gdańsk, Poland
Client evidence: 5 documented clients overall
Best for: Polish public institutions and smaller operators that want a fixed-scope build with the price published before the first call.
coompanion
#15Munich, Germany
Client evidence: 5 documented clients overall
Best for: Mittelstand and startup operations teams automating a document or ticketing workflow on tools their own staff will keep running.
Nebuli
#16London, United Kingdom
Client evidence: 5 documented clients overall
Best for: Organizations wanting a private, self-hosted generative AI workspace over their own documents rather than a public LLM.
Munich, Germany
Client evidence: 4 documented clients overall
Best for: Regulated German industrial and mid-market firms that want the model running inside their own Azure or AWS tenant.
Ergo
#18Dublin, Ireland
Client evidence: 4 documented clients overall
Best for: Irish public institutions and regulated financial firms adding AI to a Microsoft estate they already run.
Contiamo
#19Berlin, Germany
Client evidence: 4 documented clients overall
Best for: German municipal utilities, housing companies and consumer brands adding GPT-based assistants to customer service and content work.
Kruso
#20Copenhagen, Denmark
Client evidence: 4 documented clients overall
Best for: Consumer brands and public bodies adding AI search or an assistant to a commerce or content platform the same firm builds and operates.
Dublin, Ireland
Client evidence: 4 documented clients overall
Best for: Consumer-facing telecom and utility operators, and Irish public bodies, that need models tested and evidenced rather than built.
Faculty
#22London, United Kingdom
Client evidence: 3 documented clients overall
Best for: Enterprises and public bodies that need AI deployed under regulatory or safety scrutiny.
dida
#23Berlin, Germany
Client evidence: 3 documented clients overall
Best for: Organizations with a hard perception or document problem who want the method documented, not just the result.
Sngular
#24Madrid, Spain
Client evidence: 3 documented clients overall
Best for: Enterprises staffing a multi-year engineering program in infrastructure, banking or back-office automation with a single partner.
Berlin, Germany
Client evidence: 3 documented clients overall
Best for: Research institutes and public bodies with a hard technical problem and a named specialist to answer for it.
wegewerk
#26Berlin, Germany
Client evidence: 3 documented clients overall
Best for: Nonprofits and associations adding AI to a digital operation that already carries accessibility and procurement obligations.
Berlin, Germany
Client evidence: 3 documented clients overall
Best for: Public bodies and institutions that need an AI system they can operate, inspect and publish afterwards.
Grepton
#28Budapest, Hungary
Client evidence: 3 documented clients overall
Best for: Hungarian companies on Microsoft Dynamics 365 or Azure that want AI added inside the ERP, CRM and data systems they already run.
Amsterdam, Netherlands
Client evidence: 2 documented clients overall
Best for: Regulated institutions and retailers with a pricing, risk or monitoring problem that can be measured.
Dublin, Ireland
Client evidence: 2 documented clients overall
Best for: Teams that want the senior consultant doing the build, with the architecture and its limits written down before delivery.
About this list
Public-sector projects are judged by what left the building, because the administration runs what it bought. Merantix Momentum has run Hamburg's InnoTecHH innovation and governance process since 2022, with 120+ use cases evaluated in 2023, 19 in development and 8 scaled; its Federal Chancellery work ended as a proof of concept on a locally hosted Mixtral model handed over for integration, and its DZSF noise-barrier detector shipped as an open-source plugin. Fast Data Science trained the Information Commissioner's Office email classifier inside an isolated environment on data that could not be retained. Instituto de Ingeniería del Conocimiento's reservoir-inflow forecasting for Canal de Isabel II is implemented, while its SERMAS allergy-coding work is a pilot. Babel was awarded the governance and monitoring office of the Andalusia AI Center, covering model inventories and EU AI Act compliance.
For an administration buying a fixed-scope build, ask what form the handover took: an open-source release, a locally hosted model, a system with a named operator, or a proof of concept awaiting integration. Then ask who inside the organization runs it now; the curator notes several Merantix references are funded research consortia or feasibility studies. Read the tag against the evidence: Bismart's named public work for Barcelona de Serveis Municipals and Segittur is Power BI data integration, not a model; Mercury Labs' IfATE and MTC pages publish no metric, stack or live status; MS iHub's procurement-analysis platform is ongoing with no outcome stated; and Siili Solutions' production agent, Bobotin, runs at an insurer.
Intro by Gabor Kiss, curator · How we rank
Expert Insight
Why Public Sector experience matters
Procurement fluency—Tender rules, framework agreements, and evaluation criteria shape what can be bought and how. Specialists write compliant bids and structure engagements to fit thresholds; agencies new to public buying lose months learning the process on your calendar
AI Act depth where it bites hardest—Many public-sector uses sit in the EU AI Act's high-risk tier, with risk-management, logging, human-oversight, and registration duties. Specialists deliver the documentation as a work product, not a promise
Explainability as an acceptance criterion—Automated decisions touching citizens must be explainable to the citizen, the caseworker, and eventually an auditor. Specialists design plain-language explanation into the system; retrofitting it after a complaint is the expensive version
Accountability-grade delivery—Public projects end up in audit reports. Specialists document decisions, data sources, and model changes to the standard an oversight body applies, and they design human handoff into every citizen-facing flow rather than treating it as a failure state
Frequently asked questions
33 agencies in our directory combine verified project-based client evidence with documented Public Sector work. The current top-ranked are Storm Technology, Merantix Momentum, ontolux — ordered by depth of documented client evidence, then our editorial scoring (portfolio quality, credibility, completeness); placement is never paid.
Published rates across this page's agencies run €45–260 per hour (median ~€108). Project totals depend on scope — the rate index at /rates breaks the computed bands down by region, country and team size.
Every agency here documents project-based work on its own site or in a published case study, describing the engagement structure rather than a service-list claim. We never infer the model from portfolio tone, so this list only contains agencies that operate project-based engagements deliberately.
2 of the agencies on this page document an embedded / team-extension working model. See the team-extension page for the full verified list, or check the engagement chips on individual profiles.