A tender for design services is due in the morning, and the BIM execution plan is the last document still open. I am Dr Sepehr Abrishami, this is Reimagining AEC, and in a season on the ISO standards I read that plan as someone who once founded a design office of my own. This episode is about the BIM execution plan, BIM being building information modelling, as ISO 19650 Part 2 sets it out at clause five point three point two and clause five point four point one.
Here is the case we will carry throughout, and it is hypothetical. Denholm District Council is procuring multi-disciplinary design services for the Ferrybank Community Hub, a two-storey building whose concept design is complete at the end of Stage 2 of the Plan of Work of the Royal Institute of British Architects. Your organisation is an architecture-led consultancy, bidding to lead a delivery team of three task teams, architecture, structural engineering and building services, two of them from sub-consultants. The council has sent fourteen numbered exchange information requirements. Thirteen can be answered. The eleventh cannot, because it asks for the level of information need to be specified in accordance with BS EN 17412-1 of 2020, a standard that has been withdrawn. It is the evening before submission. You can write a smooth paragraph around it and hope nobody notices. You can ignore it. Or you can tell the client its own tender is wrong, which feels like the fastest way to look difficult. None of those choices is obviously right, and whichever you make, the plan you submit tonight is the plan you may be held to for the rest of the appointment.
So the question for this episode is how one plan serves both the tender and the appointment. By the end you will be able to build a pre-appointment plan from the standard itself, answer every requirement traceably, raise a defective one without absorbing it, confirm the plan after award as the standard intends, and place it among the documents that govern the appointment. We will take it in six steps. First the document and its two states. Then reading the tender, drafting the skeleton, confirming the plan, the appointment document set, and last, completing a template as a series of decisions. Ferrybank stays with us all the way.
The first step is to settle what the document is. ISO 19650-2 defines the BIM execution plan as the plan explaining how the delivery team will carry out the information management aspects of the appointment. Strip that back and you have a method statement. Not for design or construction, but for information. It says who manages information in the team, how the team divides into task teams, how the models will be federated, who produces what and when, what software, hardware and infrastructure the team will run, and where the team needs the project's own rules adjusted to deliver. Hold on to the phrase method statement, because it is the sharpest test of every sentence in the plan.
The commonest failure is to treat the plan as bid marketing. A plan padded with capability statements reads well, commits the team to nothing and defends it from nothing. A weak plan leaves the lead appointed party blamed for every information failure that follows. A serious plan creates the contractual position the team defends from. Serious means testable. Every sentence should do one of three jobs. It answers a requirement, it names a person, a tool or a date, or it records an assumption the client needs to see. Anything else belongs in the covering letter.
The standard gives the document two states with two qualifiers. The delivery team's pre-appointment BIM execution plan is the tender document. It is established by the prospective lead appointed party under clause five point three point two, and its job is to show the appointing party a credible approach and the capability and capacity to deliver it. The delivery team's BIM execution plan, once confirmed under clause five point four point one, is the appointment document. It is agreed with each appointed party and then governs information delivery under change control. One thread, two states, two jobs. Most of the confusion in practice comes from using the bare name for both, when they have different contractual status and different work to do.
Authorship needs settling too. The appointing party never writes the plan. It writes the requirements, above all the exchange information requirements, and the plan is the delivery team's answer to them. At tender the author is the prospective lead appointed party, writing for the whole proposed team, including organisations not yet engaged. That is why the standard insists on confirmation in agreement with each appointed party after award. And the plan never travels alone. Clause five point three point seven has the prospective lead appointed party compile four things into its tender response, where they are available. The pre-appointment plan, the summary of the team's capability and capacity, the proposed mobilisation plan and the information delivery risk assessment. The plan is the centrepiece. The capability summary shows the approach can be staffed, the mobilisation plan shows how the team gets running, and the risk assessment records what the plan had to assume. Evaluators read the four side by side, against the tender response requirements they set under clause five point two point three.
The document has a history. UK practice under Publicly Available Specification 1192-2, known as PAS 1192-2, already used a BIM execution plan written against employer's information requirements, and ISO 19650 carried the name across in 2018 while the words around it changed. The employer became the appointing party, and employer's information requirements became exchange information requirements, keeping the same initials. There is also a forward look, and it is only that. Draft revisions of Parts 1 and 2 were open for public comment from March to May 2026, and a draft revision of Part 3 followed, open for comment until July 2026. Among their proposals is renaming this document the information production plan, as part of a move from the language of BIM towards information management. None of the drafts is in force, and no publication date has been announced. Part 2 of 2018 remains the edition in force, and it is what your clients will cite. The document's job does not change if its name does.
At Ferrybank, the document due in the morning is the pre-appointment plan. Your consultancy writes it for a team whose two sub-consultants are not yet engaged, and only after award will it be confirmed with them. So the first step resolves. The plan is a method statement for information, written by the delivery team, in a tender state that proposes and an appointment state that governs.
The second step is reading the tender, because plans fail at evaluation when their authors answer the requirements they assumed rather than the ones the client sent. Under clause five point two point four the appointing party compiles the invitation to tender information, and the plan author needs all of it. The exchange information requirements are the question paper. The project's information standard sets the rules for structuring, classifying and coding information, and its information production methods and procedures set the rules for producing, checking and delivering it. The reference information and shared resources are what you inherit. The information protocol is the legal instrument the plan will sit beneath. And the milestones, response requirements and evaluation criteria tell you when information is due and how you will be scored.
Each exchange information requirement has five parts under clause five point two point one. The requirement itself, its level of information need, its acceptance criteria, its supporting information and its dates. The episode Information Management in ISO 19650 Part 1: Concepts and Principles explains where those parts come from. For the plan author the rule is to read all five. A requirement whose acceptance criteria you cannot meet, or whose cited standard you do not recognise, is not something to skim past. It is a finding.
Then the heart of the step. Clause five point three point two lists what the prospective lead appointed party shall consider when establishing the pre-appointment plan, and there are six areas. The wording is shall consider, not shall include. You document each in proportion to the project, and on a small appointment an area may resolve in a paragraph, but you may not ignore one. They group under three questions.
The first question is who will manage the information. The first area is the people proposed to undertake the information management function, with names and professional résumés. The function can be delegated but never deleted, and clause five point two point three tells the evaluators to assess the competence of whoever you put forward. Name real, available people, because the confirmed plan will hold you to them. And record the boundary of the role, because an organisation that offers a coordinator can find itself running the whole function unpaid.
The second question is how the information will be produced. The second area is the information delivery strategy, with four ingredients. Your approach to the requirements, objectives for collaborative production that say what good looks like beyond bare compliance, an overview of the team's organisation and commercial relationships, and its composition as task teams. Evaluators read it first, because it tells them whether the rest of the plan rests on anything. If the council wanted fortnightly exchanges and your structural task team were a four-person practice, this is where you would show the arithmetic works. The third area is the federation strategy, how the discipline models combine, how they are broken down, the coordination cadence and who runs the federated model. The fourth is the high-level responsibility matrix, which allocates each element of the information model to a task team with its key deliverables. At tender you allocate elements, not containers, and the test is that no element is orphaned. The detailed matrix comes after appointment, and the episode Roles and the Responsibility Matrix takes it further.
The third question is what has to change for delivery to work. The fifth area is the one weak plans leave empty, proposed additions or amendments to the project's information standard and to its methods and procedures. It is a formal voice back to the client at tender. If something in the standard is unworkable, or no longer current, this is where you say so in writing. The sixth area is the schedule of software, hardware and information technology infrastructure, with versions. Versions are not paperwork. Interoperability between task teams depends on them, and the confirmed plan will lock them. Be honest about what you have, and where an exchange between two packages has not been tested, say that the test happens at mobilisation. That is a promise someone can check.
One idea cuts across all six, the level of information need, the framework for saying how much information is enough across geometry, alphanumerical information and documentation. It is specified today by ISO 7817-1 of 2024, which superseded and withdrew BS EN 17412-1 of 2020. The client states it against each requirement, and the plan responds to it rather than inventing its own. The episode Level of Information Need in ISO 7817-1 covers the method.
At Ferrybank, three of the fourteen requirements shape the plan most. The third asks for federated design models at each information exchange, coordinated across the three disciplines, which drives the federation strategy and the matrix. The seventh asks for asset data drops in a named spreadsheet structure at Stages 3 and 4, which drives the methods and the delivery approach. And the eleventh cites the withdrawn standard. The second step ends, then, with a reading list for the bid. Five parts to every requirement, six areas to consider, and one requirement already flagged as a finding.
The third step is drafting the skeleton, and the first decision is structure. The skeleton is built from the standard itself rather than from a house template, which means you can rebuild it on any project, with any house template or none. It has eight sections. Section zero is document control and a short statement that this is the pre-appointment plan submitted under the tender. Sections one to six are the six areas, in the order clause five point three point two lists them, so an evaluator can map the plan to the standard without a crosswalk. Section seven closes on assumptions and issues, with cross-references to the three companion documents of clause five point three point seven.
Four drafting techniques do most of the work. The first is to answer point by point, through a response matrix in section two. One row per requirement, with the plan section that answers it, a one-line summary and a status drawn from four values. Met. Met with an amendment proposed. Assumption recorded. Issue raised. For the evaluator, that turns an afternoon of cross-reading into a scan. For you it is a completeness check, because a requirement with no row has not been dealt with, and silence in a tender response is read as acceptance. Ferrybank's matrix has fourteen rows, and thirteen read met.
The second technique is to make every promise testable. Committed, experienced and industry-leading cannot be checked, so they count for nothing. The same intent with dates, names, cadences and versions can be verified or falsified three months into the appointment, and that is what a governing document needs, because an untestable sentence is an unenforceable one. Draft a sentence, then ask what evidence would show it had been kept. If there is no answer, prune it or sharpen it.
The third is to keep the proposed apart from the assumed. Proposed is what your team commits to deliver. Assumed is what you rely on someone else to provide, such as the survey the council says exists, the common data environment configured before mobilisation, or the sub-consultant not yet under contract. Every assumption goes into section seven and is mirrored in the delivery team's risk register, because clause five point three point six expects risks arising from the team's assumptions about the requirements to be recorded there. At confirmation those assumptions become the agenda. An assumption you recorded is a negotiating position. One you did not record is your problem, at your cost.
The fourth technique is to raise defects rather than absorb them, and the commonest defect is a superseded reference. Tenders are assembled from precedents, and precedents age. A requirement drafted three years ago cites what was current three years ago. Level of information need is the live example, but you will also meet residue from the PAS 1192 era, such as employer's information requirements as a title, level of detail used as the specification method, or deprecated status codes in a document register. Each is a defect, and the plan gives you a formal route for it.
Here is that route at Ferrybank, in four short entries. An issue entry identifies requirement eleven and quotes its citation exactly. A finding states that BS EN 17412-1 was superseded and withdrawn by ISO 7817-1 of 2024. A proposed amendment in the section five register, against the information standard, asks the council to substitute the 2024 standard. And an interim assumption states that the team will apply the 2024 framework pending confirmation, with a matching entry in the risk assessment. Raising a defect is not pedantry. It is professional judgement, and a client who knows the standard reads it as competence. Absorbing the defect signs your team up to deliver it.
Most failing pre-appointment plans fail in one of five ways, and each has a structural defence. A plan reused wholesale from the last project cannot answer requirements it has never seen, and the response matrix exposes it. Scope creep on the function is defeated by the boundary recorded in section one. Promised capability that nobody has assessed is defeated by sequence, because clause five point three point three has each task team assess its capability and capacity, and the strategy should be written after those assessments, not before. Padding fails the testability question, and absorbed defects fail the fourth technique. None of these defences relies on anyone's vigilance on a deadline evening. That is the third step's resolution. A skeleton built from the standard, answered row by row, carries its own defences.
The fourth step is confirmation. The award letter has arrived, Ferrybank is yours, and the document has to change state. The word the standard chose is confirm, not reissue or finalise, and clause five point four point one attaches a condition that does the real work. The lead appointed party confirms the plan in agreement with each appointed party. Two things changed at award. The word prospective dropped away, and with it a bidder's safe distance. And the audience changed. At tender the plan spoke to evaluators. Now it binds the delivery team, and the sub-consultants who were names in a proposal are organisations whose obligations flow from this document. A plan confirmed without their agreement is one the task teams will quietly ignore.
Clause five point four point one then sets out six actions, and they map almost one to one onto the skeleton. Confirm the names of the individuals undertaking the information management function. Update the information delivery strategy, as required. Update the high-level responsibility matrix, as required, which concedes that the appointment as won rarely matches the tender as bid. Confirm and document the delivery team's information production methods and procedures. Agree with the appointing party any additions or amendments to the project's information standard. And confirm the schedule of software, hardware and infrastructure, at which point the versions stop being intentions and become commitments.
Confirmation changes the sentences. Tender language says we propose, we intend, subject to, and that is correct, because a bidder cannot commit organisations it has not appointed. Confirmed language has no proposal verbs and no conditions, and every placeholder has become a name or a date. Each softened sentence at tender is a question confirmation must answer. Who, in the end. Which version. Was the amendment agreed or refused. A confirmed plan that still says propose and intend has not been confirmed. It has been photocopied.
The fifth action closes the loop the tender opened. Every amendment proposed in section five is now agreed or refused. If it is agreed, the information standard is updated, and because that standard is an appointment document, the agreement has contractual effect. If it is refused, the requirement stands, and the recorded assumption becomes a live risk with a management action. What may not happen is silence, because an unanswered amendment leaves the plan and the information standard disagreeing for the rest of the appointment. The confirmed plan is also not an archive item. Staff change, software needs a patch, a milestone moves, and each change either reaches the plan or the plan drifts into fiction. ISO 19650-3, in its confirmation clause for the operational phase, says it plainly, that the plan is communicated to all appointed parties, kept under review and recirculated as necessary. Part 2 achieves the same through change control, which is where the fifth step goes.
Back at Ferrybank, three weeks after award, two things have happened. The council has accepted the amendment on level of information need, so the information standard now cites ISO 7817-1 of 2024, and the issue raised at tender resolves exactly as proposed. Raising it formally is why it resolved cleanly. Then the complication. The building services sub-consultant has declined the appointment over fee, and a replacement practice has joined. It authors in a different services package, so the technology schedule changes, the exchange between platforms needs a test at mobilisation, and the matrix needs a new name. Trace both events through the six actions and each lands somewhere specific.
The pitfalls at confirmation share one cause, a step taken before the agreement or settlement it depends on. Photocopy confirmation reissues the tender text under a new cover. Confirming before the level of information need is settled builds on ground that will move. And skipping the agreement of an appointed party usually happens to the one that joined late, exactly like Ferrybank's replacement practice, which is precisely the one most likely to hold assumptions nobody has checked. Unanswered amendments and names left as placeholders belong to the same family, and an informed appointing party should refuse a plan that carries them. That settles the fourth step. Confirmation is an agreement with every appointed party, run through six actions, that turns proposals into facts.
The fifth step puts the confirmed plan in company. Clause five point four point six names the completed appointment documents for the lead appointed party, and there are five. The appointing party's exchange information requirements. The project's information standard, with any amendments agreed at confirmation. The project's information protocol, likewise. The delivery team's confirmed BIM execution plan. And the master information delivery plan. Put plainly, the requirements, the rules, the legal wrapper, the method and the schedule. At Ferrybank that means fourteen requirements as amended, an information standard that now cites the 2024 standard, the protocol as issued, the plan agreed with both sub-consultants including the replacement practice, and a master plan aggregated from the three task team plans and baselined against the two information exchanges.
The same clause attaches one condition to the whole set. The documents are managed via change control for the duration of the appointment. A change to any of them is proposed, assessed for impact, agreed by the right parties and recorded, rather than simply made. The five change together, because they reference one another. If the council revises a requirement, the plan's response and the master plan's dates both move. Version history lets anyone establish what governed the appointment on a given date, which matters when delivery is disputed, and the 2019 UK guidance on concepts adds the commercial point, that how change control operates shapes a party's rights to claim additional time and cost.
One tier down, clause five point four point seven mirrors the set for each appointed party. It carries the lead appointed party's own exchange information requirements, the client's passed down with additions, the same information standard and protocol, the same confirmed plan, because there is one plan for the whole delivery team, and in place of the master plan, that party's agreed task information delivery plan. Requirements cascade and are particularised at each tier. The rules and the plan are shared. Only the schedule is sliced. The episode The Delivery Plans takes those plans in depth.
The plan is a junction. It cites four positions it does not own. The level of information need, the detailed responsibility matrix, the security position, which enters through the protocol and the triage of ISO 19650-5, and the common data environment rules and container naming, which live in the information standard. The plan references each of them, with a date, and copies none, because a plan that copies its neighbours falls out of date twice as fast. Each of those positions has its own episode, including Security-Minded Information Management in ISO 19650 Part 5 and The Common Data Environment: Process Before Platform.
Being a junction is also why sequence matters, and this is the warning at the centre of the episode. Confirm the plan before the level of information need is settled, and expect rework on every container you publish. The level of information need defines what each container must carry. While it is unsettled, task teams model to guesses. When the specification lands, the guesses are checked against it, and the gap is remodelled at the delivery team's cost, container by container, across the appointment. Ferrybank shows the pattern working. The defect was raised at tender, the amendment was agreed at confirmation, and the plan was confirmed onto settled ground. Settle first, then sign.
Small practices always ask whether this applies to them. The UK guidance on the information management function answers with a client's project manager appointed as lead appointed party with no appointed parties beneath it. That organisation is itself a delivery team, made up of its internal task teams. It still needs a plan, which the guidance calls essentially a method statement for information delivery, and a master information delivery plan, which here is effectively a task plan. The one thing it can drop is issuing its own exchange information requirements, because there is nobody to issue them to. Size changes the weight of the documents, not the set. The pattern also transfers to operation. ISO 19650-3 runs the same appointment with asset flavoured resources, an asset information standard, asset production methods and an asset protocol alongside the plan and the master plan, under change control. So the fifth step resolves. The confirmed plan is one of five documents that change together, it references its neighbours rather than swallowing them, and it is signed only once the ground under it is settled.
The sixth step is completing a template, told as decisions rather than form filling. Suppose the house template the Ferrybank team holds is a post-appointment document, written from the perspective of a contractor leading the team, so a design-led team like Ferrybank's adapts the framing and keeps the structure. Mapping the skeleton into it leaves no orphans. Section zero becomes document control, now stating that the plan is confirmed under clause five point four point one. The proposed people become named, committed people. The strategy carries across, and the response matrix becomes a permanent appendix, the fastest audit trail from requirement to answer. The amendment registers become the record of each outcome, the technology schedule is locked, and the assumptions distribute to the risk register and the change control record. Where the appointment has not fixed a value, a placeholder stays in square brackets rather than an invented one.
Every entry then falls into one of three kinds. Carried content was right at tender and is right now, such as the federation logic and the objectives, and it moves with at most a change of tense. Confirmed content was a proposal and is now a fact, names, versions, each amendment's outcome and the date each appointed party agreed. Added content could not have existed at tender, such as the reference to the baselined master plan, the detailed matrix and the mobilisation confirmations. At Ferrybank the confirmation edits run like this. The information manager, the deputy and three task team leads are named, including the replacement practice's lead, with that practice's agreement recorded. The team had also proposed, at tender, a test of the exchange between the structural analysis package and the architectural platform. That amendment was agreed, so the test becomes a scheduled mobilisation task. The amendment on level of information need was agreed, so the plan cites the 2024 standard and the tender issue formally closes. The technology schedule takes the new services package, and document control reads confirmed and agreed, with a date.
Templates age faster than standards, so the last decision is to audit the template before it binds you. This one lags in five places. It presents published status codes in A and B series, when the National Annex revised with the corrigendum of February 2021 deprecates the B codes, so a confirmed plan drops them and uses the A codes for published, contractual information. It works through level of development grades as the operative specification, when those are now history and mapping only, and Ferrybank's own amendment has already made the level of information need contractual. Its container identifier uses field names from before the revised annex renamed them. It cites PAS 1192-6 for health and safety information, superseded by ISO 19650-6 in 2025. And one reference still says employer's information requirements. Correct each one, and record why.
Then write a change note of one page, saying what changed from the tender plan entry by entry, why it changed, and who agreed, with dates. It is what the client will want when it reviews the confirmed plan against the bid it evaluated, and it is the first entry in the change control history. Teams that skip it find out months later that nobody can say when the plan last matched reality. The sixth step comes down to this. A template is completed by deciding what is carried, confirmed or added, and by correcting where it lags the standard.
Let me bring the threads together. The question was how one plan serves both the tender and the appointment. It does so by changing state deliberately. At tender it proposes, built from the six areas of clause five point three point two, answering every requirement row by row and stating what it assumes. At award it is confirmed under clause five point four point one, in agreement with every appointed party, through six actions that turn each proposal into a fact. And it then governs as one of five appointment documents under change control. The thread that joins the two states is the record. Ferrybank's eleventh requirement was raised as an issue at tender, answered as an agreed amendment at confirmation, and closed in the plan that now governs.
Three rules come out of Ferrybank's tender. Write every sentence of a plan so that someone could check, three months on, whether it was kept. When a tender cites something defective or withdrawn, raise it through an amendment and a recorded assumption, because silence is read as acceptance. And never confirm a plan onto unsettled ground, above all an unsettled level of information need, because the rework arrives one container at a time.
For reading, take clauses five point three and five point four of ISO 19650-2, with its UK National Annex, ISO 7817-1 for the level of information need, and the UK guidance on the delivery processes and on the information management function, now maintained under the IMI Framework, named after the Information Management Initiative. The Delivery Phase in ISO 19650 Part 2 places the plan in the whole delivery process, from assessment and need to project close-out. The next episode, The Delivery Plans, is about how many task team plans become one master plan whose dates hold.
Take a BIM execution plan from a project you know, at work or in research, mark each sentence as proposed or confirmed, and ask which of them anyone ever agreed. I am Dr Sepehr Abrishami, and that marking test is the one Reimagining AEC would set any plan.