AI agents for proposals and commercial communication at an engineering firm
From the email that arrives to the proposal sent and followed up — with the human decision kept at the critical points.
The client
An engineering and construction firm operating nationally and internationally, specialising in technical design, site supervision, consultancy, construction management and industrialised building solutions.
It works with property developers, architects, contractors, public bodies and industrial clients.
For confidentiality reasons, this scenario does not identify a specific organisation. It represents a real implementation developed by Strong Answer for companies with similar commercial and technical processes.
The challenge
The firm received commercial requests every day across several mailboxes:
- Proposal requests
- Technical enquiries
- Quote requests
- Tender invitations
- Requests for clarification
- Changes to project scope
- Drawings and documentation
- Follow-up requests
- Client and partner communications
Each opportunity required someone to analyse the email, consult earlier documentation and gather information from several sources. Most of that time went on repetitive work:
- Reading and classifying emails
- Searching for previous proposals
- Hunting for technical references
- Identifying missing information
- Drafting initial replies
- Producing commercial documents
- Updating the pipeline
- Chasing opportunities that had gone quiet
The information was scattered across Outlook, Teams, SharePoint, OneDrive, Word documents, spreadsheets and management applications.
The solution
We implemented a model of artificial-intelligence agents integrated with Microsoft 365.
The solution turns a commercial email into a structured process, from the moment the request arrives through to the drafting of the proposal and the follow-up of the opportunity.
Main components
Five agents, one process
Each agent handles one stage. Together they cover the full journey of a commercial opportunity.
1. Commercial triage
Analyses incoming emails and identifies company, contact, request type, engineering discipline, location, response deadline, attached documentation, priority, commercial potential and missing information. It then routes to the right person.
- New proposal request
- Tender
- Quote request
- Project change
- Technical query
- Documentation received
- Commercial follow-up
- Partner or supplier
- Administrative communication
2. Opportunity qualification
Turns the content of the email into a structured record: project name, client, location, type of building or infrastructure, estimated area, disciplines requested, project stage, deadlines, documents received and mandatory requirements. Where data is missing, it drafts an email with the questions needed.
3. Knowledge search
Searches the authorised internal information — previous proposals, similar projects, presentations, team CVs, methodologies, technical references, price tables, contract templates and tender documentation — and puts the most relevant material for the new opportunity in front of the account manager.
4. Proposal drafting
Generates a first version from the approved templates: context, understanding of the project, objectives, scope, disciplines, methodology, phases, deliverables, team, schedule, assumptions, exclusions, commercial terms and next steps. Figures and contractual terms remain subject to human validation.
5. Communication and follow-up
Prepares the covering email for review and approval. After sending, it records the opportunity in the pipeline and follows up on proposals without a reply, requests for clarification, meetings, changes and tender deadlines — creating tasks in Planner, Lists or the CRM.
One request, from start to finish
A client asks for a proposal covering engineering services and technical supervision of an industrial facility.
01
Automatic analysis
The agent identifies: a commercial request, an industrial project, a need for civil and electrical engineering, attached documentation, a ten-day response deadline and no information about the total floor area.
02
Qualification
An opportunity is created in the pipeline with the extracted data.
03
Search
The agent finds three earlier projects with similar characteristics.
04
Drafting
A first version of the proposal is generated, with methodology, phases, deliverables and references.
05
Validation
The technical director confirms the scope and the commercial department validates the figures.
06
Sending
The agent prepares the final email; the proposal goes out once approved.
07
Follow-up
With no reply, the system prepares a follow-up email five working days later.
Three levels of autonomy
What the agent does on its own, what needs approval and what it never does — agreed in writing before we start.
Automatic
The agent acts without intervention.
- Classifying emails
- Extracting information
- Organising attachments
- Creating opportunities
- Assigning owners
- Setting priorities
- Creating tasks
- Deadline alerts
Requires approval
The agent prepares and waits for a person.
- Commercial replies
- Requests for information
- Proposals
- Changes to scope
- Figures
- Contractual terms
- Sending documentation to the client
Restricted
The agent never sends these on its own.
- Financial commitments
- Contracts
- Price changes
- Final technical opinions
- Legal communications
- Decisions with contractual impact
It is this separation that makes the solution usable in a company where a badly sent proposal has contractual consequences.
What the project sets out to achieve
In a scenario of this kind, the objectives are:
- Reduce the time spent analysing commercial requests
- Cut the time needed to prepare proposals
- Centralise the company's knowledge
- Ensure consistency across communications
- Reduce forgotten opportunities
- Improve deadline compliance
- Increase commercial capacity without growing the team proportionally
- Free engineers and managers from repetitive administrative work
Illustrative indicators
The figures below are reference objectives for a project of this kind — they are not results measured at a specific client.
- A 50% to 80% reduction in the time needed to prepare the first version of a proposal
- Initial email classification in under a minute
- Less time spent searching for documents
- Follow-up on 100% of open proposals
- Fewer commercial requests left unanswered
- Greater reuse of existing knowledge and documentation
What Strong Answer does
We analyse the company's current processes and build a solution adapted to its reality. The project may include:
- Diagnosis of the workflows
- Document organisation
- SharePoint structuring
- Building the knowledge base
- Agent configuration
- Integration with Outlook and Teams
- Workflow development
- CRM or ERP integration
- Defining permissions
- Building dashboards
- User training
- Ongoing support and improvement
The implementation is not simply about installing an artificial-intelligence tool. The goal is to create an integrated operational process.
Email received → qualification → search → proposal → approval → sending → follow-up
Artificial intelligence inside the processes you already have
We build agents that work with the tools, the documents and the processes your organisation already uses.
The result is a faster, better-structured and more controlled commercial operation — while always keeping the human decision at the critical points.
Need any clarification?
Tell us your challenge. We reply quickly and with no commitment whatsoever.
Lagoa · Algarve
Monday – Friday
(09:00 – 13:00) — (14:00 – 19:00)