Updates
Dash Now Remembers How You Work
Memory, Skills, and the first look at what comes next.
GovDash users have been able to draft proposal sections faster, run compliance checks that used to take hours, and navigate RFP requirements that trip up generic AI tools. The questions we kept coming across were, “Can Dash learn from them over time? Can it carry company knowledge into every workflow so it builds on every interaction?”
Dash Memory and Dash Skills are the result of the thesis that, if Dash can build and learn from your prior interactions, it will evolve with your company. Memory establishes the preferences, standards, and company-specific context from one session, carrying them into every session that follows.
Skills define the processes your company has refined over the years that run the same way every time, for every person on the team, without anyone having to re-explain them.
Together, they close the gap between an AI that knows government contracting and an AI that knows your company.
What Customers Asked Us to Build
The most consistent request across customer conversations was the need for Dash to learn and apply firm-specific context. Users asked Dash to maintain persistent style constraints, apply firm-specific acronym definitions, understand their personas, and ensure a consistent writing cadence across all team output.
In a 2025 poll of 275 proposal professionals by Lohfeld Consulting, 42 percent named AI hallucination as their top concern, and 33 percent named generic AI writing as their second. Both challenges point to the same need: an AI that knows the company well enough to produce specific, accurate content. Memory and Skills are how Dash gets there.
How Dash Has Evolved
Dash launched as a purpose-built AI for government contracting, trained on federal procurement language and built around FAR structures, proposal workflows, and evaluation criteria that generic AI tools do not understand at a working level. That GovCon foundation is what made Dash useful from day one.
Memory and Skills build on that foundation by adding knowledge specific to your firm. There are three layers, and they work together.
Phase 1: Domain knowledge. Dash understands government contracting at a level that general-purpose AI tools cannot replicate. FAR compliance, acquisition terminology, proposal structure, and evaluation criteria language are built in. This is the baseline.
Phase 2: Company knowledge. Dash learns who you are and how you work. Your role, preferences, agency focus, and writing voice accumulate from your actual usage over time. This is Memory.
Phase 3: Process knowledge. Dash knows how your firm runs each specific task: in what order, with what standards, producing what output. This is Skills.
A proposal manager running a compliance review today gets output shaped by all three layers: GovCon expertise, their personal preferences from Memory, and their firm's specific compliance process from the Skill. That combination is not something any general-purpose AI tool can replicate, even with memory features turned on.
How Dash Memory Works
Dash Memory builds a working picture of who you are from your conversations. It is not a document library or a raw log of everything you have typed. It is a curated understanding of your role, working style, and preferences that grows more accurate over time.
In practice, after a month of regular use, Dash knows things like your preferred proposal structure (e.g., Technical Volume drafts around the PWS, Management Volume organization), specific compliance rules, and your firm’s favored language, including writing prohibitions like "leverage" as a verb and "we are pleased to." None of that needs to be typed again, and it applies to every session.
How Dash Skills Work
If Memory captures who you are and how you work, Skills capture how your firm runs each specific task.
A Skill is a reusable playbook. It encodes a multi-step process in enough detail that Dash runs it the same way every time, for every person on the team. Think of it as what you would hand a new senior proposal writer on day one: not just login credentials and templates, but the SOPs.
How your company structures a Technical Volume.
How you run color reviews.
What your compliance threshold is for a showstopper versus a score risk.
A go/no-go Skill can understand your company's specific criteria and how to weight them: incumbent status, customer relationship, NAICS fit, past performance relevance, competitive positioning, and the bid and proposal cost threshold your leadership uses to make the call.
Memory and Skills Together
Memory is passive: it accumulates from your conversations without configuration and applies your preferences in the background. Skills are explicit: you define them so processes are precise and consistent across the whole team.
When a proposal manager runs a compliance review, Memory brings their agency focus and personal style standards. The Skill brings the firm's process, severity framework, and output format. The result reflects both without the user specifying either. For teams with multiple writers on the same proposal, this produces consistent output at every desk, not because of a reminder email, but because the standards are encoded.
Writing That Sounds Like Your Company
When AI lacks company-specific context, language tends toward the generic: "innovative solutions," "proven track record," "uniquely positioned to deliver." Experienced evaluators recognize this pattern immediately. Memory and Skills give Dash the firm's actual proof points, voice, and word choices so output is specific rather than approximate.
Where This Is Going: Agents
Memory and Skills are also the foundation for what comes next. A Skill defines what Dash should do and how. An Agent runs that Skill automatically, on a trigger, without requiring you to start a prompt or session.
Memory | Skills | Agents | |
|---|---|---|---|
What it is | A growing picture of who you are and how you work | A reusable playbook for a specific firm process | An automated worker that runs Skills without being asked |
What it does | Carries your preferences, voice, and firm context from session to session | Encodes multi-step workflows so Dash runs them the same way every time | Triggers Skills automatically on a schedule or event |
How it's built | Accumulates passively from your conversations | Defined explicitly by you or your team | Configured to watch for a trigger and run a Skill |
Who it's for | Individual users | The whole team | The whole team |
Example | Dash knows you prime on OASIS+, avoid "leverage" as a verb, and structure Technical Volumes around the PWS | Dash runs your firm's compliance review: sections, severity levels, output format | Dash catches a SAM.gov amendment and runs your Amendment Analyzer Skill before standup |
Status | Live | Live | Coming soon |
For a capture team, this means an agent that catches SAM.gov amendments, runs the Amendment Analyzer Skill, and surfaces flagged changes before the morning standup. For a proposal manager, it means an agent that runs the compliance review Skill on each section as writers submit and updates the running log in real time.
The contractors building their Skill library now will be able to deploy Agents with minimal configuration when the capability goes live. Skills built today carry forward.
GovDash Memory and Skills are live in your account. Learn how to create your first skill by reading the Help Center article.
Not on GovDash yet? Book a demo today.
